High-Dose Acyclovir Compared with Short-Course Preemptive Ganciclovir Therapy To Prevent Cytomegalovirus Disease in Liver Transplant Recipients

Methods

Study Design

All patients having liver transplantation at our institution were randomly assigned to one of the two prophylactic groups. Randomization was stratified by the CMV antibody status of the recipient and the donor. Patients in the control group received 800 mg of acyclovir orally, four times daily, beginning immediately after transplantation as described by Balfour and colleagues [8]; patients continued receiving acyclovir (Zovirax; Burroughs Wellcome, Research Triangle Park, North Carolina) for 24 weeks postoperatively. The dosage of acyclovir was adjusted for impaired renal function as follows: If the creatinine clearance was greater than 50 mL/min, then patients received 800 mg of acyclovir four times a day; if the creatinine clearance was 25 to 50 mL/min, they received 800 mg of acyclovir three times a day; if the clearance was 10 to 25 mL/min, they received 800 mg of acyclovir twice a day; and if the clearance was less than 10 mL/min, they received 800 mg of acyclovir once daily. Surveillance cultures for CMV (buffy coat and urine) were obtained at 2, 4, 6, 8, 12, 16, and 24 weeks postoperatively for all study patients using the shell vial culture method.

The experimental group did not receive acyclovir, but intravenous ganciclovir (Cytovene; Syntex, Palo Alto, California), 5 mg/kg twice daily, was administered for 7 days only if surveillance cultures yielded CMV (Figure 1). Ganciclovir dosage was modified for abnormal creatinine clearance as follows: If the creatinine clearance was 80 mL/min or more, then patients received 5 mg/kg of ganciclovir twice daily; if the creatinine clearance was 50 to 79 mL/min, they received 2.5 mg/kg of ganciclovir twice daily; if the clearance was 25 to 49 mL/min, they received 2.5 mg of ganciclovir daily; if the clearance was less than 25 mL/min, they received 1.25 mg/kg of ganciclovir daily. The study was continued for 24 weeks postoperatively.

Immunosuppression

All patients received 0.10 mg/kg of tacrolimus (Fujisawa, Deerfield, Illinois) as a continuous drip for 24 hours, until they were able to take oral medications. The oral dosage of tacrolimus was 0.10 mg/kg every 12 hours. Subsequent dosage adjustments were made as indicated by clinical course and plasma levels of tacrolimus. Methylprednisolone, 1 g, was given immediately after revascularization of the graft. Methylprednisolone, 20 mg, was given intravenously immediately after transplantation and daily thereafter until the oral route was established, at which time 20 mg of prednisone was administered daily. During the subsequent months, prednisone was slowly tapered. Rejection episodes were treated with boluses of 1 g of methylprednisolone with or without steroid recycles (prednisone decreasing daily by 40 mg from an initial starting dose of 200 mg). Muromonab-CD3 (Orthoclone OKT3, Ortho Pharmaceuticals, Raritan, New Jersey) was used for steroid-resistant rejection.

Definition of Viral Infections

Cytomegalovirus Infection

Serologic test results for cytomegalovirus were determined using an enzyme immunosorbent assay, and titers of 0.79 or more were considered positive. All patients received blood products that were neither tested nor screened for CMV antibody. Primary infection was defined as isolation of virus or seroconversion in a patient who was seronegative before transplantation. Reactivation infection was diagnosed by isolation of virus in a seropositive recipient.

Cytomegalovirus Disease

Clinical diseases caused by CMV included the viral syndrome, localized CMV disease, and disseminated CMV disease. Identification of the viral syndrome caused by CMV required the following: 1) positive culture for CMV; 2) temperature of 38 °C or more with no other source to account for it; and 3) one of the following findings: leukocyte count 4000/mm³ or less, atypical lymphocytes 3% or more, and platelets 100 000/mm³ or less. Localized CMV disease was defined as tissue invasion of a single organ determined histopathologically with or without culture of the virus from tissue. Disseminated disease was defined as tissue involvement of two or more noncontiguous organ sites.

Identification of Other Viruses

Antibodies against viral capsid antigen, early antigen, and Epstein-Barr virus (EBV) nuclear antigen were determined preoperatively in all patients. Patients were defined as having a symptomatic EBV infection if they had EBV-associated lymphoproliferative disease identified by the presence of EBV DNA in tissue using nucleic acid hybridization. Antibody titers to detect asymptomatic increases in EBV were not routinely determined.

Herpes simplex virus infection was defined as the presence of typical symptomatic oral or genital ulcers. Varicella zoster virus infection was determined clinically by the presence of typical dermatomal lesions with or without viral isolation.

Statistical Analysis

Analysis was done using the Prophet System (BBN Systems and Technologies, Division of Research Resources, National Institutes of Health, Bethesda, Maryland). Baseline characteristics (age, Child-Pugh score) were compared using the Fisher exact or t-test. We estimated that at least 20 patients in each group would be needed to detect a decrease in CMV disease from 35% with standard acyclovir prophylaxis to 5% with ganciclovir (α = 0.05, power = 0.8). The incidence of infection or disease was compared using a two-tailed Fisher exact test. A Kaplan-Meier estimate was used to examine the number of days from transplant until first CMV infection for each group. The two curves were compared using the Mantel-Cox log-rank test. Similar curves were constructed for CMV disease.

Results

The study sample consisted of 47 consecutive adult male patients who received liver transplants at the Pittsburgh Veterans Affairs Medical Center during a 2-year period and who survived at least 72 hours postoperatively. These included 44 patients who had primary transplants and 3 who had re-transplants (2 transplanted once previously and another twice previously). Of 47 patients enrolled in the study, 24 were randomly assigned to the acyclovir group and 23, to the experimental group. The two patient groups were similar at entry in terms of all baseline characteristics measured (Table 1). The Child-Pugh scoring system was used to assess the severity of liver disease in the two groups before transplantation [16].

Cytomegalovirus Infection and Disease

Shedding of CMV before the onset of CMV disease occurred in 25% (6 of 24) of patients receiving acyclovir prophylaxis and in 22% (5 of 23) of patients receiving no prophylaxis (experimental group) (Figure 2). Seventeen percent (4 of 24) of the patients in the acyclovir group and 4% (1 of 23) in the experimental group did not have previous shedding and developed CMV disease as the first manifestation of CMV infection. Thus, 42% (10 of 24) in the acyclovir group and 26% (6 of 23) in the experimental group had CMV infection (16% difference; 95% CI, 10% to 42%; P > 0.2). All CMV infections were diagnosed by viral isolation. One of 6 infections in the experimental group was a primary infection, whereas 5 were reactivation infections (3 in the recipients of seronegative donor allografts and 2 in the recipients of seropositive donor grafts). Of the 10 infections in the acyclovir group, 2 were primary infections, and the other 8 were reactivation infections (4 in the recipients of seronegative donor allografts and 4 in the recipients of seropositive donor allografts). Cytomegalovirus infection was diagnosed at a mean of 34 days (range, 14 to 43 days) in the experimental group and at a mean of 39 days (range, 13 to 104 days) in the acyclovir group (Figure 3).

Cytomegalovirus disease developed in 29% (7 of 24) of the patients in the acyclovir group compared with 4% (1 of 23) in the experimental group (25% difference; CI, 4% to 46%; P < 0.05 (Figure 4). Shedding of CMV was documented in five patients in the experimental group [viremia in four and viruria in one]; all received pre-emptive ganciclovir as described in the protocol (see Figure 1). None of these five patients developed subsequent CMV disease. One patient in the experimental group did not have CMV shedding before the occurrence of CMV disease.

In the acyclovir group, 6 of 24 patients shed CMV before the onset of CMV disease (viremia in 1 and viruria in 5). Fifty percent (3 of these 6 patients) subsequently developed CMV disease. The remaining 4 patients with CMV disease in the acyclovir group did not have detectable shedding before the CMV disease. The symptomatic patients with CMV disease in the acyclovir group had disseminated CMV disease (1 patient with hepatitis and enteritis), CMV hepatitis (2 patients), CMV enteritis (1 patient), and the viral syndrome (3 patients). Mean time to the onset of CMV disease in the acyclovir group was 51 days (range, 20 to 104 days; see Figure 3). Cytomegalovirus disease (the viral syndrome) occurred at 45 days in the single patient in the experimental group.

Mortality and Graft Loss

Thirteen percent (6 of 47) of the patients died (3 in the acyclovir group and 3 in the experimental group). Deaths occurred at 34, 232, and 345 days postoperatively due to intra-abdominal abscess, invasive fungal infection, and chronic rejection with hepatitis C-related recurrent hepatitis, respectively, in the three patients in the acyclovir group. Two patients in the experimental group died postoperatively of fungal infections on days 24 and 30. Neither of these patients had evidence of CMV infection after death or at autopsy. A third patient in the ganciclovir group died 71 days after transplantation of bacterial pneumonia and fungemia. Deaths could not be attributed to CMV in either group.

Four patients in the experimental group had retransplantation for surgical complications. Second transplants were done on days 2, 3, and 10 in three patients because of primary graft failure and on day 93 in the fourth because of a leaking donor bile duct. No graft loss was attributable to CMV disease.

Other Virus Infections

Herpes simplex virus infections (orolabial) were observed in 17% (4 of 23) of the patients in the experimental and in 0% (0 of 24) of the patients in the acyclovir group (17% difference; CI, 2% to 35%; P < 0.05). These infections were observed a mean of 12 days after transplantation (range, 7 to 20 days). All patients with herpes simplex virus infection responded to a short course of oral acyclovir without evidence of dissemination. Localized varicella zoster virus infection in the left S1 dermatomal distribution developed on day 168 in one patient in the experimental group. The patient was treated successfully with intravenous acyclovir. No clinical disease attributable to EBV was observed.

Adverse Effects

Adverse events probably due to acyclovir were observed in 21% (5 of 24) of the patients receiving the drug. Three patients had lethargy and confusion, 1 had worsening renal function, and 1 had leukopenia. Acyclovir was resumed and tolerated in all but one of these patients. Six patients in the experimental group received eight courses of ganciclovir. No untoward toxicity attributable to ganciclovir was observed. In particular, no hematologic toxicity occurred in any of the patients.

Discussion

The optimal approach to antiviral prophylaxis of CMV in transplant recipients remains unclear. Uncertainties and controversies exist for the type of antiviral agent, the route of administration, the timing of prophylaxis, and the kinds of patients to receive the prophylaxis. We did a study in which patients were randomly assigned to standard therapy (high-dose acyclovir) or to an experimental group in which acyclovir prophylaxis was not used. In the experimental group, two approaches were used (see Figure 1). First, for patients who never shed CMV, no antiviral therapy was ever used. Second, for patients with CMV shedding (infection), preemptive ganciclovir therapy was used in an attempt to prevent subsequent CMV disease. The experimental group therefore received antiviral (ganciclovir) prophylaxis against CMV disease but not against CMV infection.

We found that the current standard of high-dose oral acyclovir [8] was ineffective as prophylaxis against CMV infection and disease in liver transplant recipients (Table 2 and Figure 2). With our surveillance culture protocol, no difference in the incidence of detectable shedding before the onset of CMV disease (25% compared with 22%) was noted between the patients receiving high-dose oral acyclovir and no antiviral prophylaxis, respectively. However, our study included only a small number of seronegative recipients of seropositive donor allografts, a group that had the greatest benefit in the study by Balfour and colleagues [8], and our patients were liver rather than kidney transplant recipients. However, since the publication of the report by Balfour and colleagues [8], several retrospective studies [17, 18] have failed to confirm the efficacy of high-dose acyclovir for CMV prophylaxis in renal transplant recipients.

Our study used the concept of preemptive therapy as an approach to prevention of CMV disease. Preemptive therapy has been defined as highly effective therapy administered for a brief period to selected patients who are at highest risk for serious disease [19]. Criteria used to identify high-risk patients may either be a clinical feature (for example, institution of OKT3 therapy [20]) or a laboratory marker (for example, viral excretion [13, 14]). Viral excretion, in particular viremia, has been shown in other studies [15] to be a predictor of CMV disease. We used CMV excretion as a marker to identify patients perceived to be at high risk for CMV disease in an attempt to target those patients most likely to benefit from antiviral prophylaxis (see Figure 1). Two studies in bone marrow transplant recipients have used a similar approach [13, 14]. Schmidt and colleagues [13] started ganciclovir based on detection of CMV in bronchoalveolar lavage fluid at day 35 after transplantation and reported a decrease in CMV pneumonia. Goodrich and colleagues [14] showed that institution of ganciclovir after excretion of CMV in urine, throat, or blood cultures led to a statistically significant decrease of CMV disease in bone marrow transplant recipients (P < 0.05). In both of these studies, a prolonged course of ganciclovir was used (100 days [13] and 120 days [14]). In another study [21] in bone marrow transplant recipients, seropositive patients after engraftment received intravenous ganciclovir or placebo until day 100. Cytomegalovirus disease occurred significantly less frequently in the ganciclovir group (P < 0.05); however, neutropenia developed in 30% of the patients receiving ganciclovir, leading to a statistically significant increased risk for bacterial sepsis. In addition, 40% of the patients were exposed unnecessarily to ganciclovir [21].

Our study differs from the previous studies [13, 14, 21, 22] in that a 7-day pulse course of ganciclovir was used for CMV shedding instead of long-term ganciclovir therapy. Although preemptive therapy with ganciclovir initiated for viral shedding successfully prevented subsequent therapy, viral excretion as a marker for subsequent CMV disease was not wholly adequate. Thus, in 63% (5 of 8) of patients (in 1 patient in the experimental group and in 4 patients in the acyclovir group), CMV disease occurred without previous viral shedding. These data are virtually identical to those from the studies [13, 14] in bone marrow recipients in which CMV disease occurred in 59% to 69% of the patients before or at the time of first viral excretion. Whether more sensitive laboratory markers, for example, CMV antigenemia or more frequent viral cultures (every week instead of every 2 weeks as in this study), would lead to a higher detection of viral excretion remains to be explored in liver transplant recipients.

Another potential concern with prolonged antiviral therapy is the possible emergence of drug-resistant viral strains. Cytomegalovirus strains with decreased in vitro susceptibility to ganciclovir have been recovered, albeit infrequently, from patients receiving long-term ganciclovir [23, 24] and even patients receiving acyclovir [23]. We did not encounter ganciclovir resistance in CMV isolates from our patients receiving acyclovir, although the number of isolates tested was small (n = 5). The demonstration that ganciclovir-resistant CMV may emerge under selective pressure underscores the need for instituting prophylaxis at a time when it would be maximally effective and for defining the duration for which prophylactic regimens should be used. A short course of ganciclovir is less likely to lead to emergence of resistant isolates. Finally, because no antiviral prophylaxis was used in the experimental group, we were able to evaluate the inherent risk for CMV infection (but not disease) in liver transplant recipients receiving tacrolimus; only 26% of patients developed CMV infection (see Table 2).

Our study has limitations that must be acknowledged. First, our sample size of 47 patients was relatively small compared with a few other comparative studies of antiviral prophylaxis in transplant recipients [8, 13, 14]. This small sample resulted in the wide confidence intervals for some of the observed differences. However, because most of the nonsignificant differences we found were in the direction of acyclovir being worse, even with the wide confidence intervals, we were able to rule out the possibility of a substantial acyclovir benefit. Second, we encountered a high rate of CMV seropositivity (91%) in our transplant candidates because of various epidemiologic factors including age and previous transfusion. Consequently, the number of patients in the CMV recipient-negative and the donor-positive group in our study was small. Whereas 2 of 2 recipients of seropositive organs in the acyclovir group developed CMV disease compared with 0 of 2 recipients in the experimental group, the sample size was insufficient to draw any conclusions in this group. Third, the efficacy of ganciclovir could not be assessed in another high-risk group for CMV disease, recipients of OKT3 antibodies. The use of OKT3 has been associated with a high rate of CMV disease in transplant recipients [26]. Various studies of liver transplant recipients have used OKT3 to treat rejection in up to 25% to 40% of patients [12, 27]. Using tacrolimus and low-dose prednisone as primary immunosuppression, OKT3 was used in only 4% of liver transplant recipients in an earlier study from our institution [28] and in 2% (1 of 47) of patients in this study. Given the low requirement of OKT3 in our liver transplant recipients, it is unlikely that a sufficient number of patients will ever be enrolled from this institution to assess the effect of our approach for CMV prophylaxis in recipients of OKT3.

Ganciclovir prophylaxis, however, proved efficacious in another high-risk group for CMV disease in liver transplant recipients, patients having retransplantation. Retransplantation has been shown to increase the risk for CMV disease in liver transplant patients [29]. Four of 23 patients in the ganciclovir group had retransplantation during the study period. Three of these 4 patients developed CMV shedding after the second transplant, and pulse ganciclovir therapy (for viremia in 2, viruria in 1) prevented clinical disease in all 3 patients.

We have shown that early treatment with short course, preemptive ganciclovir therapy in patients with positive surveillance cultures decreased the incidence of CMV disease after liver transplantation. Because viral excretion in our protocol was not 100% predictive of subsequent disease, other markers should be studied. By limiting ganciclovir prophylaxis to a select group, the toxicity and expense associated with prolonged ganciclovir administration can be minimized.