Telephone: 402-559-8650; FAX: 402-559-5581;
Article first published online: 7 DEC 2012
Copyright © 2012 American Association for the Study of Liver Diseases
Volume 18, Issue 12, pages 1440–1447, December 2012
How to Cite
Kalil, A. C., Mindru, C., Botha, J. F., Grant, W. J., Mercer, D. F., Olivera, M. A., McCartan, M. A., McCashland, T. M., Langnas, A. N. and Florescu, D. F. (2012), Risk of cytomegalovirus disease in high-risk liver transplant recipients on valganciclovir prophylaxis: A systematic review and meta-analysis. Liver Transpl, 18: 1440–1447. doi: 10.1002/lt.23530
This study was exempted by the institutional review board.
Andre C. Kalil and Diana F. Florescu contributed to the conception and design. Cezarina Mindru and Diana F. Florescu contributed to the literature search. Andre C. Kalil contributed to the statistical analysis. Andre C. Kalil, Cezarina Mindru, Jean F. Botha, Wendy J. Grant, David F. Mercer, Marco A. Olivera, Megan A. McCartan, Timothy M. McCashland, Alan N. Langnas, and Diana F. Florescu contributed to the interpretation of the results. Andre C. Kalil wrote the first draft. Andre C. Kalil, Cezarina Mindru, Jean F. Botha, Wendy J. Grant, David F. Mercer, Marco A. Olivera, Megan A. McCartan, Timothy M. McCashland, Alan N. Langnas, and Diana F. Florescu contributed to the critical review of the manuscript. Andre C. Kalil, Cezarina Mindru, Jean F. Botha, Wendy J. Grant, David F. Mercer, Marco A. Olivera, Megan A. McCartan, Timothy M. McCashland, Alan N. Langnas, and Diana F. Florescu contributed to the conclusions.
- Issue published online: 7 DEC 2012
- Article first published online: 7 DEC 2012
- Accepted manuscript online: 6 AUG 2012 05:19AM EST
- Manuscript Accepted: 24 JUL 2012
- Manuscript Received: 3 MAY 2012
Valganciclovir (VGC) was approved by the Food and Drug Administration in 2004 as cytomegalovirus (CMV) prophylaxis except for liver transplant recipients because of their high incidence of CMV disease with this drug. However, surveys have shown its common off-label use for CMV prophylaxis in liver transplant recipients. We aimed to evaluate the risk of CMV disease with VGC prophylaxis in liver transplant recipients. All studies that evaluated liver transplant recipients and used VGC (900 or 450 mg daily) for the prevention of CMV disease were included. Five controlled studies (n = 483) were pooled with a random effects model; five single-arm studies (n = 380) were pooled for the prevalence rate of CMV disease. The risk of CMV disease with VGC versus ganciclovir was 1.81 [95% confidence interval (CI) = 1.00-3.29, P = 0.05, I2 = 0%]. For high-risk (donor-positive/recipient-negative) patients, the risk of CMV disease was 1.96 (95% CI = 1.05-3.67, P = 0.035, I2 = 0%). The risk of CMV disease remained significant with 900 mg of VGC daily (P = 0.04) but not with 450 mg of VGC daily (P = 0.76). The risk of leukopenia with VGC was 1.87 (95% CI = 1.03-3.37, P = 0.04, I2= 0%). In single-arm trials, the overall CMV disease rate was 12% (95% CI = 9%-16%, P < 0.001), and the rate for high-risk patients was 20% (95% CI = 10%-38%, P = 0.002). In conclusion, 900 mg of VGC daily may not be safe as CMV prophylaxis in high-risk liver transplant recipients because of the significant 2-fold increase in the risk of CMV disease and the 1.9-fold increase in the risk of leukopenia. Alternative CMV prophylaxis should be used for liver transplant recipients. Liver Transpl, 2012. © 2012 AASLD.
Valganciclovir (VGC) was approved by the Food and Drug Administration in September 2004 as prophylaxis for cytomegalovirus (CMV) infections in high-risk [donor-positive/recipient-negative (D+/R−)] solid organ transplant recipients. The original clinical trial, which was called PV16000,1 enrolled patients with several types of allografts. However, liver transplant recipients showed higher rates of CMV disease and CMV tissue-invasive disease with the new drug VGC (19% and 14%, respectively) versus the control drug ganciclovir (GCV; 12% and 3%, respectively).2 On the basis of these results, the Food and Drug Administration did not approve VGC for use in high-risk liver transplant recipients.
In a more recent publication,3 a re-analysis of the PV16000 trial showed that there were statistically significant interactions with the allograft type and VGC treatment for both CMV disease (P = 0.008) and CMV tissue-invasive disease (P = 0.009); this suggested that the higher risk of developing CMV disease in liver transplant recipients was not just a chance finding. More specifically, the risk of CMV tissue-invasive disease in liver transplant recipients was 4.5 times higher with VGC versus GCV (P = 0.04).
Despite the negative results of the PV16000 trial and the Food and Drug Administration decision, many transplant centers have continued to use VGC as CMV prophylaxis in liver transplant recipients. In fact, VGC has been used in approximately two-thirds of surveyed liver transplant centers.4 Furthermore, recent CMV international consensus guidelines recommend 900 mg of VGC daily for liver transplant recipients, mainly on the basis of expert opinion (evidence level III).5
It remains unclear whether the negative results of the PV16000 trial for the liver transplant recipient group were due to a chance finding (statistical error of type I) or a detrimental effect secondary to VGC prophylaxis. We performed a systematic review and meta-analysis of all studies to evaluate the efficacy and safety of prophylactic VGC in liver transplant recipients.
MATERIALS AND METHODS
PubMed, Embase, Cochrane Library, Evidence-Based Medicine (BMJ), and American College of Physicians Journal Club databases were searched from their inception through January 2012. In addition, available abstracts from the American Transplantation Congress and the Infectious Diseases Society of America were searched. The Quality of Reporting of Meta-Analyses (QUOROM) guideline criteria were used for the search and study design methodology6 (Fig. 1). The keywords used for the search were valganciclovir, Valcyte, organ transplantation, liver, hepatic, cytomegalovirus, CMV, prophylaxis, randomized, cohort, and case-control studies. No language restrictions were applied. Two authors (D.F.F. and C.M.) performed the literature search and the study selection separately. A third author (A.C.K.) resolved any disagreements, and a final consensus was reached by all the authors.
According to the inclusion criteria, all studies that included liver transplant recipients and used VGC (900 or 450 mg daily) for the prevention of CMV disease were included in the analysis. According to the exclusion criteria, studies that did not evaluate liver transplant recipients and/or universal VGC prophylaxis or that used universal VGC prophylaxis in both arms were excluded.
The data collection included the following: authors; publication years; study designs; allograft types; sexes; ages; sample sizes; CMV donor/recipient serostatuses; types of induction therapy; types of maintenance immunosuppressive therapy; CMV end-organ disease; CMV syndrome; CMV viremia; VGC doses; types of controls; prophylaxis durations; lengths of follow-up; and leukopenia, neutropenia, acute rejection, allograft loss, and mortality rates.
CMV syndrome was defined as the presence of CMV in the blood according to polymerase chain reaction (PCR), antigenemia, or culturing plus 1 or more of the following: malaise, a fever > 38°C for 2 or more days, leukopenia, an atypical lymphocyte level ≥ 5%, thrombocytopenia, and increased liver transaminases (double the upper limit). CMV organ disease was defined as pneumonia, hepatitis, gastrointestinal disease, nephritis, or central nervous system disease as described by Kotton et al.5 Leukopenia was defined as less than 3500 cells/mm3, and neutropenia was defined as less than1500 cells/mm3 during the study follow-up. Rejection was defined as acute allograft rejection reported up to 12 months after transplantation. Mortality was defined as a death reported up to 12 months after transplantation.
The Peto method was used for the adjusted odds ratios, and the Q statistic method and the I2 method were used to assess heterogeneity. All data were pooled with a random effects model according to the DerSimonian-Laird methodology.7 The software was Comprehensive Meta-Analysis version 2 (Biostat, Englewood, NJ). Egger regression and Begg-Mazumdar methods were used to evaluate publication bias.8, 9 The Meta-Analysis of Observational Studies in Epidemiology guidelines for observational studies were followed.10 Studies that did not provide adjustments for the most important risk factor for developing CMV disease (ie, the CMV serostatus) were adjusted for the purposes of this meta-analysis. This was done through the extraction of outcomes by the specific CMV serostatus in each study.
In all, 10 studies were included in the meta-analysis: 5 controlled studies (n = 483 patients)1, 11-14 and 5 single-arm studies (n = 380 patients)15-19 (Table 1). All studies used prophylaxis for a total of 3 months.
|Study||Study Design||Time Period||Immunosuppressive Regimen: Control and VGC Patients||Diagnostic Method||Sample Size: Control Patients/VGC Patients (n)||Mean Age: Control Patients/VGC Patients (Years)||CMV Serostatus (n)||Control Arm (Dose)||Treatment Arm (Dose)*|
|Paya et al.1 (2004)||Randomized||4/2000-8/2001||Not available||PCR||59/118||45.3/45.7||High risk (177)||GCV (3 g/day)||VGC (900 mg/day)|
|Park et al.11 (2006)||Case control||1/2001-4/2003||Maintenance: FK506, MMF, steroids||PCR||49/60||50/49||High risk (24)||GCV (3 g/day)||VGC (450 mg/day)|
|Induction: ATG, OKT3||Moderate risk (85)|
|Arthurs et al.12 (2007)||Cohort||1/2000-12/2004||Maintenance: FK506, MMF, prednisone||PCR||9/58||49.5||High risk (67)||GCV (3 g/day)||VGC (900 mg/day)|
|Brady et al.13 (2009)||Cohort||8/1996-9/2006||Maintenance: FK506, MMF, prednisone||pp65 antigenemia and PCR||21/43||50/51||High risk (64)||GCV (3 g/day)||VGC (450 mg/day)|
|Shiley et al.14 (2009)||Cohort||1/2000-9/2004||Maintenance: FK506, CSA, MMF, prednisone||pp65 antigenemia||39/27||50/52||High risk (66)||GCV (3 g/day)||VGC (900 mg/day)|
|Akalin et al.15 (2003)||Case control||1/2001-3/2002||Maintenance: FK506, CSA, MMF, prednisone||Hybrid-capture RNA-DNA hybridization test (PCR)||NA/5||Not available||High risk (5)||NA||VGC (450 mg/day)|
|Jain et al.16 (2005)||Cohort||7/2001-5/2003||Maintenance: FK506, MMF, steroids||PCR||NA/203||53||High risk (73)||NA||VGC (900 mg/day)|
|Moderate risk (95)|
|Low risk (35)|
|Dupuis et al.17 (2007)||Cohort||8/2002-8/2004||Maintenance: FK506, prednisone||pp65 antigenemia and PCR||NA/85||51||High risk (46)||NA||VGC (450 mg/day)|
|Induction: ATG, basiliximab||Moderate risk (34)|
|Low risk (5)|
|Montejo et al.18 (2009)||Cohort||1/2003-2/2007||Maintenance VGC: FK506, MMF, prednisone||pp65 antigenemia||NA/23||47||High risk (23)||NA||VGC (900 mg/day)|
|Donnelly et al.19 (2009)||Cohort||1/2005-1/2006||Not available||PCR||NA/64||48||High risk (21)||NA||VGC (900 mg/day)|
|Moderate risk (43)|
Meta-Analysis of the Controlled Studies
Independently of the CMV serostatus, the risk of CMV disease in all liver transplant recipients who received prophylaxis with VGC (5 studies, n = 483) was 1.81 [95% confidence interval (CI) = 1.00-3.29, P = 0.05, I2 = 0%; Fig. 2A]. When only the high-risk (D+/R−) patients were evaluated (4 studies, n = 374), the risk of CMV disease was 1.96 (95% CI = 1.05-3.67, P = 0.035, I2= 0% (Fig. 2B). The risk of leukopenia (3 studies, n = 528) was 1.87 (95% CI = 1.03-3.37, P = 0.04, I2= 0%). The risks of acute rejection (3 studies, n = 539) and death (2 studies, n = 482) were 0.82 (95% CI = 0.55-1.22, P = 0.33, I2 = 0%) and 1.29 (95% CI = 0.58-2.85, P = 0.53), respectively. The risks of neutropenia and allograft loss could not be evaluated because of a lack of reporting in most studies.
Meta-Analysis of the Single-Arm Studies
Independently of the CMV serostatus, the rate of CMV disease in all liver transplant recipients who received VGC prophylaxis (5 studies, n = 380) was 12% (95% CI = 9%-16%; Fig. 3A). When only high-risk (D+/R−) patients were evaluated (4 studies, n = 107), the rate of CMV disease was 20% (95% CI = 10%-38%; Fig. 3B). The rate of leukopenia (2 studies, n = 108) was 15% (95% CI = 9%-23%), and the rate of acute rejection (2 studies, n = 108) was 6% (95% CI = 3%-12%). The rates of neutropenia, allograft loss, and death could not be evaluated because of a lack of reporting.
Comparison With the PV16000 Trial
When the PV16000 results1 were removed from our meta-analysis, the CMV disease rates (and 95% CIs) for high-risk liver transplant recipients who received VGC remained consistently high in all analyses: 19.9% (95% CI = 15%-25%) in controlled studies, 19.6% (95% CI = 13%-28%) in single-arm studies, and 18.6% (95% CI = 13%-27%) in the PV16000 trial (Fig. 4).
The CMV disease outcomes in the meta-analyses of the controlled studies were further evaluated by the VGC dose. The risk of CMV disease was 1.98 (95% CI = 1.02-3.83, P = 0.04, I2 = 0%) in the group receiving a daily dose of 900 mg (3 studies, n = 310) and 1.25 (95% CI = 0.31-4.99, P = 0.76, I2 = 0%) in the group receiving a daily dose of 450 mg (2 studies, n = 173). A sensitivity analysis was performed to evaluate CMV disease outcomes on the basis of studies that made adjustments for the CMV serostatus in the original publications: 2.35 (95% CI = 0.86-6.43, P = 0.09, I2 = 0%) with adjustments (2 studies, n = 243) and 1.32 (95% CI = 0.47-3.70, P = 0.60, I2 = 0%) without adjustments (3 studies, n = 240). Also, a sensitivity analysis was performed according to the quality assessment of each controlled study: no meta-analysis for randomized studies (only 1 study), 2.46 (95% CI = 0.96-6.27, P = 0.06, I2 = 0%) for cohort studies (3 studies), and no meta-analysis for case-control studies (only 1 study). The risks of leukopenia and rejection could not be evaluated by the specific VGC dose. An evaluation by the type of immunosuppression, the duration of prophylaxis, and the length of follow-up showed similar results (data not shown).
No significant publication bias was detected by an Egger regression (intercept = −0.038, standard error = 1.207, P = 0.97) or by a Begg-Mazumdar rank correlation (Kendall's τ = −0.300, P = 0.46).
Our findings suggest that the use of VGC prophylaxis to prevent CMV infections in high-risk liver transplant recipients is associated with a significant 2-fold increase in the risk of CMV disease after adjustments for the baseline CMV serostatus. The consistency and very low heterogeneity of our results make the previous hypothesis of a statistical type I error unlikely, and they further confirm the higher risk of developing CMV disease in liver transplant recipients found in the PV16000 trial with VGC prophylaxis.
The high consistency of the CMV disease rates (18%-20%) among all the study types (ie, the PV16000 trial, the meta-analysis of the controlled studies, and the meta-analysis of the single-arm studies) provides additional confirmation for the high rate of CMV disease in liver transplant recipients in studies performed at different institutions as part of the real-life use of VGC prophylaxis.
Why would liver transplant recipients be at higher risk of developing CMV disease while they were receiving VGC prophylaxis in comparison with recipients of other allografts? There are several potential explanations: the conversion of VGC to GCV may be decreased in the post–liver transplant period because of an esterase deficiency (quantitative or qualitative) secondary to temporary hepatic dysfunction, bowel dysfunction, or both.16 Also, mycophenolate mofetil (MMF) requires the liver esterase for its conversion to mycophenolic acid16; it is possible that competition between MMF and VGC for the same liver esterase could further increase its deficiency. In addition, the strong suppression of CMV replication by VGC could decrease the host's immune exposure to low-level CMV viremia and subsequently lead to an inadequate CMV-specific T cell response; this could lead to brisker CMV reactivation after prophylaxis is stopped.20 In agreement with the previously posed immunological hypothesis, a comprehensive meta-analysis recently published by our group21 showed a lower incidence of CMV disease with a lower dose of VGC (450 mg daily) versus the conventional dose of 900 mg daily. It is possible that this lower dose allows a better CMV-specific T cell response. The same study also showed that the rate of acute allograft rejection was significantly lower with 450 mg of VGC daily versus 900 mg daily; this lower rejection rate with the lower VGC dose gives further support to the aforementioned hypothesis because less CMV disease after prophylaxis (450 mg) would be expected to be associated with fewer indirect CMV effects such as allograft rejection.22
Another important finding of our study is that the rate of VGC-induced leukopenia remained significant and consistently high among all study types. This side effect is clearly unwelcome in a transplant host who is already severely immunosuppressed and may have splenic dysfunction, which can lead to more serious infections in the first few months after transplantation. We would like to evaluate the actual risk of other infections with drug-induced leukopenia, but unfortunately, most studies did not report other bacterial or fungal infections during or following the prophylaxis regimen.
The limitations of our study include the potential for publication bias. Even though our analysis did not show this bias according to 2 methods, this bias cannot be completely ruled out because of the small number of studies. Also, the low I2 values show that the studies were appropriately pooled, but some degree of heterogeneity could still have been present. The safety of the 450-mg dose needs more evidence because only 2 studies used this dose. Our findings cannot be generalized to hybrid strategies, such as the combination of different lengths of prophylaxis with preemptive therapy. Moreover, the development of CMV resistance with low-dose VGC needs to be further studied; however, because oral GCV (3 g/day) has been associated with a very low resistance rate (<2%),23 VGC at 450 mg/day would not be expected to be associated with more resistance. A few liver-kidney transplant recipients were part of the analysis, and it would be informative to have their CMV risk analyzed separately from the risk of the liver transplant recipients, but this was not possible. Lastly, a selection bias cannot be completely ruled out because of the retrospective nature of the included studies.
What are the possible solutions for CMV prophylaxis in liver transplant recipients until we accumulate more data on this important issue? There are several efficacious and safe solutions. Both high-dose acyclovir and valacyclovir have shown excellent efficacy in randomized trials and meta-analyses.22, 24, 25 Oral GCV has also shown high efficacy in preventing CMV disease without the complications recently seen with VGC (900 mg daily). However, Roche Pharmaceuticals stopped manufacturing GCV, and the generic form is not consistently available. Another alternative is the use of low-dose VGC (450 mg daily), which has been associated with a lower risk of developing both late CMV disease and leukopenia in comparison with the daily dose of 900 mg21; this was observed in an indirect treatment comparison meta-analysis, but no head-to-head comparisons have been made. Also, the use of preemptive therapy, when it is feasible, could provide another option for preventing CMV in liver transplant recipients.22
In conclusion, in comparison with GCV, VGC prophylaxis is associated with a significant increase in the risk of developing CMV disease and leukopenia for all liver transplant recipients. The very low heterogeneity and consistency of our findings suggest that 900 mg of VGC daily may not be safe for prophylaxis in liver transplant recipients. These results lend support to the original Food and Drug Administration analysis. Alternative CMV prophylaxis with 450 mg of VGC daily, GCV, high-dose acyclovir, or valacyclovir should be used for liver transplant recipients.
The authors thank Ms. Ashley Calhoon for her excellent administrative support.
- 1for Valganciclovir Solid Organ Transplant Study Group. Efficacy and safety of valganciclovir vs. oral ganciclovir for prevention of cytomegalovirus disease in solid organ transplant recipients. Am J Transplant 2004; 4: 611-620., , , , , , et al.;
- 2U.S. Food and Drug Administration. Dear healthcare professional letter for Valcyte (valganciclovir HCl tablets). http://www.fda.gov/Safety/MedWatch/SafetyInformation/SafetyAlertsforHumanMedicalProducts/ucm169502.htm. Published September 2003. Accessed July 2012.