Ganciclovir Prophylaxis To Prevent Cytomegalovirus Disease after Allogeneic Marrow Transplant

Methods

Patients

Patients who were CMV seropositive before transplant and were to receive an allogeneic marrow transplant for hematologic malignancy requiring total body irradiation or busulfan-cyclophosphamide at the Fred Hutchinson Cancer Research Center were eligible for study entry. Signed informed consent was obtained according to Food and Drug Administration and institutional guidelines.

Protocol Design

Patients were entered into this double-blind study before transplant and were followed with weekly viral cultures. Patients were assigned randomly to receive ganciclovir or placebo at the time of engraftment, which was defined as an absolute neutrophil count of 0.750 × 10⁻⁹/L or greater for 2 consecutive days. Ganciclovir was given at a dose of 5 mg/kg body weight, administered intravenously twice daily for 5 days and then once daily until day 100 after transplant. Patients were excluded from the study if they received a T-cell-depleted transplant, were younger than 2 years of age, had a serum creatinine level greater than 220 µmol/L, were allergic to acyclovir or ganciclovir, had received a marrow transplant in the previous 6 months, or had documented CMV excretion before randomization. Patients received high-dose acyclovir prophylaxis from the time of conditioning until engraftment [12]. Virus cultures were obtained from urine, blood, and throat on a weekly basis through day 100 or until leaving Seattle, whichever was longer. All patients who began to excrete virus were removed from the study and treated with ganciclovir [11].

Viral infection was defined as recovery of virus from the throat, urine, or blood. Disease caused by CMV was defined as recovery of virus from a visceral site (lung, gastrointestinal tissue) or by bronchoalveolar lavage in patients who had associated signs (pulmonary infiltrate) and symptoms consistent with CMV infection. Marrow toxicity was defined as an absolute neutrophil count of less than 0.750 × 10⁻⁹/L for 2 consecutive days. Bacterial or fungal infection was defined as the recovery of the organism from a blood culture or other normally sterile site.

Laboratory Procedures

Virus cultures included both centrifugation culture [13] and conventional culture. Conventional cultures were maintained for 5 weeks.

Statistical Design

Study end points were the development of CMV infection and marrow toxicity. Development of CMV disease and death were secondary end points. The analysis of infection, disease, toxicity, and mortality was evaluated from start of the study drug to day 100 after transplant. Mortality and disease were analyzed at 180 days after transplant. Comparison of time to specific events was done according to the method of Kaplan-Meier [14] with analysis by log-rank test [15]. Other comparisons were done using the chi-square test, Fisher exact test, Student t-test, or the Wilcoxon test, as appropriate.

Fifty patients were entered into the study at which point a scheduled interim analysis was performed. A difference in the incidence of CMV infection between study arms at P = 0.005 was required to stop the study. At the completion of the interim analysis, a statistically significant difference between the study arms for CMV infection was found at the P = 0.005 level, and the study was terminated. Seventy patients had been enrolled in the study by the time the interim analysis had been completed. Sixty-four patients were included in the final analysis.

Results

From November 1990 to August 1991, 114 consecutive CMV-seropositive patients were eligible for the study. Twenty-one patients were not enrolled, with three refusing study entry and three receiving T-cell-depleted marrow transplants. The other 15 patients were enrolled in protocols that precluded ganciclovir administration or conflicted with the blinded design of our study.

Ninety-three patients were entered into the study before transplant. Between transplant and engraftment, 23 patients became ineligible for study before randomization with acute renal failure (eight patients), hematologic relapse (eight patients), refusal (three patients), failure to engraft (two patients), and positive culture for CMV (two patients). Seventy patients were randomized at engraftment and received the study drug. When the interim analysis showed a difference between the ganciclovir and placebo arms, the study was stopped. Five of the 70 patients had not reached an end point when the result of the interim analysis became available. All five of these patients had been enrolled in the study less than 2 weeks and are not included in the final analysis but were treated with ganciclovir during the remainder of their transplant course. None of these five patients developed CMV excretion or disease during the first 100 days after transplant. One additional patient was excreting CMV the day before study entry and received two doses of study drug before culture results became available. This patient was withdrawn from the study and was treated with ganciclovir. This left 64 evaluable patients with 33 receiving ganciclovir and 31 receiving placebo (Table 1). There were no statistical differences between groups except for the day after transplant for study withdrawal and for total days on study. The median day for study withdrawal in the ganciclovir group was 88 days (range, 44 to 114 days after transplant) compared with a median of 59 days (range, 24 to 100 days after transplant) in the placebo group (P = 0.001). This difference was reflected in the total number of days on study with a difference in medians of 25 days between the two groups (P < 0.001).

Effect of Prophylactic Ganciclovir on Cytomegalovirus and Other Herpesvirus Infection

Ganciclovir prophylaxis resulted in a marked reduction in CMV infection (Figure 1). Fourteen of 31 (45%) patients on placebo excreted CMV from the start of the study drug to day 100 after transplant compared with 1 of 33 (3%) ganciclovir recipients (P < 0.001). Cytomegalovirus was recovered from urine in eight patients, blood in four patients, and throat cultures in two patients. Four of the 14 patients on placebo who excreted virus at some time after the start of study drug to day 100 after transplant developed CMV gastrointestinal disease. One patient developed CMV interstitial pneumonia 9 days after gastrointestinal disease was diagnosed. Five patients on placebo who developed CMV disease did not have detectable virus excretion before the onset of disease. None of the patients on ganciclovir excreted virus while on the study drug. One ganciclovir recipient who received 4 days of drug had virus isolated from the blood 36 days after stopping drug secondary to neutropenia. Another ganciclovir recipient had virus isolated from the blood on day 104, 40 days after termination from the study for neutropenia. Neither patient developed CMV disease.

Ganciclovir prophylaxis was also effective in suppressing herpes simplex virus excretion. Fourteen of 31 placebo recipients excreted herpes simplex virus while on study, compared with zero of 33 ganciclovir recipients. One placebo recipient had a case of varicella-zoster infection while on study.

Effect of Prophylactic Ganciclovir on Cytomegalovirus Disease

Ganciclovir prophylaxis was effective in eliminating CMV disease during the first 100 days after marrow transplant. Nine (29%) placebo recipients developed CMV disease from the start of study drug to day 100 after transplant, compared with no cases of disease in the ganciclovir group (P < 0.001). Three patients had CMV interstitial pneumonia, five had gastrointestinal disease, and one patient had pneumonia and gastrointestinal disease (Table 2). The median day to onset of CMV disease was 47 days after transplant (range, 43 to 74 days). Cytomegalovirus interstitial pneumonia was diagnosed earlier than gastrointestinal disease, with medians of 44 (range, 43 to 47 days) and 60 days (range, 45 to 74 days), respectively, consistent with previously published data [1]. Two of the nine patients died of CMV interstitial pneumonia.

When analyzed at 180 days after transplant, 10 (32%) patients on placebo had CMV disease compared with three (9%) ganciclovir recipients (P = 0.015) (Table 2). The four additional patients with CMV disease after day 100 included two patients with CMV interstitial pneumonia and one patient with gastrointestinal disease in the ganciclovir group and one patient with CMV pneumonia in the placebo group. The patient on placebo with CMV interstitial pneumonia diagnosed 128 days after transplant later contracted gastrointestinal disease 143 days after transplant and died of CMV pneumonia 226 days after transplant.

Mortality

The effect of ganciclovir prophylaxis on survival after allogeneic marrow transplant was analyzed from the start of study drug to both 100 and 180 days after transplant. No statistically significant difference in survival between the placebo and ganciclovir groups was found when analyzed to day 100. Six (19%) patients in the placebo group died and four (12%) patients in the ganciclovir group died (P > 0.2). Only one case of fatal CMV disease occurred in the placebo group in the first 100 days of transplant, compared with none in the ganciclovir group. The other nine deaths during the first 100 days after transplant, in both the placebo and ganciclovir groups were caused by bacterial or fungal infection, transplant-related mortality, or relapse (Table 3). When analyzed to day 180 after transplant, the mortality was comparable between the two groups (eight [26%] deaths in the placebo group compared with 10 [30%] deaths in the ganciclovir group) (P > 0.2). Two additional patients died of CMV pneumonia after day 100 after transplant (one each in the placebo and ganciclovir groups). The other deaths were caused by either relapse or infection (see Table 3).

Toxicity

Neutropenia was the only toxicity noted during the study. Ten of 33 (30%) patients on ganciclovir met the criteria for drug-related neutropenia compared with none in the placebo group during the period of observation (P = 0.001). The toxicity from ganciclovir occurred at a median of 66 days after transplant (range, 44 to 95 days) and at a median of 36 days after starting the drug (range, 6 to 74 days). The median duration of neutropenia was 12 days (range, 4 to 20 days). Six of 10 patients recovered their neutrophil count spontaneously to greater than 0.750 × 10⁻⁹/L after the withdrawal of ganciclovir. One patient died on the sixth day of neutropenia with Legionella pneumonia. Three patients received a course of granulocyte-macrophage colony-stimulating factor with a median duration of neutropenia of 12 days (range, 9 to 13 days). In addition to the 10 patients who met study criteria for discontinuation of the drug due to neutropenia, four patients had the drug discontinued (one patient in the placebo group and three patients in the ganciclovir group) because of concerns by the clinician or patient that the absolute neutrophil count was dropping secondary to ganciclovir. None of these patients met study criteria for drug-induced neutropenia, and none had adverse events.

Bacterial and Fungal Infection

Of the 10 patients who became neutropenic during ganciclovir treatment, four had bacterial infections, including three patients with bacteremia and one with pneumonia during the period of neutropenia. The bacteremias were caused by Streptococcus sanguis II, Moroxella (unspeciated), and Klebsiella pneumoniae. All three patients recovered. One patient was diagnosed by bronchoalveolar lavage with Legionella pneumoniae infection and died on the sixth day of neutropenia.

Additional analyses were done to determine if patients on ganciclovir were at increased risk for infection. Thirteen of 33 (45%) ganciclovir recipients had either bacterial or fungal infections during the study compared with 9 of 31 (29%) placebo recipients (P > 0.2). When infection was analyzed using neutropenia as a time-dependent variable, patients who received ganciclovir and became neutropenic had an increased risk for bacterial infection (relative risk, 4.3; P = 0.02).

Five of 10 (50%) ganciclovir recipients who died had bacterial or fungal infections as the cause of death, compared with two of eight (25%) patients in the placebo group (Table 3) (P > 0.2).

Discussion

Previous studies have shown that patients who excrete CMV are at increased risk for subsequent disease [2, 11]. Two studies have now shown that administering ganciclovir to patients who are excreting CMV will prevent CMV disease. Schmidt and colleagues [10] performed bronchoalveolar lavage on all CMV-seropositive patients on day 35 after transplant and randomly assigned patients with a positive bronchoalveolar lavage for CMV to ganciclovir or no treatment. Goodrich and colleagues [11] began ganciclovir administration after engraftment for a positive surveillance culture for CMV from the throat, urine, or blood. Both studies showed clinically and statistically significant decreases in both CMV infection and disease. Both studies, however, had approximately the same incidence of CMV disease (12% to 13%) in patients who did not have a positive surveillance culture before the diagnosis of CMV interstitial pneumonia or gastrointestinal disease.

Because a clinically significant number of patients present with CMV disease without positive surveillance cultures, we elected to try to suppress all CMV disease by giving ganciclovir to all CMV-seropositive patients immediately after engraftment. Our study suggests that ganciclovir given prophylactically to CMV-seropositive allogeneic marrow transplant recipients immediately after engraftment results in suppression of CMV infection and markedly decreases CMV disease in the first 100 days after transplant. Fourteen cases of CMV infection and nine episodes of CMV disease occurred in the placebo group with only one case of CMV excretion and no disease in the group receiving ganciclovir in the first 100 days after transplant. Our data are similar to those of the study done by Atkinson and colleagues [16] in which they showed the prophylactic administration of ganciclovir resulted in the elimination of CMV infection and disease when compared with historical controls.

Toxicity was similar to that found in previous studies. The incidence of neutropenia in the studies by Schmidt and colleagues [10] and by Goodrich and colleagues [11] were 35% and 30%, respectively. Our current study showed an incidence of 30%. Although the incidence of neutropenia is identical to our previous study, the median time to recovery above 0.750 × 10⁻⁹/L differed from 12 days in our current study compared with 4 days in our early treatment study [11]. The reason for this difference is not known. It is possible that ganciclovir toxicity is more pronounced when given earlier after transplant.

Analysis of the bacterial and fungal infections showed no statistical difference in patients who received ganciclovir compared with those who received placebo. Patients who became neutropenic from ganciclovir prophylaxis had an associated increase in the risk for bacterial infection during the period of neutropenia. One explanation is the increased duration of neutropenia seen in this study. A review of infectious complications during the period of neutropenia from our early treatment study [11] revealed an incidence of 9% (unpublished data) compared with 40% in this study. Whether the judicious use of growth factors in conjunction with ganciclovir would alleviate this problem remains to be determined.

Contrary to previous studies [10, 11], survival after transplant did not differ when ganciclovir was given prophylactically. In our previous study [11], but not this one, there was a statistical difference in survival both at 100 and 180 days after transplant due to a reduction in CMV disease in the first 100 days. In our current study, excretion of CMV was an end point. The lack of a difference in survival between the two study arms may be due to a reduction in excess deaths secondary to CMV disease by the administration of ganciclovir to all placebo patients who excreted CMV.

In our previous study, the percentage of patients dying of CMV disease without positive surveillance cultures was 5% (unpublished data) [11]. These data are comparable to our current study in which one patient on placebo died of CMV disease without having a positive surveillance culture for CMV (1 of 31; 3%). When these data are applied, it seems that ganciclovir prophylaxis in seropositive patients could reduce to near zero the 3% to 5% mortality rate due to CMV disease seen in early treatment strategies. Our data suggest, however, that the use of ganciclovir prophylaxis in all patients would result in the administration of the drug to a large number of patients (12 of 31, 39%) in our group who will never excrete virus or have CMV disease and would result in the additional risk for neutropenia, secondary infection, and death.

Thus, although ganciclovir prophylaxis will further reduce the incidence of CMV infection and disease in CMV-seropositive allogeneic marrow transplant recipients, it could subject a substantial number of patients to the toxicity of ganciclovir, including neutropenia and infection. Whether new methods such as the measurement of antigenemia [17] or polymerase chain reaction [18, 19] will better identify patients at risk for CMV disease and thus avoid the problems of prophylaxis remains to be studied. Until then, the recommendation of administration of ganciclovir prophylactic to CMV-seropositive patients undergoing allogeneic marrow transplant must be based on a careful assessment of the risks and benefits for the individual patient.