Cytomegalovirus disease has been prevented by prolonged administration (3 to 4 months) of antiviral therapies such as acyclovir [5], anticytomegalovirus hyperimmune globulin [6], or the combination of these two therapies [7]. These prophylactic strategies reduce the attack rate of cytomegalovirus disease, but this benefit is attenuated in patients who receive antilymphocyte antibody therapy [1].

Because the risk for developing cytomegalovirus disease depends on the type of immunosuppression administered after transplantation, we have proposed an alternative approach to preventing cytomegalovirus disease. This approach targets patients at greatest risk for cytomegalovirus disease for treatment with the most potent anticytomegalovirus therapy available. The antiviral therapy is administered when the risk is greatest (for example, during treatment with antilymphocyte antibodies). To distinguish this approach from preventive strategies used in all patients (nontargeted prophylaxis), we introduced the term "preemptive" therapy [8]. Preliminary studies suggested that ganciclovir administered as preemptive therapy to patients receiving antilymphocyte antibodies might reduce the attack rate of cytomegalovirus disease in transplant recipients who are positive for cytomegalovirus antibody. This multiinstitutional, randomized clinical trial was designed to assess the efficacy and safety of preemptive ganciclovir therapy in preventing cytomegalovirus disease in transplant recipients who are positive for cytomegalovirus antibody and who are receiving antilymphocyte antibodies.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

Consecutive renal transplant recipients who tested positive for cytomegalovirus antibody and who received a kidney from cytomegalovirus antibody-positive or antibody-negative donors at Albany Medical Center, Emory University Hospital, Massachusetts General Hospital, New England Medical Center, University of Chicago Medical Center, and Yale-New Haven Hospital were eligible for enrollment in the study. Patients were randomly assigned to receive either preemptive ganciclovir or no ganciclovir every day that antilymphocyte antibody therapy (muromonab-CD3 [OKT3], antithymocyte globulin, or antilymphocyte globulin) was administered. Separate randomization lists were used in each transplantation center; within each center, separate randomization lists were used for cadaveric and living, related donors. The investigators at each site knew which patients received the study drug and which patients received no anticytomegalovirus therapy. The cytomegalovirus antibody status of donors and recipients was determined using enzyme immunoassay on recipient serum obtained before transplantation and on donor serum obtained before transfusion.

Patients were excluded from the study if they were younger than 20 years of age, were pregnant, had received another organ in addition to a kidney, had received any anticytomegalovirus therapy (defined as more than 1.2 g of acyclovir per day, unselected immune globulin, or cytomegalovirus hyperimmune globulin), or refused to give consent. The committees on human experimentation at each institution approved the study.

All study participants were observed for the following outcomes during the 6 months after they received antilymphocyte antibodies: 1) cytomegalovirus disease; 2) other infectious diseases; 3) and noninfectious diagnoses. Allograft function 6 months after administration of antilymphocyte antibody was recorded as either present (for those not requiring dialysis) or absent (for those receiving dialysis). Serum creatinine levels were recorded for all patients with functioning allografts 6 months after antilymphocyte antibody therapy. Patients were evaluated for cytomegalovirus viremia and disease in three ways. First, the virology laboratory at each program tried to isolate cytomegalovirus from buffy-coat specimens at least every month using either centrifugation culture [9], conventional culture techniques [10], or both. Second, buffy-coat specimens were cultured for cytomegalovirus when any of the following signs, symptoms, or laboratory abnormalities were present: temperature greater than 38 °C, leukocyte count less than 3.0 x 10⁹ cells/L, dyspnea, abdominal pain, or gastrointestinal bleeding. Third, a biopsy specimen was obtained from any abnormal site, and all biopsy specimens were examined histologically for the presence of characteristic cytomegalovirus inclusion bodies and uptake of immunofluorescent-labeled anticytomegalovirus antibodies. Patients with other signs or symptoms suggesting infection (such as cough, headache, diarrhea, allograft discomfort, or dysuria) were evaluated using standard diagnostic protocols. At all study sites, the diagnostic protocol included at least a chest radiograph, leukocyte count, renal and liver function tests, two sets of blood cultures, urinalysis, urine culture, and cultures from other potential sites of infection. For patients who did not survive the 6-month observation period, data on the cause and date of death were collected.

Definition of Cytomegalovirus Disease

Cytomegalovirus disease was defined as either the cytomegalovirus syndrome or tissue-invasive disease developing within 6 months of antilymphocyte antibody therapy or within 1 month of discontinuing immunosuppression after allograft loss. The cytomegalovirus syndrome was diagnosed when both virologic and clinical criteria were met within a 7-day period. Virologic criteria were fulfilled when cytomegalovirus was isolated from a buffy-coat specimen or bronchoalveolar fluid. Clinical criteria were fulfilled when patients had a temperature higher than 38 °C [without antipyretic agents] for 3 or more consecutive days and within 7 days of two or more of the following: 1) leukopenia [leukocyte count <3.0 x 10⁹ cells/L on two consecutive measurements after stopping azathioprine therapy]; 2) hepatitis [serum alanine aminotransferase >1.5 times the upper limit of normal, without serologic evidence of active hepatitis B or hepatitis C virus]; 3) atypical lymphocytosis [more than 20% of leukocytes]; and 4) pneumonitis (an abnormal result on chest radiograph and no alternative explanation [including absence of Pneumocystis carinii in respiratory secretions]). These criteria were modified slightly from those used by other researchers [6] because isolation of cytomegalovirus from buffy-coat specimens or bronchoalveolar fluid predicts presence of cytomegalovirus disease, whereas isolation of cytomegalovirus from urine and saliva may not [11].

Tissue-invasive cytomegalovirus disease was diagnosed histopathologically by showing the presence of inclusion bodies characteristic of cytomegalovirus or by an immunochemical stain positive for cytomegalovirus antigens in a biopsy specimen from a lesion or an abnormal site (gastrointestinal tract, lung, or liver). The investigator at each study site made decisions about treating cytomegalovirus disease.

Study Drug Administration

The study drug was given daily (or according to the schedule in Table 1 only when the patient already had intravenous access for administration of antilymphocyte antibodies. The study medication was started within 24 hours of the first dose of each course of antilymphocyte antibodies. Patients receiving the study drug were given ganciclovir infusions based on daily serum creatinine concentrations (Table 1).

Leukocyte count, platelet count, and serum creatinine concentration were monitored daily during antilymphocyte antibody therapy in patients who received preemptive ganciclovir therapy and in controls.

Sample Size and Statistical Analysis

In the primary analysis, the proportion of patients with cytomegalovirus disease in the two study groups were compared. On the basis of our previous study [4], we assumed that cytomegalovirus disease would develop in 60% of controls. We wanted to enroll 48 patients who could be evaluated in each group (for a total of 96) to detect a decrease in the proportion of patients with cytomegalovirus disease from 60% (in controls) to 30% or less (in patients receiving preemptive ganciclovir therapy); that is, we predicted that preemptive ganciclovir treatment would yield a 50% lower attack rate (using a two-sided test, = 0.05 and power = 0.8). Initial analyses were done on an intention-to-treat basis and included all eligible patients.

We defined the primary outcome and potential predictors of cytomegalovirus disease before beginning the study. At the end of therapy and follow-up, the study nurse purged all study site and study drug information from copies of case report forms. One author evaluated the purged case report forms to determine whether each patient met the criteria for cytomegalovirus disease. The test of reliability was used to assess agreement between the diagnosis of cytomegalovirus disease reported by the investigator on the case report form and the determination made by blinded assessment. Potential predictors of cytomegalovirus disease included age (<45 years or 45 years or older), sex, study institution, renal disease (hypertension, diabetes mellitus, glomerular diseases, and other renal diseases), previous renal transplantation (none or one or more), type of donor (cadaver or living relative), donor cytomegalovirus antibody status (positive or negative), type of antilymphocyte antibody therapy (muromonab-CD3, antilymphocyte globulin, antithymocyte globulin), amount of antilymphocyte antibody therapy (high or low dose), and reason for receiving treatment with antilymphocyte antibodies (to induce or to treat steroid-resistant rejection). We defined high-dose antilymphocyte antibody therapy as follows: greater than 70 mg of muromonab-CD3, greater than 10 g of antithymocyte globulin, or greater than 10 g of antilymphocyte globulin, or more than one course of antilymphocyte antibody therapy. We defined low-dose antilymphocyte antibody therapy as 70 mg or less of muromonab-CD3, 10 g or less of antithymocyte globulin, or 10 g or less of antilymphocyte globulin.

Analyses were done using statistical software (BMDP Statistical Software, Los Angeles, California). Baseline characteristics and immunosuppression were compared using either the chi-square test, the Fisher exact test, unpaired two-sample t-test, or the Wilcoxon-Mann-Whitney test. The proportion of patients with cytomegalovirus disease in the two groups was compared using the chi-square test. Serum creatinine concentration during antilymphocyte antibody therapy was compared using analysis of variance with repeated measures (Program 2V). Cytomegalovirus-free survival time was estimated using the product-limit method of Kaplan and Meier [12]. The two curves were compared using the Mantel-Cox log-rank test [13]. Censored events included loss to follow-up, deaths in patients in whom cytomegalovirus disease was not diagnosed, and completion of the observation period without diagnosis of cytomegalovirus disease. The Cox regression model was used in multivariate comparisons [14]. Relative risk was estimated using the Cox model. All probability values were two-tailed. The financial supporters of the study had no role in protocol design, data gathering, analysis or interpretation of the results, or manuscript preparation.

Results

Top Methods Results Discussion Author & Article Info References

Between November 1990 and September 1992, 121 consecutive renal transplant recipients who tested positive for cytomegalovirus antibody in six centers were eligible to be enrolled in the study (Figure 1). Eight patients were not enrolled (3 refused to participate, 1 was unavailable for follow-up [foreign national], and 4 were not recognized as eligible for the study). One hundred thirteen patients participated (64 were randomly assigned to receive preemptive ganciclovir therapy and 49 were randomly assigned to the control group). By chance, some imbalance in the number of allocations made to the two treatment groups occurred, because the unconstrained randomization for one center included two unbroken runs of six and four consecutive assignments to the preemptive ganciclovir group. The characteristics of the two groups before treatment are shown in Table 2. Cytomegalovirus disease was not diagnosed in any study patient between the date of transplantation and the date of enrollment in the study. The groups were similar with respect to age, sex, previous renal transplantations, renal disease, type of donor (cadaver or living relative), and the cytomegalovirus antibody status of the donor (positive or negative). Forty-one patients were enrolled at Albany Medical Center, 36 at University of Chicago Medical Center, 16 at Emory Medical Center, 13 at Massachusetts General Hospital, 4 at New England Medical Center, and 3 at Yale-New Haven Hospital. The patients enrolled at the various sites had similar baseline characteristics. All 113 enrolled patients were used to calculate the proportion of patients with cytomegalovirus disease in the preemptive ganciclovir and control groups, including those with protocol deviations or those who died before the study was complete. The 8 eligible patients who were not enrolled in the study were excluded from analyses because they were not followed using the study protocol.

View larger version (16K): [in this window] [in a new window]   Figure 1. Renal transplant recipients in six transplantation centers.

All transplantation centers used similar regimens of prednisone, cyclosporine, azathioprine, and methylprednisolone for immunosuppression. Four sites used antilymphocyte globulin to induce rejection and muromonab-CD3 to treat rejection, whereas two sites used muromonab-CD3 to both induce and treat rejection. Although some site-to-site variation in use of antilymphocyte antibodies occurred, this did not cause an imbalance between the treatment groups. As shown in Table 3, we found no statistical difference between the treatment groups in the proportion of patients receiving antilymphocyte antibody to induce or treat rejection or between the proportion of patients receiving muromonab-CD3, antilymphocyte globulin, or antithymocyte globulin. We found no statistical difference in the duration of antilymphocyte antibody treatment in the two groups. No study participants received unselected immune globulin or anticytomegalovirus hyperimmune globulin before diagnosis of cytomegalovirus disease. One patient in each group was treated with a prolonged course of acyclovir (<1.2 g/d), and seven patients (four controls and three patients given preemptive ganciclovir) received acyclovir (<1.2 g/d) to treat symptomatic herpes simplex virus infections. No study patient received ganciclovir other than as preemptive therapy or to treat cytomegalovirus disease.

The outcome of all patients enrolled in the study is shown in Table 2. Fifty-five patients (86%) in the preemptive ganciclovir group completed the protocol. Minor protocol deviations occurred in 12 of 55 patients receiving preemptive ganciclovir therapy. These included 3- to 7-day delays in initiating preemptive ganciclovir therapy after starting the course of antilymphocyte antibodies (5 patients); improper dosage adjustment in renal insufficiency (2 patients); doses not administered (2 patients each missed one dose); omission of preemptive ganciclovir therapy during one of multiple courses of antilymphocyte antibody therapy (2 patients); and failure to administer preemptive ganciclovir therapy (1 patient). Eighty-four percent of the controls completed the protocol without incident; however, two pediatric patients aged 14 and 18 years were entered inappropriately. Premature termination occurred in 14% of the patients receiving preemptive ganciclovir therapy and in 16% of controls. No adverse events were associated with drug administration. Three patients died during the study. Two of the three deaths occurred in the control group 37 and 64 days after antilymphocyte antibody therapy was complete. Both deaths were attributed to sepsis and multiple-organ system failure. Cytomegalovirus was diagnosed in one of these two patients 1 month before death, whereas the other patient did not have cytomegalovirus disease before death. The third death occurred in the preemptive ganciclovir group 73 days after the completion of antilymphocyte antibody therapy and 28 days after treatment for cytomegalovirus disease. This death was attributed to myocardial infarction. Transplantations were unsuccessful in six patients in each group, all of whom stopped immunosuppression after their return to dialysis. Two patients in the preemptive ganciclovir group were lost to follow-up 32 and 78 days after completing antilymphocyte antibody therapy. Cytomegalovirus disease was not diagnosed in either patient before they were lost to follow-up.

Cytomegalovirus disease was diagnosed in 25 patients during the study. Disagreements about the diagnosis of cytomegalovirus disease occurred for only 1 patient, but review of the original medical record resolved the discrepancy (, 0.97; 95% CI, 0.92 to 1.00).

Effect of Preemptive Ganciclovir Therapy on Cytomegalovirus Disease

Table 4 shows the proportion of patients in the preemptive ganciclovir and control groups in whom cytomegalovirus disease was diagnosed within 6 months of treatment with antilymphocyte antibodies. The attack rate for cytomegalovirus disease in the control group was 2.33 times higher than in the preemptive ganciclovir group (P = 0.018; unadjusted relative risk, 0.43; CI, 0.21 to 0.89). Tissue-invasive cytomegalovirus disease was diagnosed in 5% of the patients receiving preemptive ganciclovir and in 8% of controls (P = 0.46) (Table 4). The attack rate for cytomegalovirus disease after induction was 10% in patients receiving preemptive ganciclovir therapy and 24% in the controls. The attack rate for cytomegalovirus disease after treatment of rejection was 22% in those receiving preemptive ganciclovir therapy and 64% in the controls (Table 4). Although the proportion of patients with cytomegalovirus disease varied among institutions, the variation was explained by the reason for administration of antilymphocyte antibodies, that is, whether such therapy was used predominantly to induce or predominantly to treat rejection. As Table 3 shows, patients were similarly distributed in the two groups regarding the reason for use of immunosuppression.

Figure 2 shows the Kaplan-Meier estimates of the proportion of patients without cytomegalovirus disease in the two treatment groups during the 6 months after treatment with antilymphocyte antibodies (P = 0.008 by the log-rank test) and during the 9 months after transplantation (P = 0.007 by the log-rank test). After controlling for the use of preemptive ganciclovir and the reason for using antilymphocyte antibodies (to induce or treat rejection) in a Cox proportional-hazards model, we found that preemptive ganciclovir therapy protected against cytomegalovirus disease (adjusted relative risk, 0.27; CI, 0.12 to 0.64), whereas the administration of antilymphocyte antibodies for treatment of rejection (as opposed to induction of rejection) increased the risk for the development of cytomegalovirus disease (adjusted relative risk, 3.03; CI, 1.35 to 6.78). The following potential predictors of cytomegalovirus disease were not confounders in forced-entry multivariate Cox regression and were not included in the final model: age, sex, study site, renal disease, previous renal transplantation, type of donor, donor cytomegalovirus antibody status, type of antilymphocyte antibody therapy (muromonab-CD3, antilymphocyte globulin, or antithymocyte globulin), and amount of antilymphocyte antibody therapy (high-dose or low-dose). No significant interactions were detected.

View larger version (19K): [in this window] [in a new window]   Figure 2. Kaplan-Meier product-limit estimates of the probability of remaining free of cytomegalovirus disease during the 6 months after antilymphocyte antibody therapy and during the 9 months after transplantation. The probability that cytomegalovirus disease will develop after antilymphocyte antibody therapy was different in the two groups (P = 0.008 by the log-rank test). The probability that cytomegalovirus disease will develop after transplantation differed significantly between the two groups (P = 0.007 by the log-rank test). Patients received preemptive ganciclovir therapy during each course of antilymphocyte antibody therapy, whereas controls received no specific anticytomegalovirus therapy during any course of antilymphocyte antibody therapy.

Cytomegalovirus disease was diagnosed a median of 46 days (range, 22 to 62 days) after completion of antilymphocyte antibody therapy in the patients receiving preemptive ganciclovir therapy and a median of 31 days [range, 3 to 68 days] in the controls (Figure 3); P =0.27 by the Wilcoxon-Mann-Whitney test). Cytomegalovirus disease was diagnosed a median of 61 days (range, 35 to 108 days) after transplantation in the patients receiving preemptive ganciclovir therapy and a median of 45 days (range, 39 to 92 days) after transplantation in the controls (Figure 3).

View larger version (17K): [in this window] [in a new window]   Figure 3. Cumulative frequency and timing of cytomegalovirus disease and viremia in the two study groups. In the preemptive ganciclovir group, cytomegalovirus disease was diagnosed in nine patients a median of 46 days after they received antilymphocyte antibodies. In the control group, cytomegalovirus disease was diagnosed in 16 patients a median of 31 days after they received antilymphocyte antibodies. Patients received preemptive ganciclovir during each course of antilymphocyte antibody therapy, whereas controls received no specific anticytomegalovirus therapy during any course of antilymphocyte antibody therapy.

Preemptive ganciclovir therapy reduced the occurrence of viremia from 35% (17 of 49 controls) to 17% (11 of 64 recipients of the study drug) during the 6 months after antilymphocyte antibody therapy (P = 0.03; unadjusted relative risk, 0.5; CI, 0.26 to 0.96). Viremia occurred a median of 52 days (range, 22 to 124 days) after administration of preemptive ganciclovir and a median of 31 days [range, 3 to 90 days] after administration of antilymphocyte antibodies to controls (Figure 3); P = 0.18 by the Wilcoxon-Mann-Whitney test). All patients were infected with the virus 0 to 7 days before cytomegalovirus disease was diagnosed.

Cytomegalovirus disease was diagnosed in 9 of 64 patients receiving preemptive ganciclovir therapy, and 8 of these 9 patients were treated with antiviral therapy. All 8 responded to therapy (6 were treated with ganciclovir alone, and 2 were treated with ganciclovir and anticytomegalovirus hyperimmune globulin). Cytomegalovirus disease resolved without treatment in 1 patient receiving preemptive ganciclovir therapy. Cytomegalovirus disease was diagnosed in 16 of 49 control patients. Thirteen of these 16 were treated and responded to therapy: 6 received ganciclovir alone, 5 received ganciclovir and immune globulin, and 2 received immune globulin. Cytomegalovirus disease resolved without treatment in 3 controls, 2 of whom were receiving hemodialysis therapy (but not immunosuppression) for 1 and 11 days, respectively, after allograft loss when cytomegalovirus disease was diagnosed. None of the untreated patients had tissue-invasive cytomegalovirus disease. No study patient had more than one episode of cytomegalovirus disease.

Toxicity of Preemptive Ganciclovir Therapy

Patients assigned to the preemptive ganciclovir group received the study drug for a median of 9 days (range, 0 to 16 days) during antilymphocyte antibody therapy, which was administered for a mean of 11.2 days (SE, 0.46). Controls received antilymphocyte antibodies for a mean of 11 days (SE, 0.56). The mean total dose of preemptive ganciclovir administered to the study patients was 21.2 mg/kg (range, 0 to 42.9 mg/kg). No adverse events were attributed to the study drug during or in the 6 months after administration of preemptive ganciclovir therapy. All study and control patients had leukocyte counts greater than 3.1 x 10⁹/L during the monitoring period. No patient had platelet counts less than 50 000 x 10 (⁹/L) for more than 1 day during the monitoring period. The dose of preemptive ganciclovir therapy had to be adjusted in 49 (77%) patients.

Other Events and Graft Outcome 6 Months after Antilymphocyte Antibody Therapy

The groups had a similar incidence of herpes simplex virus infection, bacterial infections, and fungal infections during the 6 months after treatment with antilymphocyte antibodies. Six months after antilymphocyte antibody therapy, patients with functioning allografts who had received preemptive ganciclovir had a lower median serum creatinine concentration (1.55 mg/dL) than did controls (1.86 mg/dL) (P = 0.028 by the Wilcoxon-Mann-Whitney test).

Discussion

Top Methods Results Discussion Author & Article Info References

Low-dose preemptive ganciclovir therapy, 2.5 mg/kg per day, administered intravenously to renal transplant recipients who were positive for the cytomegalovirus antibody only when they were receiving antilymphocyte antibody therapy decreased the incidence of cytomegalovirus disease by 52%. This protective effect occurred whether antilymphocyte therapy was administered to induce or to treat rejection. In addition, patients receiving preemptive ganciclovir therapy had improved secondary outcomes associated with cytomegalovirus infection: These included a lower incidence of cytomegalovirus viremia and better allograft function 6 months after receiving antilymphocyte antibody therapy. These beneficial effects were seen without either evidence of toxic side effects associated with preemptive ganciclovir therapy or the emergence of cytomegalovirus strains resistant to ganciclovir. Other investigators noted that cytomegalovirus disease resistant to ganciclovir therapy did not develop in solid-organ transplant recipients after previous courses of prophylactic ganciclovir therapy [15, 16]. Cytomegalovirus strains with decreased susceptibility to ganciclovir in vitro have been isolated in other immunocompromised patients after treatment with prolonged courses of ganciclovir [17, 18].

We observed a striking difference in the incidence of cytomegalovirus disease in renal transplant recipients who were positive for cytomegalovirus antibody after antilymphocyte antibody therapy to induce rejection compared with those who received such therapy to treat rejection. This observation is important for several reasons. First, observations from a single institution that showed that transplant recipients who were positive for cytomegalovirus antibody and who were treated with antilymphocyte therapy for rejection were particularly susceptible to the development of cytomegalovirus disease may be generalized to all transplant recipients [4]. These patients, who account for 10% to 20% of all renal transplant recipients, are ideally suited to a targeted (preemptive) approach to prevent cytomegalovirus disease. Second, the increased attack rate of cytomegalovirus disease occurred regardless of the type of antilymphocyte antibody administered. We might expect attack rates of cytomegalovirus disease to differ for the monoclonal and polyclonal antilymphocyte antibody preparations that lyse different populations of T lymphocytes because MHC-restricted, cytomegalovirus-specific cytotoxic T cells provide the critical defense against replicating viruses. However, recent studies in the murine model [19] and renal transplant recipients [20] consistently indicate that attack rates of cytomegalovirus disease are similar regardless of the type of antilymphocyte antibody used to treat rejection. After induction, the attack rate of cytomegalovirus disease has been reported to be independent of type of antilymphocyte antibody in some renal transplant centers [21, 22] and to depend on the type of antilymphocyte antibody in other renal transplant centers [20, 23-25]. This heterogeneity of results from clinical trials designed to compare rejection rates after induction with antilymphocyte antibodies may be due to differences in detection and diagnosis of cytomegalovirus disease as a secondary outcome. Overall, most of these studies suggest that the increase in the attack rate of cytomegalovirus disease occurs after administration of all types of antilymphocyte antibodies. The results of this study should be applicable to all transplantation centers that use antilymphocyte antibodies. Third, the higher attack rate of cytomegalovirus disease after use of antilymphocyte antibodies to treat rejection is consistent with recent results linking reactivation of cytomegalovirus to enhanced plasma levels of tumor necrosis factor [26]. Antilymphocyte antibody treatment of rejection is associated with a profound, clinically symptomatic cytokine release in excess of the cytokine release observed after induction of rejection. Preemptive ganciclovir may limit the extent of viral replication as cytomegalovirus first emerges from latency, in association with tumor necrosis factor and other cytokines elaborated during antirejection therapy.

Commenting on the cost-effectiveness of preemptive ganciclovir in renal transplantation would be premature, but the additional cost of administering 14 days of preemptive ganciclovir therapy to a 70-kg patient during antilymphocyte antibody therapy is approximately $525, not including costs of drug administration. Intravenous access and frequent measurement of serum creatinine concentration and leukocyte and platelet counts are required to monitor the safety and efficacy of antilymphocyte antibody therapy. Addition of preemptive ganciclovir requires no additional costs associated with intravenous access and laboratory tests.

Preventing cytomegalovirus disease by starting a short course of preemptive ganciclovir whenever antilymphocyte antibodies are administered must be compared with other approaches. Recently, Singh and colleagues [16] reported that the administration of ganciclovir to liver transplant recipients in whom viremia was detected before the onset of symptoms reduced the attack rate of cytomegalovirus disease from 29% to 4%. Winston and associates [27] reported that ganciclovir administered continually for the first 100 days after liver transplantation reduced the incidence of cytomegalovirus disease from 10% to 0.8% [28]. Both approaches are costly; the first requires frequent laboratory testing for the presence of cytomegalovirus, and the second requires prolonged maintenance of intravenous access for drug delivery.

One limitation of our study was that it was not powered to evaluate the efficacy of preemptive ganciclovir to prevent tissue-invasive cytomegalovirus disease. Although the attack rate of invasive disease was the same in the two treatment groups, one of the three preemptive ganciclovir recipients with invasive disease received a second course of antilymphocyte therapy to treat rejection without the assigned preemptive ganciclovir (due to a protocol violation), whereas a second recipient of preemptive ganciclovir received more muromonab-CD3 (a larger total dose administered for more days) than did any of the four controls with tissue-invasive disease. We recommend that preemptive ganciclovir be administered with each course of antilymphocyte antibody therapy, particularly second and third courses that are usually administered to treat rejection. A second limitation of our study was the early loss to follow-up of two patients who were receiving preemptive ganciclovir. Because both patients could have developed cytomegalovirus disease after being lost to follow-up, a "worst-case scenario" was constructed in which both of these patients were classified as having cytomegalovirus disease. The "worst-case scenario" attack rate for cytomegalovirus disease remained higher in the controls than in those receiving preemptive ganciclovir (P = 0.05; unadjusted relative risk, 0.53 [CI, 0.27 to 1.00]). A third limitation of this study may be the dose of preemptive ganciclovir administered. Because we observed no toxicity with the preemptive ganciclovir regimen, higher doses of ganciclovir may be more effective, particularly in the prevention of invasive disease.

How, then, should the preemptive therapy approach be used? Perhaps the most important conclusion we can draw from this study is that an antiviral therapy to prevent cytomegalovirus disease can be administered precisely when the risk increases for its development [that is, during administration of specific types of immunosuppressive therapy] but not without regard for risk or during administration of any immunosuppressive therapy. This strategy allows antimicrobial therapy to reduce the risk for infection that depends on the type and intensity of immunosuppression. The efficacy of preemptive ganciclovir to prevent cytomegalovirus disease may be improved by any of the following strategies: 1) by increasing the dose of ganciclovir (we recently showed that 5 mg/kg of ganciclovir per day may be administered safely with antilymphocyte antibodies in lung transplant recipients) [21]; 2) by combining preemptive ganciclovir with anticytomegalovirus hyperimmune globulin; or 3) by adding preemptive ganciclovir to prophylactic regimens to prevent cytomegalovirus disease, such as prolonged administration of oral acyclovir and cytomegalovirus hyperimmune globulin, only when these patients receive antilymphocyte antibodies. We have shown that targeted preventive therapy can be implemented with clinical benefit if and when intensive immunosuppression is needed to save an allograft.

Presented in part at the Twelfth Annual Meeting of the American Society of Transplant Physicians, Houston, Texas, May 1993, and at the Thirty-Third Interscience Conference on Antimicrobial Agents and Chemotherapy, New Orleans, Louisiana, October 1993.