Valganciclovir prevents cytomegalovirus reactivation in patients receiving alemtuzumab-based therapy

Alemtuzumab (Campath 1H) is a humanized monoclonal antibody targeting the CD52 antigen.¹  This antigen is present on B cells as well as T cells and highly overexpressed on the leukemic cells from patients with chronic lymphocytic leukemia (CLL). In 2000, the Food and Drug Administration approved alemtuzumab for the treatment of fludarabine refractory disease. The pivotal multicenter trial involved 93 patients given alemtuzumab at 30 mg intravenously 3 times weekly for 4 to 12 weeks.²  The overall response rate was 33%. Alemtuzumab has also been used as the initial treatment for patients with CLL, producing a response rate of 87% with 19% complete remission.³ 

Because of the significant and expected lymphopenia seen with the use of this agent, prophylactic antibiotics are routinely given to prevent Pneumocystis carinii pneumonia as well as reactivation of herpes viruses. Thus, most patients receive trimethoprim-sulfa twice daily 3 times a week and daily anti-herpes virus treatment with acyclovir, famciclovir, or valaciclovir.

Since the introduction of alemtuzumab therapy into the treatment armamentarium for CLL, it has been recognized that cytomegalovirus (CMV) reactivation may occur in patients receiving this antibody, presumably because of the extensive T-cell depletion associated with the therapy.^4,5 

A review of 78 patients with chronic lymphocytic disorders receiving alemtuzumab at M. D. Anderson Cancer Center indicated that the symptomatic reactivation rate was 20%.⁶  A further evaluation of approximately 200 patients treated with this agent at our center indicated a CMV reactivation rate of 20% to 25%. Although mortality is rare, morbidity is common with fever that often leads to hospitalization of the patient and interruption of the alemtuzumab therapy. Thus, preventing CMV reactivation would be a desirable endpoint.

Traditionally, in the setting of allogeneic stem cell transplant, CMV prophylaxis has been provided to high-risk patients with the use of daily intravenous ganciclovir; if myelosuppression was problematic, then foscarnet was used.⁷  However, daily intravenous infusions would be cumbersome over the 4- to 12-week duration of therapy typically experienced with alemtuzumab. Valganciclovir is an oral esteric pro-drug of ganciclovir.⁸  It is rapidly converted to ganciclovir by intestinal and hepatic esterases. The oral bioavailability of this product is far superior to oral ganciclovir with a 60% absorption rate when taken with food. Pharmacokinetic studies have shown that 900 mg of oral valganciclovir can produce similar area under the time curve plasma concentrations to 5 mg/kg of intravenous ganciclovir. A dose of 900 mg twice daily orally has been recommended for the treatment of CMV disease.⁹  Since once daily intravenous ganciclovir is used for prophylaxis, valganciclovir 450 mg orally twice daily (900 mg total dose) was chosen as the prophylactic dose for this study.

Patients eligible for the study were adults (age > 15 years) receiving any alemtuzumab-based regimen. Consecutive patients being evaluated for alemtuzumab-based therapy were also asked to participate in the current trial. Exclusion criteria included pregnant women, patients with active infection or patients with a creatinine clearance of more than 10 mL/min as calculated via the Cockroft-Gault equation. Patients were randomized via a computerized program to receive either valganciclovir 450 mg orally twice daily or valaciclovir 500 mg orally daily. Prophylaxis was continued throughout the duration of alemtuzumab therapy and for 2 months after therapy.

Response was defined as no reactivation of CMV during alemtuzumab therapy and for 2 months after treatment. CMV reactivation was defined as fever with positive antigenemia and no other cause for the fever.

Forty-six patients were entered in the study; 6 were inevaluable. The reasons for inevaluability in 6 patients were as follows: 2 patients developed sepsis early into therapy and came off study and did not receive any more alemtuzumab, 1 patient's physician increased his valaciclovir dose above that specified in the protocol, 1 patient was lost to follow-up, 1 patient was noncompliant, and 1 patient refused the randomization. Three patients were on the valaciclovir arm and 3 were on the valganciclovir arm. All patients signed an informed consent obtained according to institutional guidelines and in accordance with the Declaration of Helsinki. Approval was obtained from the University of Texas M. D. Anderson Cancer Center's Institutional Review Board for this study. Pretreatment evaluation included a history and physical examination, a complete blood count, differential, and serum chemistries and CMV PP65 antigenemia (IFH Kit, Product 3247; Light Diagnostics, Temecula, CA). Women of childbearing potential had a negative pregnancy test. During therapy, a complete blood count was done at least every 2 weeks, and serum chemistries were performed at least every 4 weeks. CMV PP65 antigenemia was tested every 2 weeks.

The statistical design called for enrollment of 128 patients. This sample-size estimate was based on providing 80% power to detect an 80% decrease in the incidence of CMV infection (from 25% in the control group to 5% in the treated group) at the 2-tailed 10% significance level. A 5% dropout rate was assumed, resulting in an effective sample size of 120. The patients were randomized 1:1 evenly between valganciclovir and valaciclovir. The patients were randomized via balanced block, balancing every 4. There were no stratification parameters. Three interim analyses were planned after 30, 60, or 90 patients were evaluable. A Bayesian model was used to monitor the analyses.

CMV infection rates are denoted by P and Q, respectively, for patients receiving valganciclovir and valaciclovir. Assume the prior distributions for P and Q are both Beta (0.4, 1.6), which is corresponding to a 20% infection rate, with prior information equivalent to 2 observations. During interim analyses, the distributions of P and Q will be updated by observed data. If Pr [P < Q|data] > 0.99, we will stop the trial and conclude that valganciclovir is better than valaciclovir as a prophylaxis for CMV. On the other hand, if Pr [P < Q|data] < 0.01, then we will also stop the trial and conclude that valganciclovir is not better than valaciclovir. During the final analysis, if Pr [P < Q|data] > 0.97, we will conclude in favor of valganciclovir. Otherwise, we will conclude that there is not sufficient evidence in favor of valganciclovir. By this design and the sample size of 120, the probability to falsely choose valganciclovir as superior is 5% when actually both of these 2 prophylaxes have a rate of 20% of CMV infection. However, when the infection rates are 5% and 20%, respectively, for valganciclovir and valaciclovir, the power to select valganciclovir as superior is 79%. After the first interim analysis, the study was closed.

Patient characteristics are detailed in Table 1. The median age was 58 years (range, 25–83 years). Twenty-nine patients (73%) had CLL. Excluding 6 patients being treated with alemtuzumab for minimal residual disease, Rai stage was I or II in 13 patients and III or IV in 10. The median number of prior regimens for the entire group was 2 (range, 0–10).

Alemtuzumab was given either as a single agent (n = 5) or in combination with rituximab (n = 2), pentostatin (n = 6), cyclophosphamide, vincristine, adriamycin, dexamethasone (hyper-CVAD; n = 3), methotrexate and asparaginase (n = 1), or fludarabine, cyclophosphamide, rituximab (FCR; n = 23). The number of doses of alemtuzumab delivered per month were significantly different between the regimens and are shown in Table 2.

Among the first 30 patients, CMV reactivation occurred in 6 of 15 patients receiving prophylaxis with valaciclovir versus 0 of 15 patients receiving prophylaxis with valganciclovir. The Bayesian posterior probability that valganciclovir was better than valaciclovir was 0.999. This crosses the boundary of the early stopping rule as specified in “Statistical considerations.” Evaluation of 30 patients required completion of treatment ranging from 1 to 6 months. Patients continued to be enrolled during that time and were included in the final analysis (N = 40).

CMV reactivation occurred in 7 of 20 (35%; 95% confidence interval, [14%, 56%] by normal approximation) patients receiving prophylaxis with valaciclovir vs 0 of 20 (0%; with 95% shortest confidence interval, [0%, 13.3%]) patients receiving prophylaxis with valganciclovir (P = .004). With these data, the Bayesian posterior probability that valganciclovir was better than valaciclovir was 0.999. The number of CMV-positive cells by CMV antigen test ranged from 19 to 651. All patients but one developed fever; the patient without fever was the patient with the highest CMV level of 651, but he was receiving 4 mg of dexamethasone daily at that time. Cough was also a frequent symptom, but no patient had pneumonia or other evidence of organ involvement. Median time to development of CMV reactivation was 4 weeks (range, 4–7 weeks). The median number of doses of alemtuzumab administered at the time of CMV reactivation was 10 (range, 3–18). Three patients had completed the specified number of doses of alemtuzumab for that cycle at the time of CMV reactivation; among the others, alemtuzumab treatment was interrupted in 3 patients and continued in 1 patient (1 dose). Five patients were scheduled to resume/continue alemtuzumab treatment, and all 5 received further therapy (while receiving valganciclovir). Initial treatment for CMV reactivation included intravenous ganciclovir in 5 patients, a combination of ganciclovir and foscarnet in 1 patient, and oral valganciclovir in 1 patient.

The 7 patients with reactivation of CMV all recovered, and the follow-up of those patients ranged from 2 to 25 months (median, 5 months). Three patients received no further therapy for their disease, 1 underwent allogeneic bone marrow transplant, and 2 patients had 3 more therapies each. Only 1 patient had subsequent reactivation of CMV. This patient developed CMV reactivation on his first cycle of CFAR that was relatively resistant to intravenous ganciclovir and required a combination of ganciclovir and foscarnet. Subsequent to this, he was maintained on valganciclovir but was questionably compliant. After his third cycle of CFAR, he again had 114 positive CMV cells and was re-treated with intravenous antibiotics. After CFAR, he went on to receive subsequent therapies with forodesine and then a combination of oxaliplatinum, fludarabine, cytarabine, and rituximab. No further CMV was documented, although CMV antigen analysis was obtained during any febrile episode. This patient was followed for 21 months before dying of progressive disease.

No side effects were observed with the use of valganciclovir or valaciclovir. One potential difference in toxicity between these 2 agents is myelosuppression, which is known to occur with valganciclovir. However, given the heterogeneity of alemtuzumab containing regimens, as well as the expected myelosuppression from these combination regimens, it was not possible to determine whether valganciclovir increased the incidence of myelosuppression. An attempt to examine this was made by comparing the 14 patients treated with FCR and alemtuzumab who received valganciclovir prophylaxis with the 9 patients treated with FCR and alemtuzumab who received valaciclovir prophylaxis. All patients had relapsed CLL. Two patients began treatment with an absolute neutrophil count less than 1000/μL, and both were on the valaciclovir arm so they were not included in the analysis. Table 3 shows the comparison of grade 3 or 4 neutropenia and thrombocytopenia between the 2 arms for patients beginning treatment with an ANC more than 1000/μL and a platelet count more than 100 000/μL. No significant difference was seen in the incidence of neutropenia. Although there was a trend (not statistically significant) for more thrombocytopenia in the valaciclovir arm, it should be noted that the patients randomized to that arm had a lower median platelet count before treatment. Table 4 grades the hematologic toxicity in those patients starting treatment with a platelet count less than 100 000/μL. There were 4 patients in each arm. Hematologic toxicity as per NCI Guidelines indicated grade 4 thrombocytopenia in 3 of 4 patients in the valaciclovir arm. In the valganciclovir arm, one patient started with grade 4 thrombocytopenia. Toxicity grades in the other 3 patients were 2, 3, and 4. No difference in the duration of myelosuppression was observed between arms. Hospitalization for non–CMV-related infections occurred in 4 patients receiving valaciclovir prophylaxis and 5 patients receiving prophylaxis with valganciclovir.

In healthy CMV-seropositive patients, viral reactivation rarely results in infection because viral replication is effectively suppressed by the immune system. However, in immunocompromised patients, CMV reactivation may lead to viral replication and development of CMV disease (organ involvement). Patients undergoing either solid organ or stem cell transplantation are particularly susceptible to developing serious CMV-related complications. Reactivation of CMV in patients with CLL usually does not lead to organ involvement or death; such patients are not as immunocompromised as transplant patients. Typically, patients with CLL developing CMV reactivation have fever that is unresponsive to broad-spectrum antibiotics; documentation of CMV by either polymerase chain reaction or antigenemia testing allows prompt initiation of therapy, typically with intravenous ganciclovir. In recently published trials among patients with CLL treated with single-agent alemtuzumab, the incidence of symptomatic CMV reactivation ranged from 5% to 30%.^{2,–4,6,10,11}  The wide range in reported incidence reflects several factors, including study design, patient population, and the viral detection methods used. The earlier studies likely underreported the incidence of CMV reactivation because treatment centers were not routinely monitoring for this virus (in febrile patients) when alemtuzumab was initially used.^2,11,12 

CMV reactivation is typically observed between 3 and 6 weeks after the initiation of treatment with alemtuzumab^3,–5,10 ; at this point, marked lymphopenia has developed. CMV reactivation has not been reported during the posttreatment follow-up period.^2,3,5,10  Death resulting from CMV disease has rarely occurred in patients with CLL undergoing treatment with alemtuzumab. However, this does not take into account those CLL cases where this may have occurred and are not reported.

A recent study demonstrated that CD8⁺ T lymphocytes are significantly expanded toward a CMV-specific phenotype in patients with CLL who are latently infected with CMV; this did not occur in CMV-seronegative patients or in latently infected healthy individuals.¹³  This skewed repertoire of CD8⁺ cells may be important in maintaining viral latency in CMV-infected patients with CLL and may partially explain the occurrence of CMV reactivation in patients treated with lymphocyte-depleting agents, such as alemtuzumab.

The current trial was designed to investigate whether the use of valganciclovir as prophylaxis could prevent CMV reactivation in patients receiving alemtuzumab-based regimens. The trial was stopped early after the first interim analysis based on the significant efficacy of valganciclovir. No patients receiving this agent developed CMV reactivation, whereas 7 patients receiving standard prophylaxis with valaciclovir developed CMV reactivation. No significant toxicities were seen with the use of either prophylactic agent.

This trial did not attempt to perform a cost-benefit analysis. The cost per month for acyclovir or valaciclovir prophylaxis ranges from $50 to $200, whereas valganciclovir is closer to $2500 per month. Any analysis would have to take into account the cost of hospitalization for patients with CMV reactivation.

The potential concern with the use of prophylactic valganciclovir is enhanced myelosuppression over that seen with the alemtuzumab-based regimens alone. However, given the expected myelosuppression for most of the regimens used in this trial, it was not possible to determine whether any enhanced myelosuppression was seen. In examining only the 25 patients with CLL who received the same chemoimmunotherapy regimen of FCR and alemtuzumab and then comparing patients who received valganciclovir to valaciclovir, no difference in the incidence or duration of myelosuppression was seen. In addition, no increase in non–CMV-related infections was seen with valganciclovir. There were not enough patients treated with alemtuzumab alone to know whether valganciclovir might add to the myelosuppression seen with the single agent. Nevertheless, it is clear that the use of an easily administered oral viral prophylaxis with valganciclovir is effective at preventing reactivation of CMV in patients being treated with an alemtuzumab-based regimen.

This work was supported by Roche Pharmaceuticals.

Contribution: S.O. designed the trial, wrote the paper, and analyzed the data; F.R. treated patients on the trial and reviewed the paper; T.R. designed the trial and reviewed the paper; W.W. treated patients on the trial and reviewed the paper; X.H. designed the trial and reviewed the paper; J.T. reviewed the paper and performed research; B.O. performed research; H.K. treated patients on the trial and reviewed the paper; and M.K. designed the trial, treated patients on the trial, and reviewed the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Susan O'Brien, University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Unit 428, Houston, TX 77030; e-mail: sobrien@mdanderson.org.