Increased incidence of cytomegalovirus infection in high-risk liver transplant recipients receiving valganciclovir prophylaxis versus ganciclovir prophylaxis



Optimal measures for the prevention of cytomegalovirus (CMV) in high-risk orthotopic liver transplant (OLT) patients are unknown. The charts of high-risk OLT recipients with 12 months of follow-up who were transplanted over a 44-month period were reviewed. The incidence of CMV disease in CMV-seropositive donor/CMV-seronegative recipient patients receiving valganciclovir or ganciclovir prophylaxis was compared. Sixty-six patients met the inclusion criteria and were treated with 1 of 3 prophylactic regimens: valganciclovir (900 mg daily; 27 patients), oral ganciclovir (1000 mg every 8 hours; 17 patients), or intravenous ganciclovir (6 mg/kg daily; 22 patients). Eight CMV cases occurred, all after completion of the prophylaxis. The combined incidence of CMV disease with intravenous and oral ganciclovir was lower than the incidence in valganciclovir recipients (P = 0.056; relative risk, 4.33; 95% confidence interval, 0.94–19.87). CMV disease occurred in 22.2% of valganciclovir recipients, 4.5% of intravenous ganciclovir recipients, and 5.9% of oral ganciclovir recipients. In conclusion, late-onset CMV disease occurred more frequently among high-risk liver transplant recipients treated with valganciclovir prophylaxis. The 4-fold higher incidence of CMV disease in our study supports the avoidance of valganciclovir for prophylaxis in high-risk OLT patients. Liver Transpl 15:963–967, 2009. © 2009 AASLD.

Cytomegalovirus (CMV) infection is responsible for significant morbidity and mortality among solid organ transplant recipients.1–3 Over the past decade, several pharmacological strategies have been used for prophylaxis following transplantation. Ganciclovir, in its oral and intravenous forms, has been shown to be an effective form of prophylaxis for CMV infection in high-risk orthotopic liver transplant (OLT) recipients.4–6 Unfortunately, oral ganciclovir requires frequent dosing as a result of its poor bioavailability, and intravenous delivery of ganciclovir requires long-term intravenous access. Valganciclovir, a valyl ester of ganciclovir with 10-fold greater bioavailability than oral ganciclovir, is an alternative.7 In a multicenter randomized trial of high-risk solid organ transplant recipients, valganciclovir was found to be clinically effective and well tolerated in comparison with oral ganciclovir in heart, kidney, and kidney-pancreas transplant recipients.8 In contrast, tissue-invasive CMV disease occurred more frequently in liver transplant recipients on valganciclovir versus ganciclovir.8, 9 On the basis of these results, valganciclovir did not receive Food and Drug Administration (FDA) approval for prophylaxis in high-risk liver transplant recipients, and its use in these patients is controversial. Despite the absence of FDA approval, many centers have adopted valganciclovir as standard prophylaxis for at-risk liver transplant recipients.

At our institution, valganciclovir replaced ganciclovir for CMV prophylaxis in all high-risk (donor positive, recipient negative) liver transplant recipients at the time of FDA approval for other solid organ recipients. However, the use of intravenous ganciclovir resumed after approximately 1 year when several cases of CMV disease were identified among patients on valganciclovir and concerns were raised regarding its effectiveness among liver transplant recipients. We reviewed our experience with CMV prophylaxis in liver transplant patients, comparing those receiving valganciclovir with those receiving intravenous or oral ganciclovir to better characterize the differences in outcomes observed in previous studies.


CI, confidence interval; CMV, cytomegalovirus; CrCl, creatinine clearance; D+, cytomegalovirus-seropositive donor; DM, diabetes mellitus; EtOH, alcoholic cirrhosis; FDA, Food and Drug Administration; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; IBD, inflammatory bowel disease; IV, intravenous; MELD, Model for End-Stage Liver Disease; MRSA, methicillin-resistant Staphylococcus aureus; NASH, nonalcoholic steatohepatitis; NS, not significant; NSCLC, non–small cell lung cancer; OLT, orthotopic liver transplant; PBC, primary biliary cirrhosis; PO, per os; PSC, primary sclerosing cholangitis; R−, cytomegalovirus-seronegative recipient; WBC, mean white blood cell count.


Study Design

This study was approved by the institutional review board at the study institution. A retrospective cohort study of OLT recipients at high risk for CMV was conducted at the Hospital of the University of Pennsylvania (Philadelphia, PA) between January 2000 and September 2004. High-risk patients were defined as CMV-seronegative recipients (R−) that received a CMV-seropositive donor (D+) liver. Patients were excluded for death within 30 days of transplantation or graft failure requiring retransplantation during the study period. The incidence of CMV disease in D+/R− patients that received either valganciclovir or ganciclovir prophylaxis was compared. A small number of these patients (n = 7) were participants in the PV16000 study.8

Data sources included paper charts, discharge summaries, operative reports, computerized laboratory databases, our institution's electronic transplant medical record, and hospital diagnosis codes. Our primary outcome measure was the development of CMV disease. CMV disease was defined as the presence of fever, malaise, leukopenia, atypical lymphocytosis, or thrombocytopenia in conjunction with a positive CMV antigenemia assay (pp65).8 Tissue-invasive CMV was defined as organ dysfunction not attributable to rejection in conjunction with evidence of CMV in a biopsy specimen of the affected organ.8 Other outcome measures included mortality, rejection episodes, and incidence of other infections.


High-dose intravenous methylprednisolone was used uniformly among all transplants during the first 3 days following transplantation. Maintenance immunosuppression was achieved with a combination of a calcineurin inhibitor (tacrolimus or cyclosporine) and tapering prednisone doses following transplantation. Mycophenolate was added in cases of significant renal insufficiency when full-dose calcineurin inhibitor dosing was contraindicated. Mycophenolate was discontinued when goal calcineurin inhibitor levels were achieved.

CMV Prophylaxis Regimens

At our institution, CMV prophylaxis consists of 1 of 3 regimens: valganciclovir dosed at 900 mg daily, oral ganciclovir dosed at 1000 mg 3 times daily, or intravenous ganciclovir dosed at 6 mg/kg daily. All prophylactic regimens were dose-adjusted for renal function. Prophylaxis was continued for the first 100 days after transplantation.

CMV Diagnosis

CMV antigenemia assays (pp65) were performed on patients that exhibited clinical or laboratory evidence of CMV disease as defined previously. There was no routine surveillance for patients either during or after completion of preventive therapy. Biopsies demonstrating tissue-invasive CMV infection were performed on patients when clinical suspicion of invasive disease was present.

Statistical Analysis

The Fisher exact test was used to compare the incidence of CMV disease between different treatment groups. All reported P values are 2-tailed. The Student t test was used to compare continuous variables between transplants with and without CMV disease. Analysis of variance was used to compare the 3 prophylaxis regimens for continuous variables. Chi-squared probabilities were used for comparing nonparametric patient characteristics in the 3 treatment groups.


A total of 446 patients were transplanted between January 2000 and September 2004. Seventy-two patients were confirmed to be high-risk (D+/R−). Six transplants were excluded from analysis secondary to graft failure during the study period (n = 3) or death within 30 days of transplantation (n = 3), and this left 66 patients for analysis. Twenty-seven (41%) of the transplants were treated with valganciclovir. The remainder were treated with oral ganciclovir [n = 17 (26%)] or intravenous ganciclovir [n = 22 (33%)]. All transplants were from unrelated donors. Baseline demographics were similar between the groups (Table 1).

Table 1. Patient Characteristics
 Total (n = 66)Valganciclovir (n = 27)PO Ganciclovir (n = 17)IV Ganciclovir (n = 22)P Value
  1. NOTE: There were no significant differences in patient demographics, renal function, underlying disease, or mycophenolate use between patient groups.

  2. Abbreviations: CI, confidence interval; CrCl, creatinine clearance (mL/minute) at 30 days post-transplant (Cockcroft-Gault); DM, diabetes mellitus; EtOH, alcoholic cirrhosis; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; IBD, inflammatory bowel disease; IV, intravenous; MELD, Model for End-Stage Liver Disease; NS, not significant (P > 0.05); PBC, primary biliary cirrhosis; PO, per os; PSC, primary sclerosing cholangitis; WBC, mean white blood cell count over the first 100 days post-transplant.

Age (95% CI)50 (2.3)51 (2.36)51 (3.92)48 (5.23)NS
Female sex32%32%24%39%NS
White race55%50%70%57%NS
African American race12%7%13%5%NS
Unknown race33%43%17%38%NS
MELD score (95% CI)21.5 (1.74)21.3 (2.87)22.9 (3.16)21.6 (3.7)NS
CrCl (95% CI) (cc/min)69.5 (8.89)76 (16.33)81.2 (17.47)65.8 (9.48)NS
WBC (95% CI) (× 103/μL)5.6 (0.6)5.8 (1.43)5.6 (0.88)5.3 (0.17)NS
Ascites at transplant66%64%71%65%NS
Primary liver diagnosis     

CMV Disease

Eight patients of the 66 total transplants developed CMV disease (12.1%). Of these, 6 cases were tissue-invasive (6 cases of CMV enteritis and 1 case of disseminated CMV with enteritis). The mean number of days from transplant to disease onset was 190 (range, 131–410 days). All patients developed either CMV syndrome or tissue-invasive CMV after completing a full 100-day course of prophylaxis. Six of the 8 cases of CMV disease occurred in patients given valganciclovir prophylaxis, representing a 22.2% incidence among valganciclovir recipients (Table 2). There was 1 case of CMV among the intravenous ganciclovir recipients, representing an incidence of 4.5%, and there was 1 case among the oral ganciclovir recipients, representing an incidence of 5.9%.

Table 2. Incidence of CMV Disease in Transplants Receiving Valganciclovir and Ganciclovir Prophylaxis
  1. NOTE: Valganciclovir recipients had a higher incidence of CMV disease in comparison with patients receiving ganciclovir per os or intravenously (P = 0.056; relative risk, 4.33; confidence interval, 0.94–19.87).

  2. Abbreviation: CMV, cytomegalovirus.

CMV6 (22.2%)2 (5.1%)
No CMV21 (77.8%)37 (94.8%)

The average tacrolimus dose was similar between treatment groups (P = 0.27) and between patients with and without CMV disease (P = 0.13). There was no difference in average peak (P = 0.40) or trough (P = 0.47) tacrolimus levels between CMV disease and CMV disease–free groups. Likewise, there was no difference in tacrolimus peaks (P = 0.78) or troughs (P = 0.12) between treatment groups. One patient received cyclosporine in lieu of tacrolimus. Mycophenolate was used as an additional immunosuppressive agent in 65% of transplants. Mycophenolate use was similar between treatment groups (P = 0.91) and between transplants that developed CMV and those that did not (P = 0.57). Neither immunosuppressive induction agents nor CMV immunoglobulin was used in either group. There was no difference in creatinine clearance between treatment groups (P = 0.39) or between CMV disease and CMV disease–free transplants (P = 0.26). There was no difference in the mean white blood cell count between the 3 treatment groups during the prophylaxis period (P = 0.23).

Ten patients died among the 66 transplants reviewed (15%). All patients with CMV disease responded to standard treatment with intravenous ganciclovir, and no deaths were attributed to CMV disease. One patient treated for CMV enteritis died 17 months later from metastatic lung cancer. There were no cases of ganciclovir-resistant viral infection. There were 7 episodes of rejection reported among all transplants (10.6%). Four rejection episodes occurred in patients with CMV disease (50%; P = 0.0025; relative risk, 10; 95% confidence interval, 2.7–37.8). All episodes of rejection occurred after CMV infection (Table 3). The average white blood cell count during the prophylaxis period was similar between all 3 treatment groups (P = 0.23). The incidence of other infections, ranging from urinary tract infections to sepsis, was similar between treatment groups (P = 0.19). There were no line-associated infections in the intravenous ganciclovir group. Other infections occurred more frequently in patients that developed CMV (62.5%) versus those that did not (36.7%). However, this trend did not reach statistical significance (P = 0.11).

Table 3. Patient Outcomes with CMV Disease
PatientProphylaxis (Days)Primary Liver DiseaseCMV Diagnosis (Days from Transplantation)Rejection (Days from Transplantation)Death (Cause)Other Infections (Days from Transplant)
  1. Abbreviations: CMV, cytomegalovirus; EtOH, alcoholic cirrhosis; HCV, hepatitis C virus; IV, intravenous; MRSA, methicillin-resistant Staphylococcus aureus; NASH, nonalcoholic steatohepatitis; NSCLC, non-small cell lung cancer; PBC, primary biliary cirrhosis; PO, per os.

1ValganciclovirEtOHEnteritis, hepatitis (178)Yes (915)No 
2ValganciclovirHCVCMV syndrome (152)NoNoHCV (82)
3ValganciclovirPBCCMV syndrome (411)NoNo 
4ValganciclovirNASHDisseminated (215)NoNoParainfluenza pneumonia (166), MRSA soft tissue infection (355)
5ValganciclovirHCVEnteritis (133)Yes (167)NoHCV (211)
6ValganciclovirHCVEnteritis (155)NoNoHCV (484)
7Ganciclovir POCryptogenicEnteritis (131)Yes (410)Yes (NSCLC) 
8Ganciclovir IVEtOH/HCVEnteritis (138)Yes (563)NoFungal brain abscess (30)


In a large, prospective, randomized, multicenter trial with OLT recipients, the incidence of CMV disease at 6 months was 19% in the valganciclovir group versus 12% in the ganciclovir group.8, 9 Although these results were statistically insignificant at 6 months, the majority of CMV cases in this prospective trial were tissue-invasive disease (14% in the valganciclovir group versus 3% in the ganciclovir group).9 Valganciclovir failed to receive FDA approval for CMV prophylaxis in OLT recipients on the basis of these data. Similarly, our results reveal an increased frequency of late-onset CMV disease after the 100-day prophylaxis period among D+/R− OLT recipients receiving valganciclovir prophylaxis. Of note, our results include cases of both CMV syndrome and tissue-invasive CMV disease. This differs from the results of Paya et al.'s multicenter trial, which showed a higher incidence of tissue-invasive CMV at 6 months post-transplant without a higher incidence of CMV syndrome.8, 9 In contrast to our results, other retrospective studies showed no difference in the incidence of CMV disease between valganciclovir and ganciclovir recipients.10, 11 These reports were limited by the small numbers of patients on ganciclovir, however (n = 9). The increased incidence of rejection among patients with CMV in our cohort also reiterates findings of previous studies and underscores the role for effective prophylaxis.12, 13

There were no cases of CMV disease in any of the patients during the allotted prophylaxis window. Rather, the period following discontinuation of prophylaxis revealed a 4-fold higher incidence of CMV disease among valganciclovir recipients versus ganciclovir recipients. The reason for this discrepancy is unclear. Pharmacological data for several organ transplant types, comparing oral ganciclovir to valganciclovir, showed higher systemic ganciclovir levels among valganciclovir recipients and an associated lower level of viremia during the prophylaxis window.14 However, subgroup analysis showed that the incidence of CMV disease in liver transplant recipients reached 20% after 12 months, the highest rate among organ types.14 This variable incidence of CMV seen with valganciclovir among different organ transplants raises the possibility of an issue with drug metabolism unique to liver transplant recipients.

Previous reports have suggested that the administration of mycophenolate with valganciclovir could potentially induce a competitive inhibition of the esterase responsible for the conversion of valganciclovir to ganciclovir in the gut and liver.15 This hypothesis was not supported by our results, which showed similar rates of mycophenolate use between treatment groups. Likewise, it is unlikely that malabsorption of valganciclovir was responsible for the higher incidence of CMV disease in our cohort. Earlier studies demonstrated greater concentrations of systemic ganciclovir and increased suppression of CMV viremia with valganciclovir in comparison with oral ganciclovir.14 Similarly, valganciclovir achieved an area under the curve similar to that of intravenous ganciclovir among liver transplant recipients.7 Confounding factors potentially contributing to impaired absorption of valganciclovir include comorbid conditions such as inflammatory bowel disease and chronic diarrhea related to mycophenolate use. However, in our population, there was no difference in the incidence of inflammatory bowel disease between treatment groups. Mycophenolate use was similar between groups as well. Ascites could create a larger volume of distribution, thereby lowering serum concentrations of medications. Our results showed no difference in the incidence of ascites between groups, however.

Finally, higher rates of CMV among valganciclovir recipients could be a manifestation of delayed seroconversion. Previous work demonstrated significantly lower rates of immunoglobulin G seroconversion against CMV by day 100 in patients receiving valganciclovir prophylaxis versus ganciclovir recipients.16 Conversion to CMV immunoglobulin G–positive status was shown to have a protective effect at 6 months post-transplantation in the same study. Of note, OLT recipients had the highest rates of seroconversion among organ types at all time points. In comparison, renal transplants had the lowest rates of seroconversion but did not show a difference in CMV disease incidence between prophylaxis regimens as was seen in OLT patients.8, 16 Whether rates of seroconversion were different between prophylaxis regimens in our patient population is unknown as serologies were not routinely performed following transplantation.

Our study has several limitations. Among them, the retrospective, nonrandomized nature of our data may have influenced our analysis. The single-center aspect of our study design also limits the universality of our results as well as our total sample size. The use of 3 separate CMV prophylaxis strategies raises the question of bias in the decision to use one prophylaxis strategy over another. However, the use of one prophylaxis agent over another was dependent on the set protocol at the time rather than the discretion of the clinician caring for the patient. The protocol for D+/R− patients was changed from valganciclovir to intravenous ganciclovir after the rate of CMV disease was noted to rise following the introduction of valganciclovir.

In summary, our data reveal an increased incidence of late-onset CMV disease among high-risk liver transplant recipients that were treated with valganciclovir prophylaxis. The reason for this trend is not immediately apparent. Potential explanations include altered metabolism of valganciclovir among liver transplant recipients and/or delayed seroconversion. Additionally, it is possible that the use of valganciclovir preventive therapy might specifically affect CMV-specific T cell responses such that the risk of subsequent CMV disease is greater. Future research should attempt to elucidate these issues further, potentially with a larger, prospective trial comparing intravenous ganciclovir and valganciclovir in OLT patients or with a prospective trial using a longer prophylaxis period with valganciclovir. Regular posttransplant CMV surveillance and pre-emptive treatment might also be a reasonable alternative to prophylaxis among high-risk OLT patients. Regardless of the cause, the higher incidence of CMV disease in our study warrants caution when valganciclovir is used for prophylaxis in high-risk OLT patients.