SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES
  8. Supporting Information

Mycophenolate mofetil (MMF) and sirolimus (SRL) have been used for calcineurin inhibitor (CNI) minimization to reduce nephrotoxicity following liver transplantation. In this prospective, open-label, multicenter study, patients undergoing transplantation from July 2005 to June 2007 who were maintained on MMF/CNI were randomized 4 to 12 weeks after transplantation to receive MMF/SRL (n = 148) or continue MMF/CNI (n = 145) and included in the intent-to-treat population. The primary efficacy endpoints were the mean percentage change in the calculated glomerular filtration rate (GFR) and a composite of biopsy-proven acute rejection (BPAR), graft lost, death, and lost to follow-up 12 months after transplantation. Patients were followed for a median of 519 days after randomization. MMF/SRL was associated with a significantly greater renal function improvement from baseline with a mean percentage change in GFR of 19.7 ± 40.6 (versus 1.2 ± 39.9 for MMF/CNI, P = 0.0012). The composite endpoint demonstrated the noninferiority of MMF/SRL versus MMF/CNI (16.4% versus 15.4%, 90% confidence interval = −7.1% to 9.0%). The incidence of BPAR was significantly greater with MMF/SRL (12.2%) versus MMF/CNI (4.1%, P = 0.02). Graft loss (including death) occurred in 3.4% of the MMF/SRL-treated patients and in 8.3% of the MMF/CNI-treated patients (P = 0.04). Malignancy-related deaths were less frequent with MMF/SRL. Adverse events caused withdrawal for 34.2% of the MMF/SRL-treated patients and for 24.1% of the MMF/CNI-treated patients (P = 0.06). The use of MMF/SRL is an option for liver transplant recipients who can benefit from improved renal function but is associated with an increased risk of rejection (but not graft loss). Liver Transpl 19:675–689, 2013.. © 2013 AASLD.

Although many factors, including hypertension, diabetes, and hepatitis C virus (HCV) infection, contribute to renal insufficiency after liver transplantation, calcineurin inhibitor (CNI) use likely plays a major role. CNI toxicity has been identified as the cause of end-stage renal disease in 73% of liver transplant recipients presenting with chronic renal failure.[1, 2] Mycophenolate mofetil (MMF) and sirolimus (SRL) generally have not been associated with nephrotoxicity[3, 4] and thus have been used in CNI-sparing strategies in liver transplantation.[5-9] However, a recent study in renal transplant patients randomized to SRL or cyclosporine A (CSA) found that although kidney function was similar between the groups, SRL was associated with glomerular damage and tubulotoxicity.[10] The relevance of these findings to liver transplantation is unknown.[10]

Studies examining the influence of SRL conversion on renal function in liver transplant recipients have shown conflicting results.[5-9] In 2 single-center randomized controlled trials,[7, 11] early renal function improvements were significantly in favor of the SRL group at 3 months, and modest improvements were still evident at 12 months without a significant increase in the risk of acute rejection. In a case-controlled, retrospective review of patients from a single transplant center, there was no significant difference in creatinine clearance between those converted to SRL and those on a low-dose CNI regimen, with renal function stabilizing in both groups.[5] Patients may have been converted from CNIs to SRL at a time when renal damage was irreversible. A nonsignificant improvement in the glomerular filtration rate (GFR) was noted in a meta-analysis of data from observational studies and controlled trials.[12]

Early clinical trial data indicated that the de novo use of SRL in liver transplant recipients was associated with an increased risk of hepatic artery thrombosis (HAT), graft loss, and mortality.[13] In other studies, liver transplant recipients receiving SRL (most within 24 hours) were found to have a greater risk of wound infections than those not receiving SRL.[14, 15] As a result of these adverse event reports, a black-box warning on the use of SRL in liver transplant patients was mandated for the product labeling by the Food and Drug Administration.[13] Delaying the use of SRL for 4 to 6 weeks after transplantation, however, was believed to be attractive because it might avoid the early problems associated with HAT and wound healing and allow patients to be spared from CNIs and the subsequent nephrotoxicity.[16] This prospective, open-label, multicenter, randomized, comparative trial (the Spare-the-Nephron trial) was undertaken to determine whether conversion to MMF/SRL from MMF/CNI-based immunosuppression in the early postoperative period (within 4 to 12 weeks after transplantation) would be effective in preserving renal function while providing adequate protection against acute rejection.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES
  8. Supporting Information

Study Population

Recipients of a single primary orthotopic liver transplant were enrolled if they met the screening criteria (Table 1). Eligibility for randomization between 4 and 12 weeks after transplantation included maintenance on stable doses of MMF ≥ 1 g/day at least 7 days beforehand and no more than 1 prior episode of corticosteroid-sensitive biopsy-proven acute rejection (BPAR). Excluded from randomization were patients with any corticosteroid-resistant acute rejection requiring antilymphocyte antibody therapy; active sepsis within the prior 2 weeks; a hemoglobin level < 8 g/dL, a white blood cell count < 3000/mm3, or a platelet count < 70,000/mm3 within the prior week; total bilirubin, aspartate aminotransferase, or alanine aminotransferase levels 2 times the upper limit of normal or a clinically significant elevation within the prior 2 weeks; a calculated GFR[17] < 30 mL/minute within the prior week; dialysis for >2 weeks between transplantation and randomization; a malignancy [including the diagnosis of new or recurrent hepatocellular carcinoma (HCC)]; gastrointestinal disorders that might interfere with absorption; or previous treatment with SRL.

Table 1. Eligibility Criteria
Inclusion CriteriaExclusion Criteria
• Patients 18 to 74 years of age• Recipients of single primary orthotopic liver transplantation• Patients receiving or scheduled to receive (within 12 hours of study entry) MMF/CNI• For HCV positivity, biopsy either after reperfusion or at the time of randomization• Graft with a cold ischemia time > 15 hours or from a non–heart-beating or ABO-incompatible donor• Multiple organ transplants• Malignancy in the previous 5 years other than successfully treated nonmelanomatous skin cancer or hepatoma with no extrahepatic spread• Severe diarrhea or other gastrointestinal disorders that might interfere with the absorption of oral medications• Pregnancy or lactation• Known contraindications to treatment with SRL• Treatment or expected treatment after study entry with azathioprine, everolimus, methotrexate, cyclophosphamide, enteric-coated mycophenolate sodium, alemtuzumab, or bile acid sequestrants• Treatment with any investigational agent within 30 days of study entry

Study Design

Potential patients were screened from 1 week before transplantation until 1 week before randomization and were rescreened for eligibility at the time of randomization (Fig. 1). Patients were stratified into groups by their HCV status and by whether they were receiving CSA or tacrolimus (TAC) at the time of randomization. Within each stratified group, patients were randomized (1:1) to continue receiving MMF/CNI or to be converted to MMF/SRL maintenance therapy (Fig. 1) in an open-label design. Randomization numbers were sponsor-generated in blocks with equal treatment allocation in each block to achieve equal allocation in both groups overall and within each stratification factor. Randomization codes were loaded into an independent vendor's centralized interactive voice-response system operated and maintained by the vendor.

image

Figure 1. Trial design.

Download figure to PowerPoint

For patients randomized to the MMF/SRL group, SRL (2-4 mg once daily) was initiated within 24 hours after the last dose of CSA or TAC and was adjusted to achieve and maintain trough levels of 5 to 10 ng/mL. TAC and CSA were administered in 2 doses to maintain whole blood trough concentrations of 3 to 10 and 100 to 250 ng/mL, respectively. All patients continued to receive 1 to 1.5 g of MMF twice daily. Doses were reduced for hematological and gastrointestinal toxicity. Corticosteroids and induction therapy were administered according to center practice. Treatments for underlying or concurrent diseases or adverse events were administered at the investigator's discretion.

After screening, the actual Model for End-Stage Liver Disease (MELD) score and adverse events were recorded. When eligibility was reconfirmed and patients were randomized (baseline), a complete medical history was obtained, and a physical examination performed. At baseline and at each follow-up visit (1, 2, 4, and 6 weeks after randomization and 6 and 12 months after transplantation), blood chemistry, hepatic enzyme profiles, GFR (calculated with the 6-variable Modification of Diet in Renal Disease equation),[17] creatinine clearance (calculated with the Cockcroft and Gault method[18]), urinary protein/creatinine ratio, and CSA/TAC/SRL whole blood trough levels were assessed. HCV RNA levels were collected at baseline and 6 and 12 months after transplantation. HCV recurrence was defined as the histological confirmation of an HCV infection [ie, a liver biopsy sample without evidence of ongoing acute cellular rejection (an absence of endotheliitis)] and Batts/Ludwig grade 3 inflammation or Batts/Ludwig stage 2 fibrosis that resulted in the initiation of HCV antiviral therapy. Mycophenolic acid (MPA) trough levels were measured at baseline, 2 weeks after randomization, and 6 months after transplantation. Samples were analyzed in a central laboratory. The Banff international consensus scale[19] was used to evaluate BPAR. Safety data, including adverse events, concurrent diseases, and concomitant treatments, were captured from randomization up to 24 months after transplantation.

Patients could withdraw from the study at any time for any reason. The investigator also could withdraw patients for intercurrent illness, adverse events, treatment failure after a prescribed procedure, protocol violations, and administrative or other reasons.

This study was in full conformance with the Declaration of Helsinki and the Guideline for Good Clinical Practice. The institutional review boards of the participating sites approved the protocol, and all participants provided signed informed consent. The study is registered at ClinicalTrials.gov (NCT00118742).

Study Endpoints

The primary efficacy endpoint was the mean percentage change in renal function (determined by calculated GFR[17]) from baseline (randomization) to month 12 after transplantation. The coprimary endpoint was the proportion of patients experiencing BPAR, graft loss (retransplantation or death), or death or lost to follow-up from baseline to month 12 after transplantation. Acute rejection episodes were defined as Banff histological criteria (≥grade II) for moderate or severe acute cellular rejection and a total rejection activity index score ≥ 4. Secondary endpoints are presented in Table 2. Safety endpoints included adverse events, changes in the immunosuppressive regimen, and clinically significant laboratory and vital sign abnormalities. HAT, thrombotic events, wound complications, noninfectious pneumonitis, opportunistic infections, and malignancies were the events of interest.

Table 2. Secondary Endpoints
• Mean percentage change in serum creatinine and creatinine clearance from baseline to month 12 after transplantation• Proportion of patients with BPAR from baseline to month 12 after transplantation• Proportion of patients with corticosteroid-resistant BPAR• Time from transplantation to first BPAR• Proportion of patients with treatment failure at month 12 after transplantation [defined as a composite of BPAR, graft loss (retransplantation or death), use of an additional maintenance immunosuppressive medication not specified in the assigned treatment group, discontinuation of any assigned study regimen for >14 consecutive days or >30 cumulative days, and lost to follow-up]• Proportions of deaths and patients with graft loss (retransplantation or death)• Time to death and graft loss (retransplantation or death) from transplantation to month 12 after transplantation• Proportion of patients with histologically confirmed HCV recurrence, fibrosis scores, and changes from baseline in HCV RNA levels among HCV-positive patients• Presumptive HCV recurrence in HCV-positive patients based on adverse event reporting and/or biopsy results that did not meet the criteria for histologically confirmed HCV recurrence• Mean trough concentrations 12 months after transplantation• Proportions of patients experiencing hypertension, diabetes mellitus, and hyperlipidemia after baseline and changes from baseline in lipid profiles and blood pressure at months 6 and 12 after transplantation

Statistical Analyses

The study was designed to determine whether conversion from MMF/CNI to MMF/SRL could improve renal function assessed as the mean percentage change in the calculated GFR from baseline to 12 months after transplantation (superiority) while providing a comparable rate of BPAR, graft loss, death, or lost to follow-up (noninferiority). To test both hypotheses with at least 80% power and an overall significance level of 0.05, approximately 340 patients were to enroll; a minimum of a 15% greater improvement in the calculated GFR and no greater than 12% for BPAR, graft loss, death, or lost to follow-up for the MMF/SRL group versus the MMF/CNI group were assumed.

The primary and coprimary endpoints were analyzed with both an intention-to-treat (ITT) population (all randomized patients who received at least 1 dose of the study medication and who had at least 1 postrandomization assessment) and a per-protocol (PP) population (a subset that met the major study entry criteria and received at least 30 days of treatment). Secondary endpoints were analyzed using the ITT population. Safety data were analyzed using the safety population (all patients who received any component of the assigned study regimen and who had at least 1 safety assessment after the first dose of the study medication). Demographic and baseline characteristics were summarized for the ITT population.

The primary endpoint was analyzed with an analysis of covariance model, which included the treatment, baseline HCV and HCC status, and baseline CNI as factors; the baseline calculated GFR and time from transplantation to randomization as covariates; and the interaction terms of treatment by baseline calculated GFR and treatment by baseline HCC status. The proportion of patients with BPAR, graft loss, or death or lost to follow-up from randomization to 12 months after transplantation was analyzed with the Farrington-Manning method.[20] A 1-sided 95% upper limit was calculated for the difference of 2 proportions. Noninferiority of MMF/SRL to MMF/CNI continuation was inferred if the upper limit was <0.12. For other endpoints, continuous data were analyzed using model similar to that for the primary endpoint; for categorical data, the Cochran-Mantel-Haenszel test or the chi-square test was used; and for time-to-event data, the Kaplan-Meier method and the log-rank test were used as appropriate. For the primary endpoint, patients with dialysis or graft loss were imputed with a prespecified lowest possible calculated GFR value (15 mL/minute/1.7 m2). A last observation carried forward analysis also was performed for the ITT population. For the coprimary endpoint, patients who withdrew early and did not reach the lower boundary of a 12-month visit window were treated as lost to follow-up and, therefore, met the endpoint (an event). The total number of patients in each arm and the number of patients with nonmissing assessments are presented in all tables.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES
  8. Supporting Information

Patient Disposition and Baseline Characteristics

This study was conducted at 43 centers in the United States and Canada, and 332 patients were enrolled in all. Two hundred ninety-four patients were randomized (1-22 patients at individual centers) between August 2005 and July 2007. More than 90% of the patients in each group had been receiving TAC before randomization. The patient populations are reviewed in Fig. 2. Patients were followed for a median of 519 days after randomization.

image

Figure 2. Overview of the analysis populations. *Patients who completed the treatment period (randomization to 12 months after transplantation). Reasons for withdrawal are listed in Table 4. †One patient who was randomized to MMF/SRL received only MMF after randomization. This patient was included in the safety population in the MMF/CNI group because he had been receiving MMF and TAC before randomization. This patient was excluded from the ITT population because he did not meet dosing requirements.

Download figure to PowerPoint

The 2 study groups were comparable (Table 3) except for baseline calculated GFR, which was higher in the MMF/SRL group. There was a low rate of prior rejections in both treatment groups, possibly because of the exclusion for corticosteroid-resistant rejection episodes at the time of entry. The rates of HCV positivity and concomitant HCC were comparable. In all, 178 patients (61%) completed treatment through 12 months after transplantation: 82 (55%) in the MMF/SRL group and 96 (66%) in the MMF/CNI group. The overall withdrawal rates were 45.0% in the MMF/SRL group and 33.8% in the MMF/CNI group (P = 0.05). The majority of withdrawals were due to adverse events [51/149 (34.2%) in the MMF/SRL group and 35/145 (24.1%) in the MMF/CNI group, P = 0.06; Tables 4 and 5]. There were no between-group differences in the incidence of adverse events leading to premature discontinuation of the study treatment that were statistically significantly different at the 0.05 level.

Table 3. Demographics and Baseline Characteristics of the ITT Population
 MMF/SRL GroupMMF/CNI Group
Characteristic(n = 148)(n = 145)
  1. a

    P > 0.05 for between-group comparison using the Cochran-Mantel-Haenszel test stratified by baseline HCV status, baseline HCC status, and baseline CNI.

  2. b

    P > 0.05 for between-group comparison using an analysis of variance model with terms for treatment, baseline HCV status, baseline HCC status, and baseline CNI.

  3. c

    P > 0.05 for between-group comparison using the Cochran-Mantel-Haenszel test stratified by baseline CNI.

  4. d

    P < 0.05 for between-group comparison of the treatment groups using an analysis of variance model with terms for treatment, baseline HCV status, baseline HCC status, and baseline CNI.

Sex [n (%)]a  
Male110 (74.3)110 (75.9)
Female38 (25.7)35 (24.1)
Age (years)  
Mean ± SD53.2 ± 8.853.4 ± 7.9
Median53.054.0
Minimum, maximum27, 7521, 69
Race [n (%)]a  
Caucasian125 (84.5)126 (86.9)
African American8 (5.4)11 (7.6)
Asian/Pacific Islander11 (7.4)3 (2.1)
Native American4 (2.7)3 (2.1)
Other0 (0.0)2 (1.4)
HCV-positive [n (%)]65 (43.9)61 (42.1)
HCC-positive [n (%)]44 (29.7)43 (29.7)
HCV- and HCC-positive [n (%)]29 (19.6)30 (20.7)
MELD score: physiological/natural  
Mean ± SD (n)b19.1 ± 8.46 (133)19.0 ± 7.29 (125)
Median18.019.0
Minimum, maximum2, 458, 33
MELD score: actual/exception  
Mean ± SD (n)b24.0 ± 5.72 (78)22.4 ± 6.09 (89)
Median23.022.0
Minimum, maximum6, 407, 40
Primary reason for liver transplant reported by investigator [n (%)]c  
HCV36 (24.3)31 (21.4)
HCV and HCC29 (19.6)30 (20.7)
Cirrhosis, nonspecific27 (18.2)40 (27.6)
HCC15 (10.1)13 (9.0)
Primary sclerosing cholangitis12 (8.1)8 (5.5)
Primary biliary cirrhosis10 (6.8)9 (6.2)
Alcoholic cirrhosis3 (2.0)8 (5.5)
Hepatitis B1 (0.7)0 (0.0)
Other15 (10.1)6 (4.1)
Days from transplantation to randomization  
Mean ± SDb53.9 ± 12.851.8 ± 11.3
Median54.053.0
Minimum, maximum24, 9829, 88
Use of induction therapy [n (%)]  
Antithymocyte globulin19 (13.0)17 (12.0)
Basiliximab11 (7.0)11 (8.0)
Daclizumab15 (10.0)13 (9.0)
Rejection episode before randomization [n (%)]5 (3.4)8 (5.5)
Panel reactive antibody level ≥ 20%: highest assessment [n/N (%)]3/41 (7.3)7/33 (21.2)
Patients receiving CMV prophylaxis [n (%)]47 (31.8)41 (28.3)
Calculated GFR (mL/minute)  
Mean ± SD (n)d54.3 ± 16.21 (146)50.6 ± 16.08 (142)
Median51.249.6
Minimum, maximum26, 10425, 119
Calculated creatinine clearance (mL/minute)  
Mean ± SD (n)72.1 ± 22.70 (146)68.0 ± 22.26 (143)
Median69.563.6
Minimum, maximum29, 13229, 141
Table 4. Reasons for Patient Withdrawal: All Patients
 MMF/SRL GroupMMF/CNI Group
Reason for Withdrawal(n = 149)(n = 145)
  1. NOTE: The data are presented as numbers and percentages. P values are based on the chi-square test.

  2. a

    P = 0.11.

  3. b

    P = 0.06.

  4. c

    P = 0.05.

Safetya53 (35.6)39 (26.9)
Adverse eventb51 (34.2)35 (24.1)
Death2 (1.3)4 (2.8)
Nonsafety14 (9.4)10 (6.9)
Insufficient therapeutic response6 (4.0)1 (0.7)
Refused treatment3 (2.0)5 (3.4)
Failure to return2 (1.3)2 (1.4)
Other3 (2.0)2 (1.4)
Total67 (45.0)49 (33.8)c
Table 5. Adverse Events Leading to Premature Withdrawals Occurring in ≥1% of Patients in Either Treatment Group: Safety Population
 MMF/SRL GroupMMF/CNI Group
Adverse Event(n = 148)(n = 146)
  1. NOTE: The data are presented as numbers and per-centages.

Leukopenia4 (2.7)5 (3.4)
Diarrhea4 (2.7)4 (2.7)
Neutropenia1 (0.7)3 (2.1)
Rash4 (2.7)0
HCV2 (1.4)1 (0.7)
Mouth ulceration3 (2.0)0
Peripheral edema3 (2.0)0
CMV viremia03 (2.1)
Pyrexia3 (2.0)0
Renal failure03 (2.1)
Proteinuria2 (1.4)0

Clinical Outcomes

At 12 months, patients in the MMF/SRL treatment group had a significantly greater improvement from baseline in calculated GFR (the primary endpoint; all available data in the ITT population) in comparison with those in the MMF/CNI group (P = 0.0012; Fig. 3). The difference between the groups, adjusted by the analysis model, was 24.9% [95% confidence interval (CI) = 14.8%-35.0%, P = 0.002], and this indicated that patients in the MMF/SRL group experienced greater improvements in renal function 12 months after transplantation. The last observation carried forward and PP analyses showed consistent results (P = 0.001 and P = 0.004, respectively). The level of missingness was <15%. The subgroup of patients diagnosed with HCC at baseline who were randomized to the MMF/SRL regimen had a greater improvement (percentage change from baseline to 12 months = 25.6 ± 38.4) than those receiving MMF/CNI (percentage change = −9.1 ± 41.8, P = 0.002, n = 39 in each treatment group). Both HCV-negative and HCV-positive patients in the MMF/SRL group had greater improvements in GFR versus patients in the MMF/CNI group (percentage change from baseline to 12 months, for HCV-negative patients, 14.8 ± 43.5 versus 1.1 ± 32.9, P = 0.01; for HCV-positive patients, 25.6 ± 36.4 versus 1.4 ± 48.5, P = 0.03). There were significant between-group differences in mean GFR at 6 and 12 months and in mean serum creatinine levels at 12 months (Fig. 4). The percentage changes from baseline to 12 months in serum creatinine and creatinine clearance were significantly different between the treatment groups (Table 6).[18]

image

Figure 3. Mean percentage increase in calculated GFR (6-variable Modification of Diet in Renal Disease equation) from baseline to 12 months after transplantation (the primary efficacy endpoint). The P value for the between-group comparison was calculated using an analysis of covariance model with terms for treatment, baseline HCV status, baseline HCC status, and baseline CNI as factors; baseline calculated GFR and time from transplantation to randomization as covariates; and interaction terms of treatment by baseline calculated GFR and treatment by baseline HCC status (n = 130 for the MMF/SRL group and n = 124 for the MMF/CNI group).

Download figure to PowerPoint

image

Figure 4. (A) Mean GFR and (B) mean serum creatinine level at key time points during the study for each treatment group.

Download figure to PowerPoint

Table 6. Serum Creatinine and Creatinine Clearance at Baseline and Month 12: ITT Population
 MMF/SRL GroupMMF/CNI Group 
Outcome Measure(n = 148)(n = 145)P Valuea
  1. NOTE: The data are presented as means and SDs (with n values in parentheses).

  2. a

    P values for between-group comparisons using an analysis of variance model with terms for treatment, baseline HCV status, baseline HCC status, and baseline CNI as factors and with baseline measurement and time from transplantation to randomization as covariates.

  3. b

    Cockcroft and Gault method.[18]

Serum creatinine (μmol/L)   
Baseline100.4 ± 24.88 (146)108.4 ± 29.36 (143) 
Month 12109.5 ± 102.45 (136)154.0 ± 152.73 (129) 
% change from baseline11.7 ± 107.39 (134)49.5 ± 157.26 (127)0.012
Calculated creatinine clearance (mL/minute)b   
Baseline72.1 ± 22.70 (146)68.0 ± 22.26 (143) 
Month 1278.6 ± 27.61 (136)64.7 ± 28.02 (129) 
% change from baseline14.0 ± 35.16 (134)−3.0 ± 35.86 (127)<0.001

In the coprimary endpoint analysis, the percentage of patients in the PP population with BPAR, graft loss, or death or were lost to follow-up within 12 months after transplantation was noninferior in the MMF/SRL group (16.4%) versus the MMF/CNI group (15.4%, difference in proportion = 0.9%, 90% CI = −7.1% to 9.0%; Table 7).[20] The ITT analysis showed consistent results (18.9% versus 16.6%, difference in proportion = 2.4%, 90% CI = −5.1% to 9.8%). The proportion of patients with BPAR was significantly greater in the MMF/SRL group (18/148 or 12.2%) versus the MMF/CNI group (6/145 or 4.1%, P = 0.02), whereas the proportions with corticosteroid-resistant acute rejection episodes among those with BPAR were similar between groups [3/18 (16.7%) and 1/6 (16.7%) for the MMF/SRL and MMF/CNI groups, respectively]. The majority of the BPAR episodes (21/24 or 87.5%) occurred within the first 6 months after transplantation and were treated with only corticosteroids. Two patients in the MMF/CNI group and 2 patients in the MMF/SRL group required antithymocyte globulin and/or monoclonal antibody treatments. All 12 patients (66.7%) in the MMF/SRL group who withdrew from the study treatment medication switched to treatment with a CNI.

Table 7. Coprimary Endpoint: The Proportions of Patients With BPAR or Graft Loss (Retransplantation or Death) or Lost to Follow-Up at Month 12: PP Population
 MMF/SRL GroupMMF/CNI Group
Endpoint(n = 116)(n = 123)
  1. NOTE: The between-group difference in proportions was 0.9% (90% CI = −7.1% to 9.0%); the incidence rate for the MMF/SRL group was noninferior to the rate for the MMF/CNI group because the upper CI limit was <12%. The CI was calculated with the Farrington-Manning method.[20]

  2. a

    Events were mutually exclusive; only the first event counted for each patient.

Composite endpoint [n (%)]a19 (16.4)19 (15.4)
BPAR125
Graft loss (including death)49
Lost to follow-up35

Based on Kaplan-Meier product limit estimates, the distribution of the time from transplantation to first BPAR was significantly different between the groups (P = 0.01); overall, the first episodes of BPAR occurred earlier in the MMF/SRL group. Differences between the treatment groups in the distribution of the worst Banff schema rejection severity score within 6 and 12 months after transplantation were not significant. The proportions of patients with treatment failure did not differ significantly [72/148 (48.6%) for the MMF/SRL group and 55/145 (37.9%) for the MMF/CNI group, P = 0.08].

Deaths were reported for 4 of the 148 patients (2.7%) in the MMF/SRL group and for 9 of the 145 patients (6.2%) in the MMF/CNI group within the 12 months (P = 0.10). Malignancy was the cause of death in 4 patients receiving MMF/CNI: metastatic HCC, metastatic adenocarcinoma in a patient with HCC, recurrent HCC, and non–small cell lung carcinoma. An additional patient receiving MMF/CNI died of renal failure and cardiac arrest subsequent to recurrent HCC with metastasis after the patient withdrew consent to participate in the study at day 114; that death was not captured in the database. Malignancy (metastatic HCC) was the cause of death in 1 of the 4 patients receiving MMF/SRL who died. Other causes of death were neuroleptic malignant syndrome, acute myocardial infarction, graft failure, recurrent HCV, multisystem organ failure, sepsis, septicemia, and gastrointestinal hemorrhage. The proportions of patients with graft loss (including death) within 12 months were 3.4% (5/148) in the MMF/SRL group and 8.3% (12/145) in the MMF/CNI group (P = 0.04).

The distribution of time to graft loss (including death) differed between treatment groups (P = 0.03): patients in the MMF/CNI group were more likely to die more than 6 months after transplantation than those in the MMF/SRL group. A similar trend was observed for the time to death, although the difference was not significant (P = 0.07). An exploratory analysis also examined the incidence rates of graft loss (including death) by HCC status. Patients with an initial diagnosis of HCC who received MMF/SRL had a lower rate of graft loss (1/44 or 2.3%) than those in the MMF/CNI group (9/43 or 20.9%, P = 0.004). As for patients without a diagnosis of HCC, graft loss was similar in the MMF/SRL group (4/104 or 3.8%) and the MMF/CNI group (3/102 or 2.9%).

The proportions of HCV patients with presumptive HCV recurrence (based on adverse event reporting and/or biopsy results that did not meet the criteria of histologically confirmed HCV recurrence) were slightly higher with MMF/SRL (43/65 or 66.2%) versus MMF/CNI (35/61 or 57.4%) within 12 months (P = 0.32). The mean HCV RNA levels were essentially unchanged from baseline to month 12 in the MMF/CNI group [from 5.9 ± 1.0 (n = 46) to 6.0 ± 0.9 log10 IU/mL (n = 36)] but increased from 6.0 ± 1.1 (n = 47) to 6.4 ± 0.7 log10 IU/mL (n = 27) in the MMF/SRL group (P = 0.04 for the difference in the changes from baseline to month 12 between treatment groups).

Immunosuppressant Doses and Exposure

The mean average daily dose of MMF was not statistically different between the groups throughout the treatment period (Table 8). The average daily doses of SRL, CSA, and TAC showed a trend of decreasing over time (Table 8). Corticosteroid doses (excluding treatment for rejection) decreased over time (Table 8). The mean trough concentrations are presented in Table 9. MPA levels were similar between the groups throughout the study. Therapeutic target levels of MPA have not been validated; other target levels were protocol-defined and were adhered to effectively.

Table 8. Summary of the Average Daily Doses of Study Medications and Corticosteroids: The ITT Population
Average Daily DoseMMF/SRL Group(n = 148)MMF/CNI Group(n = 145)P Valuea
  1. NOTE: The data are presented as means and SDs (with n values in parentheses).

  2. a

    P value for between-group comparisons using an analysis of variance model with terms for treatment, baseline HCV status, baseline HCC status, and baseline CNI as factors.

  3. b

    Days after transplantation.

  4. c

    Excluding treatment for rejection.

MMF (mg)   
First dose to day 90b1685.1 ± 497.1 (147)1736.7 ± 508.7 (144)0.37
Days 91 to 1801568.5 ± 552.2 (135)1590.4 ± 555.9 (139)0.72
Days 181 to 3651569.8 ± 520.3 (109)1571.5 ± 589.2 (118)0.98
CSA (mg)   
First dose to day 90 284.7 ± 146.8 (11) 
Days 91 to 180 256.8 ± 104.7 (11) 
Days 181 to 365 258.3 ± 101.8 (9) 
TAC (mg)   
First dose to day 90 6.5 ± 3.6 (134) 
Days 91 to 180 5.5 ± 3.1 (127) 
Days 181 to 365 4.7 ± 2.6 (110) 
SRL (mg)   
First dose to day 903.4 ± 1.3 (146)  
Days 91 to 1803.4 ± 1.5 (129)  
Days 181 to 3652.9 ± 1.3 (99)  
Corticosteroids (mg)c   
First dose to day 909.3 ± 7.5 (125)7.7 ± 3.9 (123)0.04
Days 91 to 1807.1 ± 7.4 (110)6.2 ± 4.7 (100)0.28
Days 181 to 3655.6 ± 4.6 (87)26.9 ± 162.1 (59)0.24
Table 9. Trough Concentrations of MPA, TAC, SRL, and CSA: Safety Population
ImmunosuppressantBaseline (4-12 Weeks After Transplantation)Visit 3 (2-3 Weeks From Baseline)Visit 6 (Approximately 6 Months After Transplantation)Visit 7 (Approximately 12 Months After Transplantation)
  1. NOTE: The data are presented as means and SDs (with n values in parentheses). The target trough levels were as follows: SRL, 5 to 10 ng/mL; TAC, 3 to 10 ng/mL; and CSA, 100 to 250 ng/mL. When there were multiple measurements within the timeframe, for baseline, the measurement closest to the date of the first trial medication was used; for postbaseline visits, the highest value was used. The following records were excluded: MPA trough levels 112% greater than the MPA concentration at 30 minutes and MPA trough levels greater than 7 μg/mL.

MPA (μg/mL)    
MMF/SRL group1.7 ± 1.3 (121)1.6 ± 1.1 (117)1.7 ± 1.4 (93)Not applicable
MMF/CNI group1.8 ± 1.2 (117)1.8 ± 1.2 (114)1.4 ± 1.0 (103) 
SRL (ng/mL)7.1 ± 5.5 (142)7.6 ± 4.2 (134)8.7 ± 4.2 (104)7.5 ± 3.7 (84)
TAC (ng/mL)9.0 ± 3.4 (127)8.3 ± 3.1 (126)8.5 ± 3.7 (119)7.2 ± 3.0 (95)
CSA (ng/mL)266.3 ± 103.1 (11)210.1 ± 103.1 (10)177.1 ± 124.2 (10)217.3 ± 155.0 (7)
Table 10. Summary of Postrandomization Adverse Events Reported by Investigators on Study Report Forms: Safety Population
 MMF/SRL GroupMMF/CNI Group
Adverse Event(n = 148)(n = 146)
  1. NOTE: The data are presented as numbers and percentages. P values are based on the chi-square test.

  2. a

    P > 0.05.

  3. b

    P < 0.001.

  4. c

    P = 0.009.

  5. d

    P < 0.008; patients with 1 or more opportunistic infection.

  6. e

    The patient underwent retransplantation because of HAT, which was not captured as an adverse event because it occurred more than 30 days after the last dose of the study medication; graft loss was a secondary endpoint and was followed up to 24 months after transplantation.

Between-group difference ≥ 5% and more frequent with MMF/SRL  
Headache39 (26.4)26 (17.8)a
Leukopenia36 (24.3)23 (15.8)a
Pyrexia31 (21.0)22 (15.1)a
Peripheral edema27 (18.2)17 (11.6)a
Rash27 (18.2)5 (3.4)b
Hyperlipidemia/hypercholesterolemia25 (16.9)7 (4.8)c
Mouth ulceration/stomatitis20 (13.5)1 (0.7)b
Between-group difference ≥ 5% and more frequent with MMF/CNI  
Opportunistic infections14 (9.5)30 (20.6)d
Vomiting12 (8.1)20 (13.7)
Hyperkalemia2 (1.4)16 (11.0)
Renal failure2 (1.4)12 (8.2)
Increased blood creatinine010 (6.8)
Hyperglycemia08 (5.5)
Other notable adverse events  
Hypertension17 (11.5)15 (10.3)
Malignancies7 (4.7)10 (6.8)
Diabetes mellitus4 (2.7)1 (0.7)
Proteinuria5 (3.4)3 (2.0)
HAT2 (1.4)1 (0.7)e

For patients who discontinued treatment during the scheduled treatment period (from randomization to 12 months after transplantation), concomitant medications taken on or after the last date of the study medication in the combined MMF/CNI group included CSA (n = 6), azathioprine (n = 1), SRL (n = 8), and TAC (n = 41). In the MMF/SRL group, concomitant medications included SRL (n = 22), TAC (n = 47), CSA (n = 13), and azathioprine (n = 1).

Safety Outcomes

Adverse events with a between-group difference in frequency ≥ 5% and other events of interest are presented in Table 10. Most adverse events were assessed by the investigator to be mild or moderate in intensity and unrelated to the study medication. Mouth ulceration, leukopenia, and hyperlipidemia/hypercholesterolemia (adverse events determined by the investigator) were reported more frequently in the MMF/SRL group (P < 0.001, P = 0.07, and P = 0.001, respectively). Filgrastim therapy was prescribed for 21 patients (14.2%) in the MMF/SRL group and for 20 patients (13.8%) in the MMF/CNI group. HAT occurred in 2 MMF/SRL patients and in 1 MMF/CNI patient. Opportunistic infections were reported in approximately twice as many patients receiving MMF/CNI, with cytomegalovirus (CMV) being the most predominant and occurring in 23 of 146 patients (15.8%) in the MMF/CNI group and in 6 of 148 patients (4.1%) in the MMF/SRL group (P = 0.001). Most of these reports involved CMV syndrome/viremia (20 cases in the MMF/CNI group and 6 cases in the MMF/SRL group, P = 0.004). Hyperkalemia, renal failure, and increased blood creatinine (adverse events determined by the investigator) were reported more frequently in the MMF/CNI group versus the MMF/SRL group (P = 0.001, P = 0.006, and P = 0.001, respectively). There were no reports of noninfectious pneumonitis in either group. Within 12 months, 3 patients per treatment group had recurrent HCC; an additional patient in the MMF/CNI group had a recurrence diagnosed at 13 months.

Table 11. Change From Baseline to Months 6 and 12 After Transplantation in Lipid Profiles and Blood Pressure: Safety Population
 MMF/SRL GroupMMF/CNI Group 
Outcome Measure(n = 148)(n = 146)P Valuea
  1. NOTE: The data are presented as means and SDs (with n values in parentheses). If there were multiple measurements within the time period, for baseline, the measurement closest to the date of the first dose was used; for postbaseline visits, the worst value was used.

  2. a

    From an analysis of covariance model to compare the least square means between groups with terms for treatment, baseline HCV status, baseline HCC status, and baseline CNI type as factors, and baseline measurement and time from transplantation to randomization as covariates.

Triglycerides (mg/dL)   
Baseline139.9 ± 79.54 (138)154.0 ± 92.79 (134) 
Change at month 654.7 ± 110.61 (125)15.6 ± 104.85 (120)0.02
Change at month 1270.4 ± 208.85 (74)15.2 ± 137.67 (91)0.07
Low-density lipoprotein cholesterol (mg/dL)   
Baseline99.4 ± 38.91 (137)98.3 ± 40.64 (131) 
Change at month 622.9 ± 49.61 (114)−8.7 ± 36.33 (115)<0.001
Change at month 1220.5 ± 50.07 (70)2.4 ± 44.15 (89)0.001
High-density lipoprotein cholesterol (mg/dL)   
Baseline49.8 ± 17.67 (138)50.8 ± 16.33 (134) 
Change at month 6−2.5 ± 16.59 (123)−6.5 ± 17.64 (120)0.07
Change at month 12−2.5 ± 30.08 (73)−5.3 ± 15.34 (91)0.55
Systolic blood pressure (mm Hg)   
Baseline130.1 ± 16.42 (147)130.0 ± 18.85 (141) 
Change at month 60.8 ± 18.54 (139)−1.8 ± 20.63 (137)0.16
Change at month 122.4 ± 19.03 (86)3.7 ± 20.53 (98)0.73
Diastolic blood pressure (mm Hg)   
Baseline80.1 ± 10.40 (147)76.8 ± 10.99 (141) 
Change at month 6−1.3 ± 11.09 (139)−0.2 ± 12.22 (137)0.48
Change at month 12−0.4 ± 12.36 (86)1.8 ± 12.70 (98)0.88

In all, 20 of the 148 patients (13.5%) in the MMF/SRL group and 7 of the 145 patients (4.8%) in the MMF/CNI group were treated with at least 1 statin. Based on fasting lipid profiles, patients receiving MMF/SRL had significantly greater percentage changes from baseline in triglycerides and low-density lipoprotein cholesterol at 6 months and in low-density lipoprotein cholesterol at 12 months (Table 11). When adverse event reports, concomitant medication use, and laboratory assessments were all used to evaluate lipid elevations, the proportion of patients with hyperlipidemia at any time after randomization was significantly higher for the MMF/SRL group (104/148 or 70.3%) versus the MMF/CNI group (73/146 or 50.0%, P = 0.001).

Based on adverse event reports, concomitant medication use, and laboratory assessments, the number of patients with new or worsening diabetes at any time after randomization was significantly higher in the MMF/CNI group (39/146 or 26.7%) versus the MMF/SRL group (21/148 or 14.2%, P = 0.01). There was no difference in the change in blood pressure from baseline between the treatment groups at 6 or 12 months (Table 11).

The median change in the 24-hour urinary protein/creatinine ratio from baseline (0.1; 25th/75th percentiles, 0.1/0.2; n = 81) to 12 months in the MMF/SRL group was 0.0 (25th/75th percentiles, −0.0/0.1; n = 35), whereas in the MMF/CNI group, the median change from baseline (0.1; 25th/75th percentiles, 0.1/0.2; n = 84) to 12 months was −0.0 (25th/75th percentiles, −0.0/0.1; n = 50; P = 0.58 for MMF/SRL versus MMF/CNI).

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES
  8. Supporting Information

In this trial, liver transplant recipients who were converted from a maintenance immunosuppressive regimen of MMF and a CNI to MMF and SRL in the early postoperative period (4-12 weeks) had greater improvements in calculated GFR in comparison with those maintained on MMF/CNI. Although baseline GFR in the MMF/CNI group was slightly higher, the primary endpoint was analyzed using an analysis of covariance model in which baseline GFR was a factor. The change in calculated creatinine clearance from baseline also was significantly greater 12 months after transplantation. Serum creatinine in CNI-treated patients increased considerably from baseline to 12 months, likely reflecting cumulative damage. Exploratory analyses suggested that patients with HCC also may have had greater improvements in renal function when they were switched from a CNI. The renal function improvements with SRL were similar to those reported with the early withdrawal of CSA followed by everolimus monotherapy in de novo liver transplant patients.[21]

The noninferiority of the coprimary endpoint (the composite endpoint) was met. The rate of BPAR increased in the MMF/SRL group (12% versus 4% for the MMF/CNI group), with the majority of episodes occurring between randomization and 6 months after transplantation; conversely, the MMF/SRL group had a significantly lower rate of graft loss (including death). In the early posttransplant period, SRL trough levels were at the lower end of the target range, whereas CNI trough levels were at the higher end, possibly explaining the difference. Higher SRL doses may be warranted in the early posttransplant period during conversion. In a fairly large number of patients receiving SRL, CNI therapy was reinstated, suggesting that patients who did not tolerate a CNI-free regimen could be salvaged.

Both HCV infection[22, 23] and HCC[22, 24] have been observed to negatively affect liver allograft survival. Because HCV RNA levels are predictive of HCV recurrence and the change from baseline to 12 months in HCV RNA levels was higher in the MMF/SRL treatment group, it is possible that SRL might have negatively affected long-term graft survival in these patients.[22, 23] Similarly, BPAR, which occurred with greater frequency in the MMF/SRL group, is a strong predictor of graft loss in patients with HCV. This study, however, was not long enough or powered sufficiently to detect a negative impact of SRL use on patient and graft survival in patients with HCV infection.

Exploratory analyses of our data suggested that patients with an initial diagnosis of HCC who received MMF/SRL had significantly better patient/graft survival; similar findings were reported in a review of the Scientific Registry of Transplant Recipients that included 2491 liver transplant recipients with HCC.[25] SRL-based immunosuppression was associated with significantly improved survival at 5 years in recipients with HCC and with a trend toward less improvement in survival rates in those without HCC. SRL-based immunosuppression may have a beneficial impact on cancer patients by inhibiting HCC recurrence and metastasis or the occurrence of other cancers.[26-29] Mammalian target of rapamycin inhibitors have been recognized as potential anticancer agents affecting processes such as angiogenesis, cell proliferation, cell survival, and molecular oncogenic signaling.[30] In a prospective, case-control review,[31] patients receiving systemic chemotherapy and SRL immunosuppression had higher recurrence-free survival rates than patients receiving the same chemotherapy regimen and MMF/CNI (P = 0.001); survival rates also were higher at 1 (94% versus 79%), 3 (85% versus 66%), and 5 years (80% versus 59%, P = 0.001). The application of SRL for simultaneously inhibiting cancer spread and de novo cancer development and improving allograft and patient survival has been suggested and needs further study.

Statin use increased in patients receiving SRL. Although statins have been hypothesized to reduce the risk of cancers on the basis of several potential mechanisms,[32, 33] large meta-analyses of randomized controlled trials[34] or in combination with case-control and cohort studies[35] have shown no reduction in cancer incidence. It is unlikely that the use of statins in this study had an impact on survival for patients with HCC.

This study was designed to convert stable liver transplant recipients receiving MMF/CNI to MMF/SRL in the early (but not immediate) posttransplant period. SRL has been associated with an increased risk of mortality and graft loss[13] and infections/wound complications when given immediately after transplantation.[14] Only 55% of patients receiving MMF/SRL and 66% receiving MMF/CNI completed the treatment up to 1 year after transplantation, with the majority of withdrawals due to intolerable adverse events. As expected,[7, 11] leukopenia, mouth ulceration, and new-onset or worsening hyperlipidemia were more frequently reported in the MMF/SRL group. The potential for negative aspects of SRL use (eg, hyperlipidemia) to affect cardiovascular morbidity, mortality, and longer term results merits consideration. HAT occurred infrequently and at a similar rate in the treatment groups. Infections (particularly opportunistic infections) occurred at a significantly greater frequency in CNI-treated patients. This difference was predominantly attributable to a reduction in the risk of CMV syndrome/viremia in the MMF/SRL group. Like reductions in the risk of CMV infections with MMF/SRL were noted in a similarly designed Spare-the-Nephron trial involving renal transplant recipients[36] and in other studies involving renal[37-39] and solid organ transplantation.[40] Mammalian target of rapamycin is activated during CMV infections, with both rapamycin-sensitive and -insensitive pathways involved,[41] and it also has been implicated in regulating memory CD8+ T lymphocyte differentiation,[42] which may contribute to the reductions in CMV infections with SRL. However, the increased risk of rejection would seem to suggest that the combination of MMF and SRL is not as potent an immunosuppressive regimen as MMF/CNI, and this perhaps underlies the greater frequency of infection in the CNI-treated patients. Patients randomized to the MMF/CNI group had a significantly higher frequency of new or worsening diabetes.

The occurrence of proteinuria in patients receiving SRL after renal transplantation has been well documented.[43, 44] It is unclear whether these findings are renal transplant–specific; in this study, there were no significant differences between the treatment groups with respect to proteinuria.

Recently, Abdelmalek et al.[45] published the results of a similarly designed trial in which 602 eligible patients received liver allografts 6 to 144 months before enrollment and maintenance CNI therapy. Patients were converted to SRL- or CNI-based therapy. Between-group changes in the baseline-adjusted mean estimated GFR at month 12 were not significant. The primary endpoint, the noninferiority of the cumulative rate of graft loss or death at 12 months, was not met. Rates of death were not significantly different, and no true graft losses were observed. Adverse events were consistent with known safety profiles. This lack of demonstrable benefit at 1 year after the conversion of immunosuppressives contrasts with the results of the present study.

The primary limitation of this study was the short time frame for follow-up. Studies that consistently follow patients on SRL for a longer term are needed to confirm that the renal preservation benefits of the MMF/SRL regimen will be maintained without complications due to hyperlipidemia or other side effects. Another limitation was the lack of adequate biopsy information to detect the overall rate and severity of HCV recurrence. Additional studies would benefit our understanding of the impact of MMF/SRL-based therapy in patients with a history of HCV. Finally, the exploratory HCC subgroup analysis was not extensive enough to provide definitive information regarding SRL effects in this population and was only suggestive.

In conclusion, these results suggest that maintenance immunosuppression with MMF/SRL could be beneficial for liver transplant recipients who tolerate these agents and for whom renal function sparing is desirable. Therapy must be initiated in the early postoperative period (within 4 to 12 weeks after transplantation) to prevent the occurrence of adverse events reported with de novo use. An advantage of SRL use appears to be a lower incidence of associated CMV infection in comparison with the MMF/CNI combination; however, patients receiving MMF/SRL may also experience BPAR more frequently. The benefits and risks of any new regimen must be weighed, and an individual patient's tolerability, needs, and efficacy requisites need to be taken into consideration. Immunosuppression should be tailored to individual patients and should not be protocol-driven. Additional randomized, prospective trials in specific patient subgroups would be optimal for providing more definitive evidence with respect to longer term outcomes, survival, HCC benefits, and adverse events.

ACKNOWLEDGMENT

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES
  8. Supporting Information

The authors acknowledge the editorial assistance of Diann Glickman, Pharm.D. (Zola Associates).

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES
  8. Supporting Information
Abbreviations
BPAR

biopsy-proven acute rejection

CI

confidence interval

CMV

cytomegalovirus

CNI

calcineurin inhibitor

CSA

cyclosporine A

GFR

glomerular filtration rate

HAT

hepatic artery thrombosis

HCC

hepatocellular carcinoma

HCV

hepatitis C virus

ITT

intention-to-treat

MELD

Model for End-Stage Liver Disease

MMF

mycophenolate mofetil

MPA

mycophenolic acid

PP

per-protocol

SD

standard deviation

SEM

standard error of the mean

SRL

sirolimus

TAC

tacrolimus

Supporting Information

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENT
  7. REFERENCES
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
lt23658-sup-0001-suppinfo.rtf7KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.