The optimal approach to the prevention and treatment of infection due to cytomegalovirus (CMV) remains uncertain despite years of experience with antiviral therapies. Two approaches to the prevention of CMV disease have emerged: “universal” prophylaxis and “preemptive therapy.” Although both can prevent tissue-invasive CMV disease, they are, in fact, quite distinct. Universal prophylaxis provides antimicrobial therapy to all at-risk patients for a defined period. Viremia occurs uncommonly during prophylaxis, but occurs in up to ∼50% of patients subsequently (“late infection”) in the highest risk recipients (seronegative recipients of organs from seropositive donors, D+/R–). Such infections may be asymptomatic, or may cause “CMV syndrome” (fever, neutropenia) or tissue-invasive disease.
“Preemptive therapy” utilizes sensitive assays (e.g. molecular, antigen detection) to monitor patients at predefined intervals to detect viral replication (viremia) before infection progresses. A positive assay triggers the initiation of antiviral therapy, a reduction in the intensity of immunosuppression and/or intensified monitoring. Preemptive therapy incurs extra costs for monitoring and coordination of outpatient care as well as clinical effects of episodes of viremia, but reduces drug exposures and avoids some of the costs and toxicities of prophylactic antiviral therapy. With intensification of immunosuppressive therapies for transplantation, notably T-cell depleting antibody “induction” therapies, more recipients now receive some period of antiviral prophylaxis-–most often for the D+/R– combination and seropositive recipients (R+) following induction therapy. The efficacy of prophylaxis is presumed to be related to the dose and duration of the antiviral agent and the susceptibility of the virus to the agent.
Each individual with CMV infection carries multiple CMV species (1). The clinical impact of sequence polymorphisms depends on whether they confer drug resistance or affect viral “fitness” and on the status of the host's CMV-specific immunity. All of the currently marketed CMV antiviral drugs, including ganciclovir, foscarnet and cidofovir, target the viral DNA polymerase. Resistance is most often associated with mutations in the viral UL97 kinase and UL54 DNA polymerase genes (1). Drug resistance generally occurs after prolonged drug exposure (usually weeks to months) with incomplete suppression of viral replication. Clinically, drug resistance is assumed in the presence of rising viral loads and clinical symptoms in the presence of appropriately dosed antiviral therapy. Not all such cases are associated with described genomic resistance mutations. The risk for CMV resistance is greatest in the immunosuppressed D+/R– solid-organ recipients with primary infection. CMV drug resistance may occur in 5–12.5% of this group and may be higher in lung transplant recipients (2,3). There are no large prospective trials examining clinical outcomes in drug-resistant CMV infections. However, clinical experience suggests that adverse clinical outcomes are observed more frequently with CMV disease that persists despite therapy-–related to the infection, or to drug toxicities, opportunistic superinfections or secondary effects including bronchiolitis obliterans or coronary vasculopathy.
One of the benefits of preemptive therapy has been assumed to be that decreased exposure to antiviral agents (most often ganciclovir or valganciclovir) would result in lower rates of antiviral resistance. This assumption is challenged by a retrospective analysis in this issue by Couzi et al. of D+/R– kidney recipients, in which the rates of CMV infection and treatment failure were “higher” with preemptive therapy than with antiviral prophylaxis (4). The rate of known ganciclovir-resistance mutations (UL97 or UL54) was also higher in patients receiving preemptive therapy. Treatment failures for CMV infection have been recognized in other studies of patients receiving both preemptive and prophylactic therapies (3,5).
This observation seems counterintuitive. However, the risk for antiviral resistance in this study was related to the peak viral load in each patient. Both the incidence of CMV infections and peak viral loads were greater in the preemptive group than in the prophylaxis group. Preemptive therapy patients with viremia were more often treated with oral valganciclovir therapy and suffered more treatment failures than in those treated intravenously. Thus, although preemptive therapy recipients had less overall drug exposure, they generally received “oral antiviral therapy during periods of higher level viral replication”-–major risk factors for resistance. The preemptive group was also more likely to receive tacrolimus than cyclosporine, had earlier disease (possibly before reductions in immunosuppression), and had a greater degree of donor–recipient HLA-mismatch. As a result, they may have been more immunocompromised than the comparator group.
What are the implications of these observations? In the attempt to reduce drug exposure (preemptive therapy) and hospitalization for intravenous treatment of CMV infections, patients may be exposed to suboptimal levels of antiviral agents during periods of active viral replication. In the seronegative graft recipient with primary infection and without antiviral immunity, viral infection may require more intensive (intravenous) therapy for rapid suppression. The adequacy of the virologic response (negative or falling viral load assay) should be documented before assuming the efficacy of therapy. These considerations mitigate toward the use of prophylaxis in place of preemptive therapy for the highest risk groups for viral replication, the D+/R– combination and R+ receiving T-cell depleting antibody-based induction therapies. An alternative is the intensification of monitoring for such patients receiving preemptive therapy. The development of clinical assays to measure pathogen-specific immune function may allow individualization of prophylactic regimens in the future.