Evidence supports the use of 12 months of cytomegalovirus prophylaxis in all at-risk lung transplants; whether cytomegalovirus serostatus can be used to further optimize this duration remains to be determined. The purpose of this retrospective study was to determine if cytomegalovirus serostatus of both donor and recipient were associated with late-onset cytomegalovirus. The primary outcome was the proportion of lung transplants that developed cytomegalovirus infection or disease during the 180-day period following 6 months of prophylaxis in each at-risk serotype. Two hundred forty-four consecutive lung transplants were evaluated, 131 were included. The proportion of recipients with cytomegalovirus differed significantly between serotypes (20 of 41 [48.8%] D+/R- vs. 19 of 56 [33.9%] D+/R+ vs. 2 of 34 [5.9%] D-/R+; p < 0.001). In a multivariate model, older age (odds ratio [OR], 1.05, 95% confidence interval [CI] 1.004–1.099; p = 0.03) and D+/R- serostatus (OR, 3.83; 95% CI 1.674–8.770; p = 0.002) were associated with cytomegalovirus. Among R+ lung transplants, D- serostatus was associated with the absence of cytomegalovirus (OR, 0.12; 95% CI 0.0263–0.563; p = 0.007). These findings suggest that in the valganciclovir era, cytomegalovirus serostatus of both donor and recipient may identify lung transplants at heightened risk for late-onset cytomegalovirus.
In 2010, an International Consensus Group on CMV management reviewed the available literature and published evidenced-based guidelines recommending at least 6 months of cytomegalovirus (CMV) prophylaxis in donor-positive/recipient-negative (D+/R-) and 3–6 months of prophylaxis in recipient positive (R+) lung transplant (LTx) patients . Following publication of these guidelines, a randomized, controlled trial demonstrated superiority of 12 months of valganciclovir (VGCV) prophylaxis over 3 months for preventing late-onset CMV disease in at-risk LTx recipients (D+/R- or any R+; . Although the aforementioned study clearly demonstrated that 12 months of CMV prophylaxis is an effective strategy in all at-risk LTx recipients, this extended duration has not been rigorously compared to 6 months of CMV prophylaxis, which had become the de facto standard of care at many LTx centers. Given the significant cost and potential toxicity associated with VGCV prophylaxis, determining the optimal duration of prophylaxis remains an important goal with broad implications for the LTx recipient.
Most R+ and many D+/R- LTx patients do not develop CMV infection or disease following completion of CMV prophylaxis suggesting that individual patient factors may predict the risk for CMV infection and disease. For example, the development of CMV-specific T cell immunity has demonstrated some success in predicting CMV risk [3-8]. It is also widely recognized that across all types of solid organ transplants, the serostatus of both the donor and recipient confer different levels of risk for the development of CMV infection or disease ([1, 9]–). R+ patients are at substantially lower risk for CMV disease than D+/R- patients. Further, R+ patients who receive a D- organ may be at lower risk for CMV when compared to those who receive a D+ organ [9-11]. In fact, one study of pediatric LTx recipients noted that following 3 months of ganciclovir prophylaxis, a numerically lower proportion of D-/R+ patients developed late-onset CMV when compared to their D+/R+ counterparts . Thus, until an accurate assay of an individual patient's CMV-specific immunity becomes universally commercially available, the CMV serostatus of both the donor and recipient may help guide the optimal duration of CMV prophylaxis in LTx recipients.
The purpose of this study was to determine if the CMV serostatus of both the donor and recipient were associated with the development of late-onset CMV in at-risk LTx recipients who completed 6 months of CMV prophylaxis with a ganciclovir-containing compound and serostatus-driven CMV hyperimmune globulin. We hypothesized that late-onset CMV infection and disease would occur in a significantly lower proportion of R+ LTx patients compared to D+/R- LTx patients, which might signal a limited benefit to extending prophylaxis beyond 6 months in this patient population.
We performed a retrospective analysis of our last 8 years of VGCV use for CMV prophylaxis using data from our center's transplant clinic database. All patients who received a lung or heart–lung transplant at our center from June 27, 2003 to May 1, 2011 were evaluated for inclusion. Study subjects were identified from the sequential list of all LTx performed at our center. Subjects were then grouped into cohorts based on CMV serostatus (D+/R-, D+/R+, D-/R+ or D-/R-). LTx recipients were excluded if they did not complete 6 months of CMV prophylaxis, if they died before 6 months post-LTx, if they had breakthrough CMV viremia (>1000 viral copies/mL) while on CMV prophylaxis, if they did not receive our center's standard immunosuppression regimen, if they had a known primary immunodeficiency or if they were at low risk for CMV disease (D-/R-). Re-transplants were also excluded because of the use of induction immunosuppression in this group.
The primary outcome measure was the proportion of LTx recipients who developed CMV infection or disease during the 180 days following completion of prophylaxis in each of the at- risk CMV serotypes (D+/R- vs. D+/R+ vs. D-/R+). A secondary outcome was CMV infection and disease free survival in all LTx recipients who completed prophylaxis, including those LTx recipients who died after completing prophylaxis but before completion of the 180-day observation period. This study was approved by the Colorado Multiple Institutional Review Board [(COMIRB) # 11–1501] and the University Hospital Human Research Review Committee.
Because augmented immunosuppression is a risk factor for CMV, we also assessed the potential impact of acute rejection (AR) on the development of CMV infection or disease . In LTx recipients who developed CMV, the incidence of AR was assessed during the 90-day period preceding the development of CMV. In those who did not develop CMV, the incidence of AR was assessed during the 90 days preceding the average day posttransplant of CMV onset among those recipients who developed CMV. This 90-day period was considered the pre-CMV observation period.
CMV prophylactic regimen
The CMV prophylactic regimen used during the entire study period is shown in Figure 1. D+/R- LTx recipients received 100 days of intravenous (IV) ganciclovir (5 mg/kg twice daily × 14 days followed by once daily) followed by completion of 6 months of CMV prophylaxis with VGCV (450 mg twice daily). CMV hyperimmune globulin was given by the following schedule in D+/R- LTx patients: 150 mg/kg on post-LTx day 1, then 100 mg/kg every 2 weeks × 4 doses, then 50 mg/kg every month × 2 doses. All R+ LTx patients received 30 days of IV ganciclovir (5 mg/kg once daily) followed by completion of 6 months of CMV prophylaxis with VGCV (450 mg twice daily). CMV hyperimmune globulin was given by the following schedule in R+ LTx patients: 150 mg/kg on post-LTx day 1, then 100 mg/kg on post-LTx day 15 and 30. If LTx recipients were treated for AR at any time following completion of 6 months of CMV prophylaxis they were prescribed VGCV 450 mg twice daily for an additional 3 weeks. Recipients requiring treatment with lympholytic agents or plasmapheresis received a single dose of CMV hyperimmune globulin (100 mg/kg) in addition to 3 weeks of VGCV. Doses of ganciclovir and VGCV were adjusted for renal function and/or toxicity.
Monitoring and diagnosis
Following completion of CMV prophylaxis, quantitative CMV polymerase chain reaction (PCR) (Roche Lightcycler, Branchburg, NJ, USA; range of positivity 501–1000 000 viral copies/mL) was performed on whole blood at all routine clinic visits (monthly) and for signs and symptoms of CMV. Asymptomatic LTx recipients with <1000 viral copies/mL were considered negative and followed routinely. CMV infection was defined as >1000 viral copies/mL. Asymptomatic LTx recipients with 1000–3000 viral copies/mL were tested weekly until the progression or resolution of viremia occurred. No changes to a recipient's CMV prophylaxis or immunosuppression regimen were undertaken for asymptomatic LTx recipients with <3000 viral copies/mL. Those with >3000 viral copies/mL were treated with antiviral therapy. In symptomatic LTx recipients, those with >1000 viral copies/mL were subsequently evaluated for tissue invasive disease. This algorithm was validated in previous studies at our center [13, 14]. No CMV-specific immune assays were utilized. All LTx recipients underwent surveillance bronchoscopy at post-LTx month 1, 2, 3, 6 and 12. In the absence of systemic anticoagulation, transbronchial biopsies were obtained with each bronchoscopy and underwent immunohistochemistry staining for CMV pneumonitis. Culture of the bronchiolar lavage fluid for CMV was performed following each bronchoscopy. In addition, LTx recipients exhibiting signs and symptoms consistent with CMV pneumonitis or AR also underwent bronchoscopy with tissue culture and immunohistochemistry. Based on the low incidence of breakthrough CMV observed in our prior experience, CMV PCRs were not routinely performed in LTx patients receiving CMV prophylaxis . Only those patients with signs and symptoms of CMV had CMV PCRs obtained while receiving CMV prophylaxis. Routine testing for hypogammaglobulinemia was not performed. CMV disease was defined as the presence of histologic evidence of tissue invasive CMV disease on biopsy of the lung, liver or other tissues, a positive tissue culture, a positive bronchoalveolar culture or a syndrome characterized by fever, leukopenia and a positive PCR.
All LTx recipients were initiated on standard triple-drug immunosuppressive therapy with cyclosporine, azathioprine and prednisone post-LTx. Recipients were switched to tacrolimus for any clinical episode of AR with a positive biopsy grade (A1 or higher), for a second episode of moderate AR on surveillance biopsy (biopsy grade A2 or higher) or for manifestations of cyclosporine toxicity (neurotoxicity or renal insufficiency) . Induction therapy was not given. LTx recipients with biopsy proven AR or suspected AR (fall in forced expiratory volume in one second [FEV1] of >10% and all other causes ruled out) received methylprednisolone (10 mg/kg once daily for 3 days; . Those with steroid resistant AR received rabbit antithymocyte globlin. LTx recipients with antibody-mediated rejection received methylprednisolone and/or plasmapheresis.
The proportion of LTx recipients who developed CMV infection or disease was compared between CMV serostatus groups using a chi-squared test or Fisher's exact test, where appropriate. Freedom from CMV infection or disease was compared between groups using the Kaplan–Meier method with log-rank statistics. The number of AR episodes occurring during the pre-CMV period was compared between groups using either a one-way analysis of variance (ANOVA) or unpaired t-test, where appropriate. Continuous variables are presented with their respective standard deviation (SD). A stepwise logistic regression was performed to determine if CMV serostatus was independently associated with the development of CMV infection or disease. Variables entered into the model included age, gender, indication for transplant, single or bilateral lung transplant, tacrolimus use, thymoglobulin use, plasmapheresis, incidence of AR and CMV serostatus. Variables having a p value of less than 0.1 were retained in the model. A Kaplan–Meier survival model was used to determine the impact of serostatus on CMV infection and disease free survival; the associated survival proportions are presented with their respective standard error (SE). Time-to-event analysis was censored at 180 days following completion of prophylaxis. A p value of less than 0.05 was considered statistically significant. Data were analyzed using MEDCalc (version 184.108.40.206 [MedCalc Software, Mariakerke, Belgium]).
Two hundred forty-four LTx recipients were evaluated for inclusion in the analysis (Figure 2). One-hundred ten LTx recipients were excluded from the analysis; 36 for early termination of prophylaxis, 31 for D-/R- CMV serostatus, 18 for death before 6 months post-LTx, 16 for lung retransplant, 5 for breakthrough viremia on prophylaxis, 3 for use of a nonstandard immunosuppressive regimen (inhaled cyclosporine in all 3 cases) and 1 for known common-variable immunodeficiency. Three additional LTx recipients who completed prophylaxis died before completing the 180-day postprophylaxis observation period; thus, 131 LTx recipients were included in the primary outcome analysis. These 3 recipients were included in the secondary endpoint analysis of CMV free survival (n = 134). Of the 5 LTx recipients who developed breakthrough CMV while on prophylaxis, 3 were D+/R-, 1 was D+/R+ and 1 was D-/R+. The incidence of breakthrough CMV in LTx recipients who completed 6 months of prophylaxis and were not otherwise excluded was 6.7% for D+/R- (3 of 45), 1.7% for D+/R+ (1 of 59) and 2.9% for D-/R+ (1 of 35). No subject included in the analysis was lost to follow-up.
Reasons for early discontinuation of prophylaxis were: leukopenia/neutropenia (n = 24), impaired renal function (n = 7), intolerance (n = 3), increased liver function enzymes (n = 1) and thrombocytopenia (n = 1). Seventeen of the LTx recipients who discontinued prophylaxis early were D+/R+, 15 were D+/R- and 4 were D-/R+.
Baseline demographics of the study cohort are shown in Table 1. The most common indication for transplant was chronic obstructive pulmonary disease followed by interstitial lung disease. Most patients received a single lung transplant (70.2%). Fifty-six (42.7%) were D+/R+, 41 (31.3%) were D+/R- and 34 (26.0%) were D-/R+.
Table 1. Baseline demographics at completion of CMV prophylaxis
n = 131
CMV = cytomegalovirus.
Female, n (%)
Indication for transplant, n (%)
Chronic obstructive pulmonary disease
Interstitial lung disease
Type of transplant, n (%)
CMV donor/recipient serostatus, n (%)
Tacrolimus, n (%)
The proportion of LTx recipients who developed CMV infection or disease differed significantly between groups during the 180 days following completion of CMV prophylaxis (20 of 41 [48.8%] D+/R- vs. 19 of 56 [33.9%] D+/R+ vs. 2 of 34 [5.9%] D-/R+, chi-squared test; p < 0.001; Table 2). When the proportion of LTx recipients who developed CMV infection or disease was compared specifically between the D+/R- and D+/R+ cohorts, there was no significant difference between these two groups (20 of 41 [48.8%] D+/R- vs. 19 of 56 [33.9%] D+/R+; Fisher's exact test; p = 0.15; Table 2). Freedom from CMV infection and disease across serotypes is shown in Figure 3.
Table 2. Development of CMV infection or disease in each at-risk serotype
D+/R- (n = 41)
D+/R+ (n = 56)
D-/R+ (n = 34)
1D+/R- versus D+/R+ versus D-/R+, p = n/a (data does not meet assumptions of chi-squared test).
2D+/R- versus D+/R+, p = 0.25 (Fisher's exact test).
3D+/R+ versus D-/R+, p = 0.08 (Fisher's exact test).
4D+/R- versus D+/R+ versus D-/R+, p < 0.001 (chi-squared test).
5D+/R- versus D+/R+, p = 0.15 (Fisher's exact test).
6D+/R+ versus D-/R+, p = 0.002 (Fisher's exact test).
CMV = cytomegalovirus; D = donor; R = recipient.
There was no significant difference in the number of AR episodes that occurred during the pre-CMV period between groups (0.54 ± 0.74 D+/R- vs. 0.5 ± 0.99 D+/R+ vs. 0.47 ± 0.71 D-/R+; ANOVA; p = 0.95). Two LTx recipients received antithymocyte globulin for treatment of steroid resistant AR during the pre-CMV period (1 D+/R- and 1 D+/R+). Despite the reintroduction of CMV prophylaxis, both of these recipients developed CMV disease. Two D+/R+ and 2 D+/R- LTx patients were treated for antibody-mediated rejection; 3 of these LTx recipients went on to develop CMV infection (2 D+/R- and 1 D+/R+).
D+/R- versus R+
When the proportion of LTx recipients who developed CMV infection or disease was compared between D+/R- and R+ LTx patients, significantly more D+/R- developed CMV infection or disease (20/41 [48.8%] D+/R- vs. 21/90 [23.3%] R+; Fisher's exact test; p = 0.005; Table 3). There was also a significant difference in the proportion of LTx recipients who developed CMV disease between these two groups (8/41 [19.5%] D+/R- vs. 6/90 [6.7%] R+; Fisher's exact test; p = 0.04). In the multivariate model, older age (odds ratio [OR], 1.05; 95% confidence interval [CI] 1.004–1.099; p = 0.03) and D+/R- CMV serostatus (OR, 3.83; 95% CI 1.674–8.770; p = 0.002) were associated with the development of CMV infection or disease. In the multivariate model which included only CMV disease as the dependent variable, CMV serostatus was the only variable retained in the model and D+/R- serostatus was significantly associated with the development of CMV disease (OR, 3.39; 95% CI 1.094–10.533; p = 0.03).
Table 3. Factors significantly associated with CMV in the multivariate model
CMV = cytomegalovirus; D = donor; R = recipient.
D+/R- versus R+
CMV infection or disease
D+/R- versus R+
D-/R+ versus D+/R+
CMV infection or disease
R+: Impact of donor serostatus
Among R+ LTx patients, those who received a D- organ developed significantly less CMV infection or disease than those who received a D+ organ (2/34 [5.9%] D-/R+ vs. 19/56 [33.9%] D+/R+; Fisher's exact test; p = 0.002). No D-/R+ LTx developed CMV disease. In the multivariate model which included only R+ LTx patients, donor CMV serostatus was the only variable retained in the model and D-/R+ serostatus was significantly associated with the absence of CMV infection and disease (OR, 0.12; 95% CI 0.0263–0.563; p = 0.007).
CMV free survival
CMV infection and disease free survival differed significantly across the three at-risk CMV serotypes (p < 0.001, log-rank test, figure not shown). Of the 3 additional LTx recipients who were included in this model, 2 were D+/R+ and 1 was D+/R-. None of these 3 recipients developed CMV infection or disease before their death. At day 180 following completion of CMV prophylaxis, the CMV free survival proportions were 51.7 (7.8) for D+/R-, 66.2 (6.3) for D+/R+ and 94.1 (4.0) for D-/R+.
In this study, the development of late-onset CMV infection or disease differed significantly based on both the donor and recipient CMV serotype following 6 months of CMV prophylaxis with a ganciclovir-containing compound and serostatus-driven CMV hyperimmune globulin. Taking into account the recently published randomized, controlled trial by Palmer et al. which investigated the optimal duration of CMV prophylaxis post-LTx, this study highlights the potential for further optimization of prophylaxis regimens based on CMV serostatus . Although some of the results of this study might have been expected based on our previous report, they suggest that CMV serostatus of both the LTx donor and recipient may identify individuals at heightened risk for CMV disease and infection post-LTx in the VGCV era .
In this analysis, the overall proportion of D+/R- LTx patients who developed late-onset CMV infection or disease during the 180 days following 6 months of CMV prophylaxis approached 50%: 29.3% developed CMV infection and 19.5% developed CMV disease. Our results would seem to confirm the findings of Palmer et al. regarding the need for an extended length of prophylaxis in D+/R- LTx patients . It is noteworthy that D+/R- LTx patients received a more intensive CMV prophylaxis regimen compared to R+ patients (Figure 1). The potential for this more intensive regimen to impede the development of CMV-specific immunity against the donor strain in these CMV naïve patients should be considered. Data derived from the LTx population suggested that development of CMV-specific immunity is not impaired in patients receiving VGCV; however, data of this nature is not available for patients receiving concomitant prophylactic CMV hyperimmune globulin. When considering whether D+/R- LTx patients included in this study could have benefited from a less intensive CMV prophylactic regimen, it is worth noting that the majority of cases of breakthrough CMV occurred in D+/R- LTx patients. Therefore, although the potential for impaired development of CMV-specific immunity in patients receiving a more intensively prophylactic regimen cannot be ruled out, our findings do not seem to support the use of a less intensive prophylactic regimen in D+/R- LTx patients.
Among R+ LTx patients, transplantation of a D+ organ was associated with a higher risk for CMV infection and disease than was transplantation of a D- organ. None of the D-/R+ LTx patients included in this analysis developed CMV disease and only 5.9% developed CMV infection. The marked difference in the observed rates of CMV infection or disease between D-/R+ and D+/R+ patient may be related to R+ LTx patients having only partial immunity to the CMV strain carried by the D+ organ . Whereas recipients of D- organs have preexisting T cell-specific immunity to protect against reactivation of their native CMV strain, many recipients of D+ organs may require several months to develop T cell-specific immunity that confers protection against CMV superinfection by the donor strain. Further, it appears that transmission and reactivation of multiple CMV strains is possible, and that episodes of CMV disease associated with multiple genomically distinct strains are linked to higher viral loads and delays in viral clearance [18, 19].
Although numerically different, the proportion of LTx recipients who developed CMV infection or disease did not differ significantly between the cohort of D+/R- and D+/R+ patients. Whether this lack of statistical difference is because of the fact that D+/R- and D+/R+ LTx recipients have a comparable risk for CMV, or because our study was underpowered to detect a difference between these two cohorts, remains to be determined. Given the retrospective nature and modest effect size observed in our study (proportion with CMV infection or disease = 48.8% for D+/R- vs. 33.9% for D+/R+) one might conclude that a greater number of subjects, perhaps from multiple centers will be required to determine if the risk for CMV is indeed comparable between these two groups.
This study has several limitations. Its observational, single-centered design limits the conclusions that can be drawn from the results presented. Although we controlled for potential confounders in the multivariate logistic regression model, unidentified confounders could have potentially impacted the differences in CMV infection and disease that were observed. Because the results of the aforementioned study by Palmer et al. were not reported by individual at-risk CMV serotypes, we are unable to benchmark our results against this randomized, controlled trial. The use of CMV hyperimmune globulin for CMV prophylaxis in our population may limit the generalizability of the study results to centers with differing protocols. Finally, because we sought to determine the occurrence of CMV infection or disease over a 180-day period only in those LTx recipients who completed 6 months of CMV prophylaxis, our findings may be limited by the select population we evaluated.
These findings suggest that in the VGCV era, CMV serostatus of both the donor and LTx recipient continue to be associated with the development of CMV disease and infection post-LTx. This further suggests that individual patient factors such as CMV-specific immunity may be the critical determinant of the duration of CMV prophylaxis and monitoring post-LTx. Our results highlight the potential importance of both the LTx donor and recipient CMV serostatus for further stratifying the risk for late-onset CMV. A randomized, controlled trial stratified by CMV serostatus or driven by the level of CMV-specific immunity will be needed to determine the optimal duration of CMV prophylaxis among each at-risk CMV serotype in lung transplant.
The authors of this manuscript have conflicts of interest to disclose as described by the American Journal of Transplantation. Kelly E. Schoeppler has received research funding from CSL Behring. Martin R. Zamora has served as a consultant for CSL Behring.