- Top of page
A randomized controlled trial of CMV-IVIG (cytomegalovirus-intravenous immunoglobulin) for prevention of Epstein Barr virus (EBV) posttransplant lymphoproliferative disease (PTLD) in pediatric liver transplantation (PLTx) recipients was begun in Pittsburgh and subsequently expanded to four additional sites. Protocol EB viral loads were obtained in a blinded fashion; additional loads could be obtained for clinical indications. Patients were followed for 2 years post-LTx. Eighty-two evaluable patients (39 CMV-IVIG, 43 placebo) developed 18 episodes of EBV disease (7 CMV-IVIG, 11 placebo) including nine cases of PTLD (three CMV-IVIG, six placebo). No significant differences were seen in the adjusted 2-year EBV disease-free rate (CMV-IVIG 79%, placebo 71%) and PTLD-free rate (CMV-IVIG 91%, placebo 84%) between treatment and placebo groups at 2 years (p > 0.20). The absence of significant effect of CMV-IVIG may be explained by a lack of efficacy of the drug or limitations of sample size.
- Top of page
Epstein Barr virus (EBV) is second to cytomegalovirus (CMV) as the most frequent and important viral pathogen affecting pediatric recipients of liver transplantation (PLTx). The spectrum of EBV infection after transplantation includes asymptomatic seroconversion, a fever syndrome and posttransplant lymphoproliferative disease (PTLD) which encompasses a range of disorders from infectious mononucleosis-like lesions to lesions indistinguishable from lymphoma (1). Clinically important EBV infections and PTLD occur five to seven times more frequently in patients who are seronegative prior to transplantation compared with patients who are seropositive at the time of LTx (2,3). As many as 60% of EBV seronegative children undergoing organ transplantation will seroconvert during the first 3 months following solid organ transplantation (4). Although many children who seroconvert will not have clinically recognizable disease, cumulative rates for PTLD as high as 20% have been reported 7 years after pediatric liver transplantation (PLTx) under cyclosporine-based immunosuppression (5). Reported rates of illness attributable to EBV in older series may have underestimated the true rate because of the lack of sensitive diagnostic tools (e.g. polymerase chain reaction [PCR] detection).
Although the outcomes of EBV-related disease, including PTLD, appear to be improving, development of EBV-related complications is still associated with an unacceptably high rate of morbidity and mortality (6). Accordingly, prevention of EBV infection is desirable. However, few data are available regarding strategies to prevent EBV infection among transplant recipients (7–9). Since infection with EBV before transplantation is associated with a five- to sevenfold reduction of symptomatic EBV infection after transplantation (including PTLD) (3), preventive strategies for EBV might include active immunization. However, an EBV vaccine is not currently available and is unlikely to be in the near future (10). A possible alternative is passive immunization with intravenous immunoglobulin (IVIG). CMV-IVIG (Cytogam®) is a pooled IVIG product with a high titer of activity against CMV. Initial analysis of the anti-EBV antibody titers in CMV-IVIG showed 10- to 100-fold greater titers compared with standard IVIG (unpublished data—MedImmune, Inc. Gaithersburg, MD, USA). Accordingly, this product appeared to be a good candidate to determine whether passive prophylaxis could result in a decreased frequency and severity of EBV infection after PLTx. The purpose of this study was to compare the incidence of EBV disease and PTLD during the first 24 months post-LTx, in patients randomized to receive either CMV-IVIG (Cytogam) or a saline placebo.
- Top of page
This study began in 1995 as an investigator-initiated single-center, prospective, double-blind, randomized, placebo-controlled trial at The Children's Hospital of Pittsburgh. In an effort to enhance accrual, the study was expanded to a multicenter trial in 1998 by the inclusion of the following centers: Children's Memorial Hospital (Chicago), Wyler's Children's Hospital (Chicago), Mt. Sinai Medical Center (New York City) and University of North Carolina Medical Center (Chapel Hill). EBV seronegative children <18 years of age undergoing their first LTx were eligible. EBV seropositive infants <12 months of age were also eligible since the presence of EBV antibodies in these children likely represented the presence of maternal antibody. Patients were stratified by age (<12months, ≥12 months). Patients were ineligible if they were seropositive for EBV and older than 12 months of age, undergoing re-transplantation, or were undergoing or had previously undergone any other type of organ transplant. Enrollment was closed in December 2001. The study protocol was approved by the Institutional Review Board of each participating center.
Informed consent was obtained from parents of eligible children before or at the time of LTx. After consent was obtained, patients were randomized to receive either CMV-IVIG or placebo. Randomization was carried out in blocks of four for each stratum (<12 months, ≥12 months of age) at each site using randomization packets that were prepared by the coordinating center. CMV-IVIG was dosed according to the approved regimen for the prevention of CMV infection in LTx recipients (11). The initial dose of 150 mg/kg of CMV-IVIG was provided within 120 h of transplant. Repeated doses of CMV-IVIG of 150 mg/kg were given at 2, 4, 6 and 8 weeks after transplantation; 100 mg/kg was infused at 12 and 16 weeks posttransplantation. Patients in the control group received an intravenous dose of saline as a placebo at the same time-points until discharge. For logistical and ethical reasons, once discharged, patients in the placebo group did not receive saline infusions. However, treating physicians continued to be blinded to randomization group. Intravenous ganciclovir (10 mg/kg per day) was provided for CMV prophylaxis for the first 14 days after transplant. Children who were CMV seronegative prior to transplant who received organs from a seronegative donor were not treated with ganciclovir.
Each participating center used a tacrolimus-based immunosuppressive regimen in combination with corticosteroids. None of the participating centers used anti-lymphocyte antibody regimens as induction therapy during the study period. Modifications in immunosuppressive regimens, including treatment of rejection, were made on an individual basis by the child's managing transplant team without knowledge of group assignment.
EBV serology was requested for all donors and recipients at the time of transplantation at each participating center. An aliquot of blood was obtained for the measurement of an EB viral load before the first infusion of study drug. Monthly samples were then obtained for the first 6 months posttransplant as well as 12 and 24 months posttransplant. EB viral load measurement was also performed as part of the clinical evaluation for symptomatic illness. EB viral loads were performed at a central lab by one of the investigators (DR) using previously published methods (12), and reported as genome copies per 105 peripheral blood lymphocytes (PBL). The performance of surveillance EB viral load measurement outside of the study, though available, was discouraged.
Patients were followed until they reached a study endpoint (symptomatic EBV disease or PTLD) or 24 months after transplant, whichever came first. Patients requiring re-transplantation or who were withdrawn by a parent or investigator were removed from the study with analysis censored from that time forward. Patients removed from the study for any reason within 1 month of enrollment were considered non-evaluable and were completely censored from the analysis. Data points, including transfusion history, immunosuppression, episodes of rejection, need for OKT3, as well as diagnosis and outcomes of infectious episodes were prospectively collected using standardized case report forms.
Evaluation of infectious episodes
Consideration of EBV infection is a routine part of the evaluation of fever in PLTx recipients. Accordingly, in addition to standard viral and bacterial cultures, diagnostic tests aimed at the identification of EBV disease and PTLD were routinely obtained at each participating center as part of standard clinical care. Patients identified as having an illness consistent with EBV infection underwent further evaluation to identify occult sites of disease, including computed tomography (CT) scan of the chest and abdomen, endoscopic evaluation of the gastrointestinal tract and lymph node excision for histologic evaluation, as clinically indicated. Evaluation for the presence of EBV was performed on all tissues biopsy specimens obtained from patients with suspected EBV disease at each center using Epstein Barr-encoded RNA (EBER) staining (13). Results of these evaluations were recorded on the case report forms.
Definition of EBV infection and disease
Symptomatic EBV infection was defined as either seroconversion, development of a positive viral load ≥ 200 genome copies per 105 PBL (12), or histologic evidence of EBV infection (by EBER) in the presence of typical symptoms or laboratory findings (e.g. fever, leukopenia, atypical lymphocytosis, exudative tonsillitis and/or adenopathy). EBV disease was further characterized as either ‘proven’, ‘probable’ or ‘possible’. ‘Proven’ EBV disease required histologic confirmation using the EBER stain. ‘Probable’ symptomatic EBV infection was diagnosed if there was evidence of EBV infection in the presence of typical symptoms, in the absence of alternate explanation. ‘Possible’ symptomatic EBV infection was made if there was evidence of EBV, presence of typical symptoms and an inability to exclude the diagnosis despite the presence of alternate explanation. Episodes of ‘probable’ and ‘possible’ EBV disease were further classified as: viral syndrome, mononucleosis or adenitis/adenopathy. The diagnosis of PTLD required proven histologic evidence using standardized definitions (1). Histology was reviewed at the Children's Hospital of Pittsburgh.
EBV disease case-review process
All probable and possible episodes of EBV disease were reviewed by a case-review committee comprising of the senior investigator from each participating center. An abstract describing each episode of probable or possible EBV disease was prepared and distributed to each committee member who initially made an independent diagnosis using the pre-set study definitions. Cases with discrepant diagnoses were resolved by conference. All determinations were made without knowledge of treatment group assignment.
Definitions of CMV disease
CMV disease was defined as a CMV-positive culture from any site after transplantation in association with either CMV syndrome (characterized by fever, leukopenia and/or atypical lymphocytosis) or evidence of invasive CMV disease (defined as positive results of viral culture or histological examination of a tissue biopsy specimen which demonstrated the presence of viral inclusions) (14). CMV pneumonia was diagnosed if there was a CMV-positive culture or cytology of a bronchoalveolar lavage fluid specimen associated with lower respiratory tract disease in the absence of identification of any other viral, fungal or bacterial pathogen.
The log-rank test was the primary statistical test of equivalence of the treatment and placebo arms. This test compared the time free from EBV disease (symptomatic EBV infection or PTLD) over the study period. Log-rank was used because of the expectation that some patients would die or not complete the 2-year follow-up period without developing EBV disease. This method allowed for adjustment for this occurrence. Subjects who did not meet the event were censored at the time of death (for subjects who died) or at the time of last contact (for subjects who discontinued prematurely). Subjects who experienced multiple events were counted as having met the event at the time of the first occurrence. The primary endpoints of this study were 2-year EBV disease and PTLD-free rates adjusted for censoring. The secondary endpoint of this study was the adjusted 2-year CMV disease-free rate.
Prior to initiation of the study, a power analysis was performed based upon the assumption that a 2-year EBV disease incidence rate of 25% and 10% would be observed in the placebo and CMV-IVIG cohorts, respectively. Under these conditions and using a one-sided test with a type I error rate of 5%, it was predicted that accrual of 40 evaluable patients into each arm would allow an analysis that would have about 72% power to detect the specified EBV disease incidence rates with an alpha level of 0.05. To allow for loss of evaluable patients (due to early death or parental or investigator withdrawal from study), a target enrollment population of 120 patients was set at the onset of the study.
An interim analysis was performed in January of 2001. The principle investigators for the overall study and for each site were blinded to the full details of this analysis. The provisions for stopping the study were tied to safety issues. No safety issues requiring termination of the study were identified. The interim analysis identified that statistical significance had not been achieved and continued study accrual was recommended. However, by December of 2001, enrollment was closed due to lower than expected enrollment numbers in the months following the interim analysis.
- Top of page
Ninety-seven children were enrolled between 1995 and 2001; 82 of the 97 were evaluable. Non-evaluable patients included those who were EBV seropositive at the time of transplant (n = 2) or who were withdrawn from the study before completion of first posttransplant month for any additional reason (e.g. withdrawal of parental consent, re-transplantation, death) (n = 13). Of the remaining 82 patients, 39 were randomized to CMV-IVIG and 43 to placebo. Ninety percent of the children received their first infusion of study drug within the first 3 days following transplantation. A summary of the demographic data from the patients in the two treatment groups is shown in Table 1. Groups were equivalent for age, gender and ethnicity and underlying disease leading to transplantation at the time of randomization. They were also equivalent for primary immunosuppression as well as exposure to anti-T-cell antibody therapy as treatment for rejection (data not shown). A trend toward increased frequency of living related donors was noted in the placebo group (33%) compared with patients receiving CMV-IVIG (15%) (p-value = 0.12). A similar trend toward an increased rate of EBV donor seropositivity was also noted in the placebo group (33 vs. 15%; p-value = 0.12). However, EBV donor status was unavailable for 50 of 82 donors. Thirty-two of the CMV-IVIG patients (82%) and 36 of the placebo patients (84%) completed the study (i.e. completed 2-year follow-up or reached a primary study endpoint). The reasons that the remaining 14 children did not complete the study are also shown in Table 1. Results of participants who did not complete the study are included until the time that these patients were censored from the analyses.
Table 1. Comparison of demographic data and patient survival for patients treated with CMV-IVIG and placebo
|No. of subjects enrolled||39||43|
|No. of subjects who completed study||32||36|
|No. of subjects who did not complete study||7||7|
| Re-transplant||2 (5.1%)||1 (2.3%)|
| Subject lost to follow-up||1 (2.6%)||3 (7.0%)|
| Subject died||1 (2.6%)||1 (2.3%)|
| Subject/guardian withdrew consent||1 (2.6%)||2 (4.7%)|
| Investigator withdrew consent||2 (5.1%)||0|
| < = 1||21 (53.8%)||24 (55.8%)|
| Male||21 (53.8%)||24 (55.8%)|
| White/non-Hispanic||28 (71.8%)||32 (74.4%)|
| Black||3 (7.7%)||4 (9.3%)|
| Hispanic||5 (12.8%)||4 (9.3%)|
| Asian||1 (2.6%)||2 (4.7%)|
| Other||2 (5.1%)||1 (2.3%)|
|EBV donor serostatus|
| EBV donor seropositive||6 (15.4%)||14 (32.6%)**|
| EBV donor seronegative||5 (12.8%)||7 (16.3%)|
| EBV donor serostatus unknown||28 (71.8%)||22 (51.2%)***|
|Patient and graft survival|
| 2-year patient survival||97%||97%|
| 2-year rejection-free survival||43%||32%|
Eighteen cases of EBV disease, including 9 cases of PTLD, were diagnosed. All 9 cases of PTLD involved EBV-infected B cells and presented a polymorphic histology. The remaining 9 cases of EBV disease included 3 episodes of enteritis (1 proven and 2 possible), 5 episodes of viral syndrome (2 probable and 3 possible) and 1 episode of hepatitis (proven). EBV disease developed in seven of 39 recipients of CMV-IVIG compared with 11 of 43 children receiving placebo. EBV-associated PTLD was diagnosed in three of 39 recipients of CMV-IVIG compared with six of 43 placebo patients. The adjusted 2-year EBV disease-free and PTLD-free rates were 79% and 91% compared to 71% and 84% for CMV-IVIG and placebo groups, respectively, (p > 0.20). Analysis by age at the time of transplant also failed to identify differences in the time free from EBV disease or PTLD for those who received CMV-IVIG or placebo (Figures 1 and 2). As noted, donor EBV serology was unknown in 50 (61%) transplants. For the remaining 32 subjects, donor EBV serology did not significantly affect EBV disease or PTLD-free rates (Table 2).
Figure 1. Kaplan-Meier curve for time to EBV disease (evaluable subset). (A) Results for all evaluable patients. Corrected 2-year EBV disease-free rates were 82.1% for CMV-IVIG cohort compared to 74.4% for children receiving placebo (p > 0.20). (B) Results for children < 1 year of age. Corrected 2-year EBV disease-free rates were 75% and 62% for children receiving CMV-IVIG and placebo, respectively (p > 0.20). (C) Results for children ≥ 1 year of age. Corrected 2-year EBV disease-free rates were 83% and 82% for children receiving CMV-IVIG and placebo, respectively (p > 0.20).
Download figure to PowerPoint
Figure 2. Kaplan-Meier curve for time to PTLD (evaluable subset). (A) Results for all evaluable patients. Corrected 2-year PTLD-free rates were 91.2% for CMV-IVIG treated patients compared to 84.0% for children receiving placebo (p > 0.20). (B) Results for children < 1 year of age. Corrected 2-year PTLD disease-free rates were 88% and 81% for children receiving CMV-IVIG and placebo, respectively (p > 0.20). (C) Results for children ≥ 1 year of age. Corrected 2-year PTLD disease-free rates were 94% and 88% for children receiving CMV-IVIG and placebo, respectively (p > 0.20).
Download figure to PowerPoint
Table 2. Effect of pretransplant EBV serostatus* on the incidence of EBV disease and PTLD at 2 years of follow-up
|Treatment group||N||# with EBV disease||2-year EBV disease-free rate (%)#||p-value||# with PTLD||2-year PTLD-free rate (%)**||p-value|
|Donor EBV seropositive|
| CMV-IVIG||6||1||75%|| ||0||100%|| |
| Placebo||14||6||46%)||p = 0.18||2||82%||p = 0.33|
|Donor EBV seronegative|
| CMV-IVIG||5||1||80%|| ||0||100%|| |
| Placebo||7||2||71%||p = 0.83||1||86%||p = 0.40|
|Donor EBV serostatus unknown|
| CMV-IVIG||28||5||79%|| ||3||86%|| |
| Placebo||22||3||84%||p = 0.64||3||84%||p = 0.79|
Study groups were also compared for the incidence of CMV disease. A total of 13 cases of CMV disease were diagnosed among the 82 evaluable children. CMV disease developed in five of 39 recipients of CMV-IVIG compared with eight of 43 children receiving placebo. The adjusted 2-year CMV disease-free rates were 86% compared to 81% for CMV-IVIG and placebo groups, respectively, (p > 0.20) (Figure 3). Adjusted 2-year CMV disease-free rates were not improved by the use of CMV-IVIG in high risk (donor seropositive/recipient seronegative) recipients (50% vs. 70%, for CMV-IVIG and placebo, respectively (p > 0.20). Two-year CMV disease-free rate in children who were CMV seropositive prior to transplantation were 90% versus 73% for patients treated with CMV-IVIG compared to placebo (p > 0.20) Similarly, the use of CMV-IVIG did not affect 2-year disease-free rates for children < 1 year of age (83% vs. 74%, for CMV-IVIG and placebo, respectively (p > 0.20)) or for those ≥ 1 year (89% for both those receiving CMV-IVIG and placebo).
Figure 3. Kaplan-Meier curve for time to CMV disease (evaluable subset). Corrected 2-year CMV disease-free rates were 86% for CMV-IVIG treated patients compared to 81% for children receiving placebo (p > 0.20).
Download figure to PowerPoint
The potential impact of CMV-IVIG on the development of either PTLD or CMV disease was evaluated. The 2-year adjusted PTLD and CMV disease-free rate was 78% versus 68% for recipients of CMV-IVIG and placebo, respectively (p > 0.20). Similar rates were observed in children < 1 year of age (71% vs. 60%) and ≥ 1 year of age (84% vs. 77%) (p > 0.20). The study groups were also compared for overall survival and incidence of rejection. A single death occurred in each group (97% survival at 2 years); neither death was due to EBV or CMV disease. The rejection-free rate at 2 years was 43% among recipients of CMV-IVIG and 32% among recipients of placebo (p > 0.20).
- Top of page
Symptomatic EBV disease, including PTLD, is recognized as a major cause of morbidity and mortality following organ transplantation, especially in patients experiencing primary EBV infection following transplantation. Therefore, interest has focused on prevention of EBV infection and disease in pediatric transplant recipients. The absence of an EBV vaccine prompted the consideration of passive immunization to prevent the development of EBV disease, including PTLD. The potential efficacy of this approach was supported by data from a severe combined immunodeficiency (SCID) mouse model (15,16), as well as anecdotal human experience.
Results of the current study did not demonstrate a statistically significant decrease in EBV disease or PTLD using CMV-IVIG. The inability to demonstrate a benefit could have been due to a lack of efficacy of this therapy, changes in clinical management as the study progressed and/or inadequate statistical power of the current study. For our original sample size calculation, the EBV disease-free rate in the placebo group was based on historical experience, while a prediction was generated for the effect of CMV-IVIG. Clearly, this latter prediction was in error and the study was underpowered. Based upon rates of EBV disease and PTLD observed in the current study, we would have had to accrue 195 patients in each group to achieve 70% power for EBV disease and 579 per group for PTLD. Accrual of such a large number of PLTx recipients was not then, and is not now, feasible.
Although statistically significant differences were not observed, rates of EBV disease and PTLD were somewhat lower in recipients of CMV-IVIG than in those who received placebo. This was particularly true for children less than 1 year of age, where 25% of children receiving CMV-IVIG developed EBV disease compared with 38% receiving placebo. In contrast, the rate of development of EBV disease at 2-year follow-up was essentially identical for children ≥ 1 year of age. While differences in the rates of development of PTLD in the children < 1 year of age were less dramatic, the advantage again favored the recipients of CMV-IVIG (12% vs. 19%). Only 3 children older than 1 year developed PTLD, making an analysis of difference between the two treatment groups in this older age cohort impossible. Confirmation of these potential differences would require performance of an adequately powered study.
In 1998, McDiarmid and colleagues published experience using EB viral load monitoring to inform preemptive treatment of PLTx recipients with rising EB viral loads (17). This study, along with several subsequent studies (18–20) demonstrated apparent decreases in the rate of EBV disease and PTLD for patients whose levels of immunosuppression were decreased in response to rising EB viral loads. To assess the potential confounding impact of use of this strategy on the outcome of the current study, the medical records and all EBV viral loads obtained from patients participating in the study at the Children's Hospital of Pittsburgh were reviewed. As shown in Table 3, an increase in the number of EBV viral loads drawn outside of the study protocol, preemptive reductions in immunosuppression made in response to these levels and a decline in the number of cases of PTLD were seen in the cohort of subjects enrolled after November 1998 compared to those enrolled before this time. Although similar comparative data are not available for research subjects outside of Pittsburgh, these findings suggest that increasing use of this strategy on patients on the research protocol could have accounted for the decreased rates of EBV disease and PTLD observed in the latter part of this study, thereby both potentially confounding our results and impacting our ability to achieve statistically significant results with CMV-IVIG.
Table 3. Potential effect of EBV viral load assessment outside of study protocol on the outcome of the CMV-IVIG versus placebo trial
|Pittsburgh patients||Pre-November 1998||Post-November 1998||p-value|
|Patients enrolled||29||22|| |
|# Patients with any viral load drawn||11||15||0.061|
|Total # viral loads drawn||23||101|| |
|Patients receiving ‘preemptive therapy’||2||7||0.032|
|Cases of PTLD||5||1||0.222|
CMV-IVIG is approved for the prevention of CMV disease in PLTx recipients. As part of the current study, the impact of CMV-IVIG on the incidence of CMV disease was also investigated. With the exception of CMV seronegative recipients of organs from CMV seronegative donors, all patients participating in this trial received 2 weeks of intravenous ganciclovir as CMV prophylaxis. In this setting, CMV-IVIG did not have a significant impact on the development of CMV disease. This was even true in the high-risk CMV donor seropositive/recipient seronegative patient. A trend toward a higher 2-year CMV disease-free rate was observed in children who were CMV seropositive prior to transplantation treated with CMV-IVIG. However, a significant beneficial effect for the combined endpoint of development of either PTLD or CMV disease was not observed, and there were no differences in the incidence of rejection or the mortality rate in either treatment group.
Several specific factors potentially weaken the current study. Beyond the previously mentioned issues with study size, the fact that donor EBV serostatus was available in only 32 of 82 evaluable patients adds additional potential problems with the interpretation of this study. It is possible that despite the use of randomization, a higher rate of EBV mismatched patients might have been randomized to one of the two treatment cohorts, thereby adversely affecting observed results. Given the absence of EBV donor status on more than 60% of the patients, one can only hope that the randomization process corrected for this variable that cannot be otherwise accounted for. A second potential weakness is the fact that more than half of the cases of EBV disease observed in this study were graded as only ‘possible’. Despite development of specific definitions and processes for evaluations of EBV disease, such cases cannot be excluded from results but leave a troubling concern over their potential impact in the analysis of the outcome of this study. Avoidance of these ‘possible’ cases of EBV disease as endpoints for analyses would require an even larger study to identify cases and controls with more definite endpoints, making the conduct of such a study even less feasible.
In summary, the current study did not demonstrate efficacy of CMV-IVIG in the prevention of EBV disease or PTLD in PLTx recipients. The observed trends would appear to support performance of a larger study. Unfortunately, the large sample size requirements will likely prevent the subsequent initiation of a similar study, especially in an era where disease prevalence may be decreasing. At present, our study lends further support to other uncontrolled observations regarding the strategy of using EB viral load monitoring to inform preemptive reduction in immune suppression to prevent EBV disease and PTLD. However, additional experience directed at determining the precise role, if any, of the separate components of this approach (e.g. reduction of immune suppression versus the use of antiviral therapy) is needed to determine the most effective and efficient implementation of this strategy.