Cytomegalovirus (CMV) infection is common after lung transplantation. We performed a prospective trial of valganciclovir prophylaxis in lung recipients with outcomes compared to matched historical controls. The valganciclovir group (n = 40) (including D+/R– and R+ patients) was prospectively enrolled, and received oral valganciclovir 900 mg once daily for 12 weeks. Historical controls (n = 40) received 12 weeks of daily intravenous ganciclovir if D+/R– or 12 weeks of oral ganciclovir if R+. CMV viral load testing was done at two-week intervals until 6 months posttransplant. Baseline demographics and immunosuppression were comparable in the two groups. The incidence of CMV viremia was 16/40 (40.0%) in the valganciclovir arm versus 18/40 (45%) in the ganciclovir arm (p = NS). The incidence of symptomatic CMV disease was 8/40 (20%) versus 7/40 (17.5%), respectively (p = NS). In both groups viremia, while on prophylaxis, was uncommon (valganciclovir: 0/40 and ganciclovir: 2/40). Peak viral load and time to viremia were similar in the two arms. High rates of viremia and symptomatic disease occurred in the D+/R– patients after discontinuation of prophylaxis. Genotypic CMV sequence analysis demonstrated low rates of ganciclovir resistance in both groups. Valganciclovir prophylaxis had similar efficacy to either intravenous ganciclovir (D+/R– patients), or oral ganciclovir (R+ patients) in lung recipients.
Cytomegalovirus (CMV) remains an important cause of morbidity and mortality in lung transplant recipients and is second only to bacteria as the most common opportunistic pathogen (1,2). In fact, the rate of CMV infection and disease in lung transplantation is higher than in other solid organ transplant groups. Lung and heart–lung recipients who are not receiving antiviral prophylaxis have a reported incidence of CMV disease ranging from 38–75% (2). Invasion of the allograft resulting in CMV pneumonitis is not uncommon and may result in substantial mortality. In addition to directly attributable effects, CMV may also have an immunomodulatory effect in transplant recipients, and active CMV disease has been found to be an independent risk factor for the development of other infectious complications such as bacteremia, invasive fungal disease and Epstein-Barr virus related posttransplant lymphoproliferative disease (3–5). In addition, CMV may be an important risk factor for the development of bronchiolitis obliterans syndrome (BOS) postlung transplant (6,7). Given the potential for adverse consequences of CMV, strategies aimed at preventing the development of active infection are preferable.
One option for the prevention of CMV in lung transplant recipients is to use prolonged ganciclovir (either intravenous or oral) prophylaxis with or without immune globulin (8). However, oral ganciclovir has a low bioavailability necessitating three times a day dosing to achieve adequate plasma levels. In addition, concern exists regarding the development of antiviral resistance due to suboptimal ganciclovir levels achieved with oral ganciclovir (9). Oral valganciclovir is an oral prodrug of ganciclovir that is rapidly hydrolyzed to ganciclovir after oral ingestion. It has a bioavailability of about 60% compared to 6–9% for standard oral ganciclovir (10). Many lung transplant programs have adopted valganciclovir as an alternate to prolonged courses of intravenous or oral ganciclovir. We prospectively assessed virologic and clinical outcomes using a 3-month course of valganciclovir prophylaxis in lung transplant recipients compared with standard ganciclovir prophylaxis.
Patients were enrolled at two Canadian transplant centers (center A and center B) and consisted of prospectively enrolled cases and matched historical controls. Prospectively enrolled patients were recruited from 2002 and 2003. Historical controls were from between 1999 and 2002. Prospectively enrolled patients were given valganciclovir prophylaxis and controls had received either IV or oral ganciclovir prophylaxis as described below. Controls were matched for CMV pretransplant donor (D) and recipient (R) serostatus, and by transplant center. No other matching criteria were used. All patients had to be able to take oral therapy within 2 weeks of transplant. Exclusion criteria included previous transplantation and CMV D–/R–. The institutional review board at each center approved the study and informed consent was obtained from patients. Standard immunotherapy at center A included induction antilymphocyte globulin for 3–5 d. Induction therapy was not routinely used at center B. Standard immunosuppression consisted of cyclosporine with targeted trough levels of 400–450 ng/mL for 0–3 months posttransplant, then 200–250 ng/mL thereafter (center A), or 250–350 ng/mL for 0–3 months, 250–300 ng/mL for 3–6 months, and 200–250 ng/mL for 6–12 months (center B). Prednisone and azathioprine (1.5–2 mg/kg/d) or mycophenolate mofetil (MMF) were used in addition to cyclosporine. Standard therapy for rejection was pulse methylprednisilone 500–1000 mg intravenous daily for 3 d. Antilymphocyte therapy was used for steroid-resistant rejection. All patients received prophylaxis for Pneumocystis jirovecii with trimethoprim/sulfamethoxazole daily or three times a week.
All patients in both arms received 2 weeks of intravenous ganciclovir (10 mg/kg/d). In the historical control arm, this was followed by 12 weeks of oral ganciclovir (CMV R+ patients; 1 g po t.i.d.) or intravenous ganciclovir (D+/R– patients; 5 mg/kg IV daily). In the prospective valganciclovir arm all patients, regardless of D/R serostatus, received 12 weeks of oral valganciclovir prophylaxis (900 mg once daily). All doses were adjusted appropriately for renal function. CMV immune globulin was not used for prophylaxis in any patient.
All patients in both groups had CMV antigenemia testing at a minimum of 2-week intervals until 6 months posttransplant. These tests were part of the clinical monitoring practice already in place with results available to clinicians. Plasma samples from each of these time points were saved for viral load testing. Viral load testing (quantitative PCR) was done using the Roche Cobas Amplicor assay (Roche Diagnostic Systems, Inc., Branchburg, NJ) according to manufacturer's instructions (lower limit of detection ∼400 copies/mL) (11). The results of viral load testing were not available to treating physicians. Although treatment was not mandated as part of the trial, generally any patient with symptoms or any patient with a positive antigenemia exceeding five or more positive cells/slide received treatment with full dose intravenous ganciclovir until antigenemia became negative at the discretion of the treating physician.
Genotypic resistance testing was done for all patients in both groups with detectable viremia. Testing was done on the first and last sample with detectable virus. Briefly, samples were first screened for UL97 mutations using a seminested procedure that amplifies a region encompassing codons 363–698 (12). The sequenced amplicon covered all reported UL97 mutations associated with ganciclovir resistance. In samples that contained a known ganciclovir-resistance UL97 mutation, the UL54 gene was sequenced after a nested PCR procedure covering all reported mutations associated with drug resistance (codons 184–1017).
Both virologic and clinical outcomes were evaluated. The primary outcome was the incidence of viremia (defined as detection of DNA by PCR) in the first 6 months posttransplant. Other outcomes assessed included time to viremia, peak viral loads and symptomatic CMV disease.
CMV infection was defined as any detectable virus by quantitative PCR and included patients with and without symptoms. CMV disease was defined according to standard criteria and included CMV syndrome and tissue invasive disease (13).
CMV Syndrome: CMV infection accompanied by fever ≥38°C (100.4°F) and one or more of malaise, arthralgias, myalgias, leukopenia, thrombocytopenia, or transaminase elevation.
Tissue-invasive CMV: Compatible symptoms or signs of organ involvement and evidence of localized CMV infection in a biopsy or other appropriate specimen. Patients with a positive bronchoalveolar lavage culture for CMV were counted as CMV pneumonitis if they had a compatible clinical and radiological picture.
Categorical variables were compared using χ-square or Fisher's exact test. Continuous variables were compared with Mann-Whitney U test. Time to viremia analysis was done using the Kaplan-Meier method with log-rank statistic. All statistical analysis was done with SPSS version 11.0 (SPSS, Inc., Chicago, IL).
A total of 80 patients were included in the study; this included 40 patients receiving valganciclovir prophylaxis and 40 historical controls who had received ganciclovir prophylaxis. Baseline characteristics and immunosuppression are shown in Table 1 and were comparable in both groups.
Table 1. Demographics and characteristics of patients in the two arms
Valganciclovir (n = 40)
Ganciclovir (n = 40)
ap = not significant for all comparisons.
Mean age ± SD (years)
48.6 ± 12.9
50.6 ± 11.8
Single vs. double lung transplant
Idiopathic pulmonary fibrosis
α-1 antitrypsin deficiency
Donor (D) / Recipient (R)
Cyclosporine/prednisone/ mycophenolate mofetil
for acute rejection
A total of 32 patients were enrolled at center A (16 cases and 16 controls) and 48 patients at center B (24 cases and 24 controls). Patients at the two centers were similar in terms of baseline demographic characteristics, CMV serostatus and underlying lung disease (data not shown). Center A routinely used induction antilymphocyte therapy (23/32; 71.9%), whereas center B did not (0/48) (p < 0.01). Center A more commonly used MMF (instead of azathioprine) than did center B (31/32 vs. 4/48) (p < 0.01). To control for differences in centers, patients and historical controls were matched by center.
Cytomegalovirus infection and disease
Although antiviral prophylaxis was only until 3 months posttransplant, the primary outcomes were assessed at 6 months (Table 2) since viremia and disease are common after discontinuation of prophylaxis. The incidence of detectable viremia (viral load > 400 copies/mL; includes both CMV infection and symptomatic CMV disease) within the first 6 months posttransplant was (16/40) 40% (95% CI 26–56%) in the valganciclovir arm and (18/40) 45% (95% CI 31–60%) in the ganciclovir arm (p = NS). Time to first detectable viremia by Kaplan–Meier analysis was 149 ± 18 d in valganciclovir group versus 138 ± 29 d in ganciclovir group (p = NS) (Figure 1). While on prophylaxis, viral suppression was good in both arms of the study. There were no cases of detectable viremia in the first 3 months in the valganciclovir arm and two cases (5.0%) in the ganciclovir arm. A scatter plot of peak viral load results is shown in Figure 2. The mean peak viral load in the valganciclovir patients was 8890 ± 18 600 copies/mL (median 0; 25th percentile 0; 75th percentile 8742 copies/mL) and was 8393 ± 17 270 copies/mL (median 0; 25th percentile 0; 75th percentile 9215 copies) in the ganciclovir arm (p = NS). When only viremic patients were analyzed, the mean peak viral load was 22 225 ± 24 151 copies/mL (median 13 000; 25th percentile 3430; 75th percentile 36 500 copies/mL) and 18 652 ± 21 947 copies/mL (median 9365; 25th percentile 4105; 75th percentile 31 325 copies/mL) respectively (p = NS).
Table 2. Outcomes in the first 6 months posttransplant
Valganciclovir (n = 40)
Ganciclovir (n = 40)
ap = not significant for all comparisons.
Symptomatic CMV disease CMV disease
Peak Viremia (copies/mL, mean ± SD)
8890 ± 18 600
8393 ± 17 270
CMV viremia in D+/R– subgroup
Symptomatic CMV disease in D+/R– subgroup
CMV disease occurred in 8/40 (20%; 95% CI 10–35%) patients on valganciclovir and 7/40 (17.5%; 95% CI 8–32%) patients on ganciclovir. No cases of disease occurred while on prophylaxis. Time to onset of CMV disease was 151 ± 13 d in the valganciclovir arm (median 152 d) and 136 ± 31 d in the ganciclovir arm (median 129 d) (p = NS). The most common type of disease in both arms was CMV viral syndrome. Tissue invasive disease occurred only in two patients (CMV pneumonitis; both patients in ganciclovir arm). These two patients were treated with CMV immune globulin in addition to ganciclovir therapy.
Although the primary endpoint was at 6 months, clinical data up to 1-year posttransplant were analyzed as a secondary outcome. CMV infection between months 6–12 was diagnosed in 12/40 (30%; 95% CI 18–46%) in the valganciclovir arm (median time of onset 224 d) and 14/40 (35%; 95% CI 22–51%) in the ganciclovir arm (median time of onset 208 d) (p = NS). The majority of these represented relapsing infection in patients who were previously viremic in the first 6 months posttransplant. Only one new onset CMV infection occurred in each group. Symptomatic disease occurred in 3/40 (7.5%; 95% CI 2–21%) and 5/40 (12.5%; 95% CI 5–27%), respectively (p = NS). All of the disease cases except one patient had a previous episode of CMV infection in the first 6 months.
Of the 14 patients that had both acute rejection and CMV infection, CMV infection preceded the first rejection episode in three cases and was after the rejection episode in 11 patients. In the seven patients with both CMV disease and acute rejection, CMV disease preceded the rejection episode in one patient and was after the rejection episode in six patients. The incidence of CMV infection was 37.8% in patients with a history of acute rejection vs. 48.8% in those without (p = NS). The incidence of CMV disease was 18.9% and 18.6%, respectively (p = NS). The use of antilymphocyte therapy was not found to be statistically associated with CMV infection or disease (data not shown).
CMV D+/R– subgroup
The high-risk D+/R– subgroup was analyzed separately as well. Ten patients received IV ganciclovir prophylaxis and 10 patients received valganciclovir prophylaxis. Again, both regimens were effective for suppressing viremia while on prophylaxis (Figure 3). However, following discontinuation of prophylaxis there was a very high incidence of viremia (8/10 (80%) patients in both groups) and a high incidence of CMV disease (4/10 (40%) patients in both groups). No significant differences were observed with valganciclovir vs. IV ganciclovir. In viremic patients, peak viral loads were 29 844 ± 30 581 copies/mL in valganciclovir arm vs. 22 189 ± 21 788 copies/mL in IV ganciclovir arm (p = 0.57).
All patients with CMV disease within the first 6 months were treated initially with intravenous ganciclovir. Two patients were treated with foscarnet during their first episode of CMV disease due to lack of virologic response to ganciclovir (i.e. failure of the antigenemia result to decline with full dose intravenous ganciclovir). All patients with detectable viremia had genotypic resistance testing from samples taken at the time of first and last detectable viremia within the first 6 months. UL97 sequencing was performed on a total of 49 samples. Two patients were found to have a UL97 mutation consistent with ganciclovir resistance (one patient in ganciclovir arm (CMV D+/R–) and one patient in valganciclovir arm (CMV D+/R+); mutations Met460Ile and Thr659Ile, respectively). Sequencing of the UL54 gene in these patients revealed no significant mutations. One of these patients responded to ganciclovir while the second required foscarnet. A third patient (noted above; valganciclovir arm, CMV D+/R–) was treated with foscarnet due to a lack of clinical response to ganciclovir. Although, no resistance mutations were detected from the samples from the first 6 months posttransplant in this patient, follow-up testing on a subsequent sample (at 10 months posttransplant) confirmed a UL97 mutation at codon 594 (Ala to Val) known to confer ganciclovir resistance and a UL54 mutation at codon 412 (Phe to Leu) suspected to confer both ganciclovir and cidofovir resistance (14). The two patients in the ganciclovir arm who had breakthrough viremia while on prophylaxis were not found to have a mutation conferring resistance.
During the 6-month follow-up, only a single death was reported (in the valganciclovir group due to cardiac disease). Both regimens were well tolerated. The most significant adverse event was neutropenia. Significant neutropenia (absolute neutrophil count < 1 billion/L) requiring discontinuation of antiviral prophylaxis and G-CSF occurred in three patients receiving valganciclovir (7.5%) and one patient receiving ganciclovir (2.5%) (p = NS). No induction of rejection was observed with G-CSF use.
Lung transplant recipients are at particularly high risk of CMV disease. This had led to many centers adopting universal antiviral prophylaxis regimens for all lung transplant patients at risk of CMV disease. We compared the virologic and clinical efficacy of a 3-month prophylaxis regimen of valganciclovir with ganciclovir (oral ganciclovir for R+ patients and IV ganciclovir for D+/R– patients). Both regimens achieved good viral suppression while on therapy with no detectable viremia observed in the valganciclovir arm and only two cases of detectable viremia in the ganciclovir arm. However, viremia was common after discontinuation of therapy. The rate of subsequent viremia was similar in the two arms, and the peak viral loads in patients who were viremic were comparable. The occurrence of symptomatic CMV disease was similar as well. A particularly high rate of viremia and disease (twice the percentage of the entire cohort) was observed in the D+/R– arm after discontinuation of therapy. However, it is reassuring that valganciclovir was comparable to intravenous ganciclovir in terms of viral suppression in this very high-risk D+/R– subgroup of lung transplant recipients. The high rate of CMV following discontinuation of prophylaxis especially in, but not limited to, D+/R– patients suggests that continuation of primary prophylaxis for a longer period or the use of alternative antiviral strategies may be beneficial.
Oral ganciclovir generally has poor bioavailability (6–9%) requiring large doses and a three times a day dosing interval (10). Valganciclovir is an oral prodrug of ganciclovir. After ingestion it is hydrolyzed to ganciclovir by intestinal and hepatic esterases. The estimated oral bioavailability of valganciclovir is 50–60% (10). Valganciclovir was compared with oral ganciclovir in a large multicenter trial of 364 CMV D+/R– liver, kidney, heart and pancreas recipients (15). Notably, lung transplant patients were excluded from this study. This trial demonstrated comparable efficacy of the two regimens as well as good viral suppression while on therapy. Again, viremia and disease were most commonly seen after discontinuation of prophylaxis. There is limited published literature on the use of valganciclovir in lung transplant recipients.
In general, large randomized clinical trials of CMV prophylaxis in lung transplant recipients are lacking. Therefore, the optimal prophylaxis strategy is unknown (16). In studies that evaluated ganciclovir prophylaxis (usually intravenous ganciclovir), high rates of CMV infection and disease have been reported usually after discontinuation of prophylaxis (17–21). Duration of prophylaxis in these studies has ranged from several weeks to months. In a review of ganciclovir prophylaxis in lung transplant recipients, the average incidence rates from the combined studies were 60% for CMV infection and 43% for CMV disease (16). The use of CMV immune globulin in addition to ganciclovir therapy has been advocated by some centers. This is based on the fact that primary CMV infection, in which endogenous antibody protection is lacking, is often associated with severe disease with high mortality in some settings. Data from some series have demonstrated improved outcomes with a possible beneficial impact on BOS, but randomized controlled trials are lacking (22,23). The centers participating in our study do not routinely use immune globulin for CMV prophylaxis.
Recently, Zamora et al. evaluated valganciclovir prophylaxis in lung transplant recipients (24). All patients received either 30 d of intravenous ganciclovir and three doses of CMV immune globulin (if R+) or 90 d of intravenous ganciclovir with seven doses of CMV immune globulin (if D+/R–). In 90 patients who completed prophylaxis with subsequent valganciclovir out to either 180, 270 or 365 d posttransplant, the incidence of CMV disease was significantly lower compared to a historical control group that received acyclovir after the initial course of intravenous ganciclovir (2.2% vs. 20%). Longer courses of prophylaxis were associated with less CMV disease: freedom from CMV infection and disease was between 90–95% for those receiving between 180–365 d of prophylaxis versus 64% for those receiving 100–179 d of prophylaxis versus 59% for those receiving <100 d of prophylaxis (p < 0.02). Our study also supports the notion that relatively long courses of prophylaxis are likely required in lung transplant recipients given the high rate of viremia we observed after discontinuation of prophylaxis. This is true both in D+/R– patients and R+ patients. An important difference between our study and the one by Zamora et al. was that patients in the latter study were given 30 d of intravenous ganciclovir and 90 d of intravenous ganciclovir if D+/R– and all patients were given CMV immune globulin. In our study, we used oral valganciclovir in all serostatus subgroups after 2 weeks of intravenous ganciclovir, and no immune globulin was used for prophylaxis.
Ganciclovir-resistant CMV is an emerging problem in transplant recipients. Currently available therapies for ganciclovir-resistant CMV include cidofovir and foscarnet, both of which have significant toxicity and are poorly tolerated in transplant recipients. In general, risk factors for resistance include CMV D+/R– serostatus, more potent immunosuppression, and prolonged use of oral ganciclovir (9). However, lung transplant recipients appear to be at especially high risk of developing ganciclovir-resistant CMV (23). Resistance in this setting has been described with both pre-emptive and universal prophylaxis approaches (25,26). Theoretically, if good viral suppression is achieved with an antiviral prophylaxis regimen, the emergence of resistance should be relatively uncommon. It is encouraging that we found very little resistance in this cohort of lung transplant recipients. This is in keeping with data from the multicenter valganciclovir trial in D+/R– organ transplant recipients (excluding lung) (27). In the 239 patients receiving valganciclovir prophylaxis, no cases of clinical or genotypic ganciclovir resistance was found. However, lung transplant recipients were excluded from that trial. Also, since resistance may more commonly be seen in the setting of relapsing disease (i.e. beyond 6 months), larger studies of valganciclovir prophylaxis in lung transplant recipients with longer follow-up will be required determine the rate of resistance in this patient population.
The primary limitation of our study included the use of historical controls. Clearly a randomized controlled trial would have been a superior study design; however, with the widespread availability of valganciclovir, such a study would have been difficult to carry out at our centers. We believe that the control and intervention groups were quite comparable. The immunosuppression protocols remained the same at the two centers for the time period of the study, and patients were comparable with respect to all other baseline characteristics. In addition, the use of serial CMV viral loads is an objective outcome, which is not as subject to potential biases such as interpretation of CMV disease symptoms. All viral load testing was done in the same laboratory with a standardized, commercially available assay conducted by technologists blinded to the clinical status of the patients.
In summary, this study demonstrates that valganciclovir is comparable to ganciclovir (IV and oral) for CMV prophylaxis in lung transplant recipients. Good viral suppression is achieved while on prophylaxis, but both viremia and disease are common after the discontinuation of prophylaxis. The rate of ganciclovir resistance on either regimen was low during the first 6 months. This study suggests that a longer duration of prophylaxis should be considered and further studied in lung transplant.
This study was supported by an unrestricted grant from Roche Pharmaceuticals.