Long-term outcome of mycophenolate mofetil rescue therapy for resistant acute allograft rejection in pediatric liver transplant recipients
Abstract
Mycophenolate mofetil (MMF) has been used to rescue liver allografts with steroid-resistant rejection (SRR). However, the long-term outcome of these patients is not known. This study evaluates the long-term outcome of MMF rescue therapy for SRR in pediatric liver allograft recipients. Twenty-six children (who received 28 liver transplants), including 16 girls, were given MMF for SRR. The median age at transplant was 1.7 (range 0.4-13.6) years. Primary immunosuppression was cyclosporine-based in 22 and tacrolimus-based in 6. All patients except one had been converted to tacrolimus prior to MMF, having already received a median of 2 (1-5) courses of high-dose intravenous methylprednisolone. The median time to MMF rescue therapy was 1.8 (0.4-35.8) months. Twenty-one of 28 episodes of SRR responded to MMF therapy. The median follow-up was 8.8 (7.7-11.5) years. In responders, there was 1 death from posttransplant lymphoproliferative disease, and no grafts were lost to chronic rejection. In the 7 nonresponders, 3 grafts were lost to chronic rejection with 2 patient deaths. Surviving children are clinically well with good liver function, and 17 remain on MMF. Three children have glomerular filtration < 80 mL/minute/1.73 m2. Side effects of MMF were seen in 12 patients; diarrhea (n = 5) and leukopenia (n = 5) being the most common. MMF was found to be effective in treating SRR in pediatric allograft recipients, with good long-term graft function and an acceptable side-effect profile. Liver Transpl 14:1303–1308, 2008. © 2008 AASLD.
Hepatic allograft rejection and its treatment remain important causes of patient morbidity after liver transplantation. With current immunosuppressive regimens, the incidence of acute allograft rejection and steroid-resistant rejection (SRR) is reported to be between 40% and 60% and between 6% and 30%, respectively, with primary tacrolimus-based regimens experiencing lower rejection rates than cyclosporine-based regimens.1-3 In order to avoid side effects associated with potent polyclonal (for example, anti-thymocyte globulin) or monoclonal antibodies (OKT3 and interleukin 2 receptor antibodies), agents such as mycophenolate mofetil (MMF) and lately mammalian target of rapamycin inhibitors have been used to rescue grafts with SRR.4-6 However, there is little data on the long-term outcome of MMF rescue therapy for resistant rejection in pediatric liver transplant recipients.
The aim of our study was to evaluate the use of MMF as rescue therapy for resistant rejection and to determine its use in preserving allograft function and its side-effect profile with a follow-up period of more than 5 years.
Abbreviations
AST, aspartate aminotransferase; 51C EDTA, 51chromium ethylene diamine tetraacetate acid; CNI, calcineurin inhibitor; EBV, Epstein-Barr virus; GFR, glomerular filtration rate; LASSR, late acute steroid-resistant rejection; MMF, mycophenolate mofetil; MPA, mycophenolic acid; PFIC, progressive familial intrahepatic cholestasis; PTLD, posttransplant lymphoproliferative disease; SRR, steroid-resistant rejection.
PATIENTS AND METHODS
This is a retrospective study on pediatric liver transplant recipients who received MMF as rescue therapy for SRR from September 1996 to December 1999 and were subsequently followed up prospectively until August 2007 for the outcomes of patient survival and graft function as well as complications associated with the use of MMF. A waiver for the need for ethics approval for this review was obtained from the Hospital Research Ethics Committee.
The standard immunosuppressive regimen used in our unit at that time was cyclosporine-based triple therapy (cyclosporine, azathioprine, and prednisolone). Recipients of living-related allografts or retransplants received tacrolimus-based dual therapy (tacrolimus and prednisolone) as primary immunosuppression. Treatment of acute rejection consisted of high-dose methylprednisolone (10 mg/kg/day) for 3 consecutive days. If patients failed to show a biochemical response following high-dose steroids, a second liver biopsy was performed to determine the cause of the persistent biochemical abnormality. If there was evidence of ongoing cellular rejection, patients were then converted to tacrolimus, if not already on it, and/or received a second course of high-dose methylprednisolone. In addition, children who experienced multiple (at least 2) episodes of cellular rejection also were converted to tacrolimus-based immunosuppression. Initial-target 12-hour tacrolimus trough levels were around 10 ng/mL. Tacrolimus whole blood levels were assayed with the Abbott IMx microparticle enzyme immunoassay.
The indication for the addition of MMF as rescue therapy for resistant rejection was biopsy-proven cellular rejection that had failed to improve following at least one 3-day course of intravenous pulse methylprednisolone (10 mg/kg/day).
MMF was added at a dose of 10 mg/kg/day in divided doses, and the dose was increased to a maximum of 40 mg/kg/day over 2 weeks. Mycophenolic acid (MPA) blood levels were not assayed, as this test was not available at the time of this study.
A positive response to the addition of MMF was defined as normalization of serum aspartate aminotransferase (AST; <50 U/L) or a decrease in AST to less than twice the upper limit of normal (AST 50-100 U/L).
RESULTS
Patient Demographics
Twenty-six children (who underwent a total of 28 liver transplants), including 16 girls, formed the study population. All except 1 received deceased donor liver transplants. The median age at transplant was 1.7 (range 0.4-13.6) years. The most common indication for liver transplant was biliary atresia (n = 10). Primary immunosuppression was cyclosporine-based triple therapy in 22 and tacrolimus-based dual therapy in 6. All children except 1 had been converted to tacrolimus therapy prior to the addition of MMF. In this cohort of children, the median time from liver transplant to the addition of tacrolimus was 0.5 (range 0.25-12) months in those who had received primary cyclosporine-based immunosuppression. The median time to MMF rescue therapy was 1.8 (range 0.4-35.8) months, with patients have receiving a median number of 2 (range 1-5) courses of high-dose intravenous methylprednisolone pulse therapy prior to the addition of MMF.
The histological features before MMF rescue therapy in relation to the immediate response to MMF are shown in Table 1.
| Histology | n | Initial Outcome |
|---|---|---|
| Acute cellular rejection | 14 | Twelve showed a positive response (1 child had near normal AST, but MMF was also stopped). |
| Acute cellular rejection and features of perivenular cell dropout | 11 | Six showed a positive response. |
| Acute rejection and ischemic changes | 1 | There was no response; the child subsequently underwent dilation of an anastomotic stricture. |
| Acute rejection and lobular hepatitis | 1 | There was a positive response; the child later was diagnosed to have de novo autoimmune hepatitis 3 years post-transplant. |
| Rejection and vanishing bile ducts | 1 | There was no response; the patient died from chronic rejection. |
- Abbreviations: AST, aspartate aminotransferase; MMF, mycophenolate mofetil.
Long-Term Outcome in Relation to the Initial Response to MMF Therapy
Twenty-one of 28 episodes of SRR (in 20 children/21 grafts) showed an initial positive response to MMF therapy. AST normalized completely in 15 patients. The median (range) time to normalization of AST was 0.5 (0.1-1.8) months. AST normalized in another 3 patients but was subsequently noted to be intermittently elevated. Another 3 patients experienced improvements in AST levels to between 50 and 100 U/L, without completely normalizing. However, no further escalation of immunosuppressive therapy was required.
Of these 20 children, 3 required a further single dose each of methylprednisolone therapy for biopsy-proven acute rejection (at 2, 3, and 5 months, respectively, after the addition of MMF). One of these children had stopped MMF temporarily because of neutropenia. She experienced an elevation of transaminases shortly after, and liver biopsy showed features of ongoing rejection. MMF was restarted subsequently at a lower dose without further problems.
Two grafts were lost as a result of biliary complications (n = 1) and hepatic artery thrombosis (n = 1). Two children died. One of the children who died received a liver allograft for acute liver failure. Following successful rescue therapy with MMF, she was well until 14 months post-transplant when she experienced Epstein-Barr virus (EBV) seroconversion and presented with signs of acute liver failure, requiring retransplantation. There were no features of posttransplant lymphoproliferative disease (PTLD) at retransplantation. She went on quickly to develop hyperacute rejection in her second graft and underwent a third liver transplant within 1 month of the second transplant. She also experienced SRR in her third allograft. Liver biopsy prior to the addition of MMF in this third allograft revealed features consistent with ductopenic rejection. Despite having being given anti-thymocyte globulin prior to her third transplant, as well as cyclophosphamide and interleukin 2 receptor antibodies (Simulect, Novartis) following the addition of MMF, she progressed to worsening jaundice and graft dysfunction, with death at 7 months after the third transplant. The second child who died succumbed to PTLD (B-cell lymphoma). This child experienced EBV seroconversion 37 months after the addition of MMF. MMF was stopped, and tacrolimus was continued. PTLD was diagnosed when the child presented 19 months later with generalized lymphadenopathy.
Of the surviving 18 children, 2 have continued follow-up in their country of origin since 1.1 and 5.5 years post-transplant. At their last review, they had normal AST and graft function and were still on MMF. The remaining 16 children have a median follow-up of 9.2 (range 7.7-11.5) years from liver transplant. Twelve remain on MMF (and tacrolimus), and 10 have normal serum transaminases. One child with associated biliary complications has an elevated AST (> 100 U/L). Another child has an elevated AST (between 50 and 100 U/L), and the latest liver biopsy showed nonspecific changes not suggestive of allograft rejection. Four children stopped MMF, 1 child because MMF was not available in her country of origin (she was converted back to azathioprine) and 3 because of side effects. One child had leukopenia and stopped MMF 7 months after starting it. His AST remains intermittently elevated (50-100 U/L). One child stopped MMF 3 years later because of protein losing enteropathy. His AST remains normal. The third child developed PTLD 5 years 2 months after MMF, and both tacrolimus and MMF were stopped. This child remains on low-dose prednisolone with normal AST.
In the remaining 7 episodes of SRR (in 7 children), there was no initial response to MMF therapy (Table 2). At follow-up, 2 of these children died (patients 1 and 2). Both deaths were related to graft loss from chronic rejection. One child was mentioned previously (patient 1). Patient 7 developed leukopenia 10 days following the addition of MMF, which was then stopped. His AST was on a downward trend during this time but did not completely normalize until after MMF had already been discontinued. It is therefore uncertain how much the addition of MMF to his immunosuppressive regimen contributed to his biochemical response. MMF was subsequently restarted 5.7 years later because of poor renal function (Table 3) without him experiencing leukopenia the second time.
| Patient | Liver Disease | Immunosuppression (After MMF Was Added) | Outcome |
|---|---|---|---|
| 1 | Third liver graft* | Cyclophosphamide and Simulect (anti-thymocyte globulin, tacrolimus, and prednisolone given as induction) | Chronic rejection and death |
| 2 | Biliary atresia | Methylprednisolone and tacrolimus† | Chronic rejection and retransplant 4 months later (chronic rejection in second allograft and subsequent death) |
| 3 | Biliary atresia | Methylprednisolone and tacrolimus† | Chronic rejection and retransplant 4 months later; small bowel PTLD discovered at retransplant; patient well following retransplant |
| 4 | Acute liver failure | Methylprednisolone and tacrolimus† | Associated hepatic artery thrombosis and anastomotic stricture; patient well; AST 50-70 U/L |
| 5 | Biliary atresia | Methylprednisolone, Simulect, and tacrolimus† | Patient well; AST 50-70 U/L |
| 6 | Progressive intrahepatic cholestasis (PFIC1) | Sirolimus (tacrolimus and MMF stopped) | Chronic diarrhea and pancreatic insufficiency; patient well; AST 50-100 U/L |
| 7 | Glycogen storage disease type 1 | Tacrolimus (MMF stopped) | Patient well; AST normal; MMF recommenced 5.7 years later because of low GFR |
- Abbreviations: AST, aspartate aminotransferase; GFR, glomerular filtration rate; MMF, mycophenolate mofetil; PFIC1, progressive familial intrahepatic cholestasis 1; PTLD, posttransplant lymphoproliferative disease.
- * The first allograft was lost to chronic rejection, and the second allograft was lost to hyperacute rejection,
- † A further 3- to 5-day course of methylprednisolone (10-15 mg/kg/day) was given.
| Patient | Duration of MMF Treatment Prior To Discontinuation | Rejection Response | Time from Discontinuation to Restarting MMF (Years) | Time from Transplant (Years) | GFR (mL/minute/1.73 m2) |
|---|---|---|---|---|---|
| 1 | 2 months | Responder | 5.5 | 8.5 | 56* |
| 2 | 1.5 years | Responder | 5 | 11.4 | 55 → 110 |
| 3 | 3 years | Responder | None | 11.5 | 64 |
| 4 | 5.3 years | Responder | 1.5 | 9.5 | 70 → 85 |
| 5 | 10 days | Nonresponder | 5.7 | 7.9 | 33 → 96 |
| 6 | MMF not discontinued | Responder | — | 10.1 | 55 → 77 |
- Abbreviations: GFR, glomerular filtration rate; MMF, mycophenolate mofetil.
- * There was an improvement in serum cystatin C; GFR was not yet repeated.
Renal Function
Renal function was assessed with serum cystatin C levels or with 51chromium ethylene diamine tetraacetate acid (51Cr EDTA) clearance. Sixteen children had at least 1 51Cr EDTA glomerular filtration rate (GFR) measurement. Six children had GFR < 80 mL/minute/1.73 m2, 5 of whom had previously stopped MMF because of side effects. MMF was reintroduced in combination with a lower tacrolimus dose in 4 children, with subsequent improvements in GFR in 3 (Table 3). The child in whom MMF had not been stopped received a reduced dose of tacrolimus with a concomitant increase in MMF.
The median time of the latest 51Cr EDTA assessment from transplant in the 16 children was 8.1 (range 5.2-11.5) years. Overall, the median GFR in the group was 92 (range 56-150) mL/minute/1.73 m2, with 3 patients having a GFR < 80 mL/minute/1.73 m2 (56, 64, and 77, respectively). Of the remaining 5 patients, 4 had consistently normal serum cystatin C (0.74-0.94 mg/L, normal range 0.55-1.15 mg/L), and 1 child had a borderline value of 1.15 mg/L.
Late Acute Rejection and Other Associated Complications
Of these 26 children, the majority (n = 20) experienced SRR within 6 months of liver transplant. Clinical details of the 6 children who had SRR beyond 6 months after liver transplantation are shown in Table 4.
| Patient | Time (Months) Post-Transplant at Which LASSR Occurred | Associated Conditions Prior to LASRR | Outcome |
|---|---|---|---|
| 1 | 7.9 | PTLD at 3 months; patient taken off tacrolimus and converted to cyclosporine; 3 episodes of steroid-responsive rejection | Patient well; normal AST; ischemic changes on biopsy due to hepatic artery thrombosis with good collaterals; de novo autoimmune hepatitis |
| 2 | 14.5 | 3 episodes of steroid-responsive rejection; biliary stricture post-dilatation | Patient well; borderline elevated AST; most recent biopsy showed nodular regenerative hyperplasia |
| 3 | 14.5 | PFIC with chronic diarrhea; patient unable to achieve adequate tacrolimus levels | Patient well; persistently elevated AST; patient taken off MMF and converted to sirolimus |
| 4 | 17.3 | 4 episodes of steroid-responsive rejection | Patient well; normal AST; no further episodes of rejection over next 7.7 years |
| 5 | 27.7 | Biliary anastomotic stricture, reconstruction done 1 month post-transplant; patient underimmunosuppressed because of EBV seroconversion 21 months post-transplant; 2 episodes of steroid-responsive rejection | Patient well; elevated AST; de novo autoimmune hepatitis |
| 6 | 35.8 | 3 episodes of steroid-responsive rejection; biliary complications and hepatic artery stenosis | Elevated AST; protein losing enteropathy; MMF stopped |
- Abbreviations: AST, aspartate aminotransferase; EBV, Epstein-Barr virus; LASRR, late acute steroid-resistant rejection; MMF, mycophenolate mofetil; PFIC, progressive familial intrahepatic cholestasis; PTLD, posttransplant lymphoproliferative disease.
Furthermore, 8 of these 26 children had associated biliary or hepatic artery complications. These included biliary anastomotic strictures (n = 5), hepatic artery stenosis (n = 1), or a combination of both biliary and hepatic artery complications (n = 2). Although a number of them received intervention prior to the episodes of SRR, their presence could be the cause of persistent biochemical abnormalities.
Of this population of children, 3 have developed de novo autoimmune hepatitis, 2 of whom belong to the late acute rejection group.
EBV Infection
Eight (30.8%) children became EBV-seropositive following transplantation, 3 before the addition of MMF and 5 after. Of these, 4 developed PTLD. One child developed PTLD before the addition of MMF; the data is shown in Table 4 (patient 1). The other 3 developed PTLD after the addition of MMF. One child succumbed to B-cell lymphoma, and one is well following treatment with rituximab and chemotherapy and is on low-dose prednisolone as the only immunosuppressive agent. The third child was diagnosed to have small bowel PTLD only at retransplant (described previously) and is currently well. One child developed acute liver failure following EBV seroconversion, and this was attributed to acute EBV hepatitis. She lost her graft as previously mentioned.
MMF-Related Side Effects
Twelve children experienced side effects that could be related to MMF. These included diarrhea (n = 5), vomiting (n = 1), abdominal pain (n = 1), protein losing enteropathy (elevated stool alpha-1-antitrypsin levels; n = 1), and leukopenia (n = 5). MMF was stopped in 6 children: the child with leukopenia, the child with protein losing enteropathy, the child with progressive intrahepatic familial cholestasis and chronic diarrhea, and the 3 children who developed PTLD. MMF was stopped and restarted successfully in the remaining children.
DISCUSSION
MMF is effective as additional immunosuppression in pediatric liver allograft recipients with SRR, with an acceptable side-effect profile and good long-term graft outcome. Overall response was good, with 75% of the rejection episodes in the described cohort experiencing a positive response. Of the patients with histological features of perivenular cell dropout, 54.5% responded to MMF, whereas in those who did not have this histological feature, a positive rejection response was seen in 85.7%. This difference, however, was not statistically significant, possibly because of the small sample size. It has previously been noted that perivenular hepatocyte dropout is one of the characteristics of irreversible graft rejection, much like the loss of bile ducts in vanishing bile duct syndrome.7 Thus, this 54.5% response rate could mean either that MMF was sufficiently potent to reverse allograft rejection at a fairly advanced stage or that its presence merely served as a negative prognostic factor for response. At the same time, patients who did not show an initial response to MMF had a high chance of progressing to chronic rejection (3/7, 42.9%) resulting in the need for retransplantation or patient death. Without MMF, this group of 26 children would certainly have been candidates for the potent polyclonal/monoclonal antibodies or higher maintenance doses of tacrolimus, with their associated adverse effects. The latter would have important implications in children.
The children in this study are likely to have had a propensity for allograft rejection in view of the frequency of rejection episodes prior to the occurrence of SRR, as well as the fact that 2 of these children developed recurrence of chronic rejection in their retransplants. The high incidence of EBV infection would also indirectly reflect that the net immunosuppression that these children received was likely to have been on the higher side and not a consequence of MMF per se. EBV seroconversion and the subsequent development of PTLD are of particular importance to pediatric recipients as they are more likely to be EBV-naïve and receive donor grafts from adults who are likely to be EBV-positive. Studies in pediatric liver transplant recipients show that the incidence of PTLD ranges from 7% to 13.7%, being higher in the tacrolimus-converted group than the primary tacrolimus group.8, 9 In our study, 4 of 26 children developed PTLD (occurring between 3 and 62 months post-transplant), a figure that is probably not unexpected because these 26 children with resistant rejection probably represent the highest risk group for PTLD.
Being able to avoid higher doses of tacrolimus is an important factor in the long term in preventing the renal dysfunction that has been associated with its use. A decreased GFR has been described to occur in 20% to 70% of children under tacrolimus-based immunosuppression.10, 11 In our study, 17 of 21 children (81%) had normal renal function with a median follow-up of 8.8 years. It is of interest to note that 5 of the 6 children with impaired renal function had discontinued MMF earlier and had been maintained on tacrolimus. MMF was reintroduced successfully in 4 children in an attempt to prevent calcineurin inhibitor (CNI)–related nephrotoxicity, with improvements in GFR in 3. This shows that it is possible to reintroduce the drug in patients who were previously intolerant of it. Importantly, none of these children experienced acute allograft rejection as a result of tacrolimus dose reduction during this time. The efficacy of MMF in preventing CNI-related nephrotoxicity, as well as the potential risk of allograft rejection, has been described.12, 13 In the last year, our unit has been screening transplant recipients using serum cystatin C instead of 51Cr EDTA. 51Cr EDTA assessments are used for patients with borderline results or previously known low GFR. It has been shown that serum cystatin C is a reliable marker for renal dysfunction. A level > 1.06 mg/L was 91% sensitive and 81% specific in predicting a GFR < 80 mL/minute/1.73 m2.14
At the time of this study, plasma MPA levels were not monitored as the assay was not available. Subsequently, a study done by our unit showed that predose MPA levels correlated both with therapeutic efficacy and with risk of side effects.15 The relative risk of rejection was 2.5 with an MPA level < 1 mg/L, whereas levels above 3 to 4 mg/L were associated with a greater than 3-fold increase in infection or leukopenia. In addition, it was noted that MPA levels were not correlated with the dose of MMF but were influenced more by the immunosuppressive comedication, age of the patient, renal function, and serum albumin levels. The ability to monitor MPA levels would have allowed better tailoring of the MMF dose, thereby maximizing efficacy and minimizing side effects. Currently, monitoring MPA levels is part of routine clinical care in our unit.
In conclusion, MMF was generally well tolerated with side effects similar to those that have been reported by other authors.4, 5 Our study has shown that MMF is both effective and safe as add-on therapy to rescue liver grafts with SRR in children under CNI-based immunosuppressive regimens, with good long-term patient and graft outcome.
