We investigated relationships between cytomegalovirus (CMV) seropairing and CMV prophylaxis on graft outcome in recipients of solid organ transplants. Transplants carried out from 1985 to 2002 and reported to the Collaborative Transplant Study were analyzed. In cadaver kidney recipients, CMV prophylaxis was significantly associated with improved graft survival only in the seronegative-recipient/seropositive-donor combination (at 3 years: 79.4% with prophylaxis vs. 73.5% without prophylaxis; RR 0.80, p < 0.0001). Among patients who had a functioning graft at 1 year, significantly fewer patients who received CMV prophylaxis received rejection treatment in the preceding year (26.3%), compared with patients who did not receive prophylaxis (32.4%) (p = 0.0001), suggesting an inhibitory effect of CMV prophylaxis on acute rejection. Significant improvements in graft survival after CMV prophylaxis were found also in CMV-negative recipients of CMV-positive heart, and lung or heart-lung transplants, but not liver transplants. The age of the recipient had a differential effect on graft and patient survival after CMV prophylaxis. Use of antilymphocyte antibodies or mycophenolate mofetil was not associated with an enhanced CMV effect on graft outcome. These results may contribute to a better understanding of the influence of pretransplant CMV serology on the effect of CMV prophylaxis.
Cytomegalovirus (CMV) disease is a major cause of morbidity and mortality in solid organ transplantation and affects greater than 50% of transplant recipients (1). Generally, CMV infection occurs within the first 3 months of transplantation. Approximately 60% of healthy adults are asymptomatic CMV carriers, as evidenced by seropositivity for antibodies to CMV, and iatrogenic transmission can occur from donated solid organs to seronegative recipients (2). Seropositive recipients are at risk of reactivation of a latent virus by immunosuppressive therapy, as well as of reinfection by a virus in the allograft. The most important risk factor for development of CMV disease is the serostatus of the donor and recipient. Symptomatic disease occurs most frequently in seronegative recipients of an organ from a seropositive donor (3). The type and intensity of immunosuppression also has a strong influence on the risk of developing CMV disease, with the use of anti-T-cell antibodies being associated with increased incidence and severity of CMV disease (4–6).
Cytomegalovirus infection and disease may influence both graft function and survival and also patient survival. It has been difficult to examine whether CMV infection causes allograft rejection or whether rejection in CMV-infected patients is a consequence of reduced immunosuppressive therapy in patients treated for symptomatic CMV disease. A mechanism by which CMV may promote rejection is the triggering of inflammatory processes including increased expression of adhesion molecules, cytokines, and HLA class II antigens, especially in endothelial cells (2,7). Some clinical studies demonstrate an association between CMV infection and acute rejection episodes in renal transplant recipients (8–10), whereas other studies have failed to show this correlation (11,12). Although conflicting results have been obtained regarding the relationship between CMV seropairing and acute rejection or graft survival in renal transplantation (13–15), in general it appears that seronegative recipients of a seropositive kidney have worse graft and patient outcomes and are associated with increased costs (14).
An implication of these findings is that CMV prophylaxis may improve graft and patient survival, but there is no consensus on either the necessity or the efficacy of CMV prophylaxis. Current prophylactic approaches vary considerably among transplant centers because of the absence of large multicenter randomized trials evaluating the efficacy of various regimens, variable protocols used in single-center studies, and differences among the types of solid organ transplants. Nevertheless, prophylaxis with anti-CMV immunoglobulin, acyclovir, and gancyclovir has been shown to reduce CMV-associated morbidity among recipients of renal and other transplants (15,16). In some single-center studies, CMV prophylaxis had beneficial effects on graft function (17,18), but a meta-analysis that included only prospective randomized studies showed no effect of prophylactic treatment on allograft rejection, loss, or transplant-related death (12). In general, prophylactic antiviral therapies have not been studied in sufficiently large cohorts after solid organ transplantation to assess their effects on graft outcome.
To examine whether CMV prophylaxis influences graft outcome after solid organ transplantation, we have investigated relationships between CMV seropairing and CMV prophylaxis on graft and patient survival in recipients of solid organ transplants drawn from the Collaborative Transplant Study (CTS). The large database allows an analysis of the variables affecting CMV prophylaxis that can be used to guide therapeutic decisions in transplant recipients.
Transplants reported to the CTS by 435 centers from 44 countries were included in the analysis (19). Information on the pretransplant CMV serostatus of recipients and donors was recorded at the time of initial registration, shortly after transplantation. Clinical follow-up information was requested at 3, 6 and 12 months, and yearly thereafter. The transplants were performed between 1985 and 2002. Only those patients whose pretransplant CMV serostatus and that of the donor were known were included. Information on CMV prophylaxis was available for a subset of these patients. Demographics of kidney recipients with or without CMV prophylaxis are shown in Table 1. The distribution of factors was similar in the two groups, with the exception of transplant region. Whereas approximately 60% of patients in both subgroups were transplanted in Europe, the fraction of patients transplanted in North-America was greater, and that of patients transplanted in other continents smaller among patients with prophylaxis than among patients without prophylaxis. The factor ‘region’ was included as a covariable in the multivariate analysis.
Table 1. Demographics of patients with or without cytomegalovirus-prophylaxis
First cadaver kidney
+Prophylaxis n = 12 117
−Prophylaxis n = 24 165
HLA-A + B + DR
The main outcome measures were graft survival and patient survival. Functional graft survival was also determined after censoring patients dying with a functioning graft. In renal transplant patients, graft loss was determined by return to dialysis or death. In patients with other transplants, graft loss was determined by removal of the transplant or death. In addition to functional graft survival, which was considered an approximation of the rate of acute rejection, the impact of CMV prophylaxis on acute rejection was assessed indirectly by determining the incidence of patients who required antirejection therapy in the first year among those with a functioning graft at the end of year 1.
The Kaplan-Meier method was used to estimate graft and patient survival and results are indicated as mean percentage ± SE. For statistical evaluation of differences in outcome between patients with or without CMV prophylaxis, multivariate analysis was performed using a Cox regression model. The covariables recipient and donor age, gender, and race, year of transplantation, geographical location of transplant center (continent), original disease leading to organ failure and necessitating transplantation, cold ischemic preservation time, percent panel-reactive antibodies, HLA-A, -B, -DR mismatches, and immunosuppressive protocol (intent to treat) were considered. Although, as shown in Table 1, donor age was nearly identical in patients with or without CMV prophylaxis, there was, as expected from the epidemiology of CMV, a difference between the age of CMV-positive donors (mean 38.6 year) and CMV-negative donors (mean 33.2 year). Particular care was used to ensure that the results of the analysis were not unduly influenced by this factor. Cox regression results are indicated as relative risk (RR) and 95% confidence interval (CI), and a p-value < 0.05 was considered significant.
Effect of CMV seropairing on graft and patient outcome after CMV prophylaxis in cadaver kidney transplants
In recipients of first cadaver kidney transplants who received no CMV prophylaxis, graft survival was highest in seronegative recipients of a kidney from a seronegative donor (81.7%± 0.6% at 3 years) and poorest in seronegative recipients with a seropositive donor (73.5%± 0.8%) (Figure 1). Compared with the recipient negative/donor negative group, Cox regression analysis showed a relative risk (RR) of 1.40 (CI 1.26–1.54) for the recipient-negative/donor-positive group (p < 0.0001), RR 1.13 (CI 1.04–1.23) for the recipient positive/donor positive group (p = 0.0031), and 1.14 (CI 1.04–1.25) for the recipient positive/donor negative group (p = 0.0054).
A strong association of CMV prophylaxis with graft survival was found only in the seronegative-recipient/seropositive-donor combination (79.4%± 0.6% with prophylaxis vs. 73.5%± 0.8% without prophylaxis). Multivariate Cox regression analysis showed a significantly reduced relative risk of graft failure in patients receiving CMV prophylaxis: RR 0.80 (CI 0.73–0.89; p < 0.0001) (Figure 2A). In this group, CMV prophylaxis also was associated with improved functional graft survival (84.9%± 0.5% with prophylaxis vs. 81.2%± 0.8% without prophylaxis; RR 0.86 (CI 0.78–0.95; p = 0.0017), as well as patient survival (92.0%± 0.4% with prophylaxis vs. 88.1%± 0.6% without prophylaxis; RR 0.71 (CI 0.61–0.83; p < 0.0001). In CMV-seronegative recipients of a second graft from a CMV-seropositive donor, CMV prophylaxis was associated with improved 3-year graft survival from 64.9%± 2.7% to 75.5%± 1.9% (RR 0.74, CI 0.57–0.97; p = 0.0309) (Figure 3).
Among first transplants from CMV-positive donors to CMV-positive recipients, we noted a small improvement effect of CMV prophylaxis on 3-year graft (79.0%± 0.7% vs. 77.2%± 0.4%) and patient survival (90.0%± 0.5% vs. 88.6%± 0.3%), however, this difference did not reach statistical significance (Figure 2D). No effect of CMV prophylaxis was seen in the positive-to-positive retransplant group (not shown).
The introduction of more effective immunosuppressive regimens in the last decade has improved transplant outcomes. To control for this effect, the year of transplantation was included in the multivariate analysis and, in addition, we performed a separate analysis of CMV prophylaxis in renal transplants performed from 1997–2002. The positive effect of CMV prophylaxis in CMV-seronegative recipients of kidneys from CMV-seropositive donors was also present in transplants performed during recent years: For first cadaver kidney transplants performed between 1997 and 2002, 3-year graft survival rates were 84.0%± 0.9% and 76.3%± 1.9%, respectively, in 2205 patients with CMV prophylaxis and 674 patients without CMV prophylaxis (RR 0.68, CI 0.55–0.83; p = 0.0003).
Among CMV-negative recipients of a CMV-positive kidney who had a functioning graft at 1 year, significantly fewer patients who received CMV prophylaxis received rejection treatment in the preceding year (26.3%), compared with patients who did not receive prophylaxis (32.4%, p < 0.0001). Combined with the improved functional graft survival rate (which provides an approximation of the rate of immunological graft loss) observed in patients receiving CMV prophylaxis, these data suggest an inhibitory effect of CMV prophylaxis on acute rejection. In addition, among patients who did not receive CMV prophylaxis, the incidence of rejection treatment was significantly increased only in the recipient-negative/donor-positive group of patients (p < 0.0001 vs. each of the other three groups) (Figure 4) which, together with the graft survival data shown in Figure 1, indicates that an increased incidence of acute rejection contributed to reduced graft survival in patients with this CMV serocombination.
The effect of CMV prophylaxis in the recipient-negative/donor-positive serocombination was not confined to cadaver kidney recipients and was also observed in recipients of a first related donor kidney transplant. In these patients, 3-year graft survival in patients who received prophylaxis was higher (87.5%± 1.2%, n = 905) compared with those who were not treated (82.9%± 1.7%, n = 575; RR 0.71, CI 0.53–0.96, p = 0.0243). There was also a significant improvement in functional graft survival (90.3%± 1.1% vs. 87.2%± 1.5%; RR 0.68, CI 0.52–0.88, p = 0.0040), but not in patient survival (95.9%± 0.7% vs. 94.5%± 1.0%; RR 0.71, CI 0.41–1.21, p = NS).
Effect of CMV prophylaxis in other organ transplant recipients
In heart transplant patients, as in renal transplant recipients, a statistically significant association of CMV prophylaxis with graft outcome was evident only in the seronegative-recipient/seropositive-donor group, with an improvement in graft survival after prophylaxis (71.1%± 1.6% vs. 78.7%± 1.0% at 3 years; RR 0.71, CI 0.60–0.84, p = 0.0001) (Figure 5). Significantly fewer patients who received CMV prophylaxis were treated for acute rejection (46.1%) in the preceding year, compared with patients who did not receive prophylaxis (55.6%) (p = 0.0037). Paralleling the findings in renal recipients, the effect of CMV prophylaxis in the seropositive-recipient/seropositive-donor group was much less pronounced: 73.6%± 0.9% graft survival at 3 years without prophylaxis (n = 2412) vs. 76.3%± 1.2% with prophylaxis (n = 1473; RR 0.85, CI 0.73–0.98, p = 0.0240). In recipients of lung or heart-lung transplants, graft survival in the recipient-negative/donor-positive group was also significantly better after CMV prophylaxis: RR 0.53, p < 0.0001 (Table 2). However, in liver transplant recipients, in contrast to other transplant groups, CMV serostatus was not associated with graft survival; CMV prophylaxis in the recipient-negative/donor-positive group was not significantly associated with graft survival either (Table 2). We noted a reduction in the rate of treated rejection in the first year from 27.1% to 17.5%, however, the difference did not reach statistical significance (p = 0.12).
Table 2. Graft and patient survival and rejection treatment in cytomegalovirus (CMV)-seronegative recipients of CMV-seropositive donor organs
Type of transplant
3-year graft survival* (%± SE)
Rejection treatment** (%)
+ CMV prophylaxis
− CMV prophylaxis
+ CMV prophylaxis
− CMV prophylaxis
*Relative risk (RR) and p-value obtained by Cox regression.
**Rate of rejection treatment during the first post-transplant year in patients with a functioning graft at 1 year. The information was available on a subset of patients.
79.4 ± 0.6
73.5 ± 0.8
(n = 5426)
(n = 2908)
(n = 2368)
(n = 1276)
87.5 ± 1.2
82.9 ± 1.7
(n = 905)
(N = 575)
(n = 322)
(n = 260)
78.7 ± 1.0
71.1 ± 1.6
(n = 1679)
(n = 851)
(n = 757)
(n = 340)
58.7 ± 2.3
44.7 ± 4.9
(n = 544)
(n = 109)
(n = 95)
(n = 14)
68.1 ± 2.4
73.4 ± 2.8
(n = 429)
(n = 297)
(n = 114)
(n = 70)
Effect of type of CMV prophylaxis on graft outcome
To compare different prophylactic agents in relation to graft outcome, we examined results from 1995 to 2002. This time period was chosen in order to avoid bias arising from use of immunoglobulin predominantly in earlier years, widespread use of acyclovir and gancyclovir only in recent years, and improved graft outcomes (unrelated to CMV prophylaxis) in recent years. As in all multivariate analyses contained in this report, year of transplantation was included as a covariable also in this subset analysis in order to exclude the possibility that recent improvements in graft outcomes resulted in erroneous associations. A comparison of the efficacy of acyclovir, gancyclovir, and anti-CMV immunoglobulin between 1995 and 2002 showed the three agents to be associated with improved graft survival to approximately the same degree in cadaver kidney or heart recipients (Figure 6). For kidney recipients, Cox regression analysis showed a relative risk (RR) of 0.65 (CI 0.52–0.82) (p = 0.0002) for gancyclovir, RR 0.65 (CI 0.50–0.85) (p = 0.0013) for acyclovir, and RR 0.53 (CI 0.41–0.70) (p < 0.0001) for immunoglobulin. For heart transplants, the corresponding values were RR 0.60 (CI 0.43–0.83) (p = 0.0019) for gancyclovir, RR 0.42 (0.19–0.92) (p = 0.0290) for acyclovir, and RR 0.65 (CI 0.42–0.99) (p = 0.0454) for immunoglobulin.
Immunosuppression and CMV prophylaxis
Because immunosuppression with MMF (20,21) or induction treatment with ATG/OKT3 (4–6) are considered risk factors for CMV disease, we examined the effect of these agents on graft survival after CMV prophylaxis. The 1995–2002 period was chosen for the reasons outlined earlier. Our analysis showed that graft outcome in CMV-negative recipients of CMV-positive kidneys who received prophylaxis was improved to a similar extent in patients who received an MMF-containing regimen (5.1% improvement of the graft survival rate at 1 year, RR 0.69) vs. those who received an azathioprine-containing regimen (5.9%, RR 0.75). Cytomegalovirus prophylaxis was associated with a similar enhancement of graft survival in patients who received ATG or OKT3 (6.6%, RR 0.65) as compared with patients not treated with antibodies (6.6%, RR 0.67).
Age-related effect of CMV prophylaxis
The age of the recipient emerged as an influencing variable for the association of CMV prophylaxis with transplant outcome in recipient-seronegative/donor-seropositive transplants (Table 3). In pediatric recipients of a cadaver renal transplant, CMV prophylaxis was associated with improved functional graft survival (83.9% with prophylaxis vs. 73.5% without prophylaxis, RR 0.64, CI 0.45–0.89; p = 0.0079) but not patient survival (95.7% with prophylaxis vs. 97.0% without prophylaxis, p = NS), whereas in patients older than 60 years, the main effect of CMV prophylaxis was on patient survival (83.1% with prophylaxis vs. 71.4% without prophylaxis, RR 0.57, CI 0.41–0.79; p = 0.0008). Cytomegalovirus prophylaxis had an increasing effect with age on patient survival in other age groups. A similar effect of age was noted in cardiac transplant recipients, in whom graft survival essentially equates with patient survival (Table 3).
Table 3. Effect of age on functional graft survival and patient survival in cytomegalovirus (CMV)-seronegative recipients of CMV-seropositive donor organs
Age strata (years)
3-year functional graft survival1,2 (%± SE)
3-Year patient survival2 (%± SE)
1Patients dying with a functioning graft were censored.
2Relative risk (RR) and p-value obtained by Cox regression.
83.9 ± 2.0
73.5 ± 3.5
95.7 ± 1.1
97.0 ± 1.3
(n = 402)
(n = 169)
82.9 ± 0.9
80.4 ± 1.2
95.7 ± 0.5
93.1 ± 0.8
(n = 2134)
(n = 1211)
86.7 ± 0.8
82.7 ± 1.1
89.7 ± 0.7
85.6 ± 1.0
(n = 2340)
(n = 1232)
86.7 ± 1.7
83.8 ± 2.4
83.1 ± 1.9
71.4 ± 2.9
(n = 521)
(n = 289)
81.6 ± 3.0
80.7 ± 5.5
(n = 178)
(n = 53)
81.8 ± 2.4
76.4 ± 3.2
(n = 318)
(n = 190)
79.7 ± 1.3
70.8 ± 2.1
(n = 968)
(n = 517)
72.0 ± 3.5
59.7 ± 5.4
(n = 190)
(n = 86)
A role for prophylaxis with antiviral agents or anti-CMV immunoglobulin in controlling the direct infectious manifestations of CMV in solid organ transplant recipients has been established (22,23). However, few studies have formally addressed the effect of CMV prophylaxis on indirect consequences of CMV replication such as acute rejection and graft loss. Our large-scale retrospective analysis shows that CMV prophylaxis is associated with increased graft and patient survival in recipients of kidney, heart, and lung transplants. In general, the survival benefit of CMV prophylaxis was confined to the seronegative-recipient/seropositive-donor population. In patients with a functioning graft at the end of 1 year, CMV prophylaxis was also associated with a reduced incidence of rejection treatment during the preceding year, and in kidney recipients the analysis of death-censored functional survival showed a lower graft loss rate in patients who received CMV prophylaxis.
Several retrospective and prospective studies have demonstrated that CMV infection is a risk factor for poor graft outcome (8,9,24,25), whereas other studies have failed to demonstrate this relationship (26) or suggested that CMV infection may be a consequence of treatment of rejection rather than a cause of rejection (11). The reality could encompass both possibilities: CMV could trigger rejection in the recipient-negative/donor-positive combination, leading to intensified immunosuppression, which in turn could facilitate progression to clinical CMV disease. Several mechanisms may account for a link between CMV infection and acute rejection. In addition to having a direct cytopathic effect in renal parenchymal and endothelial cells (27), CMV infection leads to up-regulation of adhesion molecules, increased expression of HLA antigens, and activation of cytotoxic T cells, which could enhance alloantigen recognition, increase cellular infiltration of the allograft parenchyma, and trigger acute rejection (2). Cytomegalovirus -seronegative patients who receive an allograft from a seropositive individual are a population generally considered to be at high risk for symptomatic CMV disease (28). Our study shows that graft and patient survival were worse in patients receiving a kidney from a seropositive than a seronegative donor, with the effect most pronounced in seronegative recipients. Similar findings were obtained in another study that examined the relationship between CMV antibody status and graft outcome (14), whereas other investigators reported that donor serology had the greatest influence on outcome, regardless of recipient serology (15,29,30).
Our demonstration that CMV prophylaxis is significantly associated with improved graft and patient survival in the donor-positive/recipient-negative group of patients underlines the enhanced transplant risk posed by CMV infection in this patient serogroup. Our results provide indirect evidence to support a role for CMV infection in acute rejection. An examination of the survival curves reveals that the effect of CMV prophylaxis on functional graft survival occurs mainly within the first 6 months of transplantation. The 1-year functional graft survival rate was approximately 5% higher and there was a reduction in first-year rejection treatment of patients with a functioning graft at the end of 1 year in the CMV prophylaxis group, which suggests that the decreased graft failure rate resulted from a reduction in acute rejection by CMV prophylaxis. The impact of CMV prophylaxis only in this serocombination is similar to the finding from a recent study demonstrating that prophylaxis with valacyclovir significantly reduced the incidence of graft rejection only in seronegative recipients of a seropositive kidney transplant, although there was no improvement of 12-month survival rates (17).
The age of the patient had a differential influence on transplant outcome after CMV prophylaxis. Whereas in pediatric kidney recipients CMV prophylaxis was associated with a significant reduction in functional graft loss but with no impact on patient survival, this effect gradually reversed with increasing age, so that in patients aged older than 60 years, CMV prophylaxis was associated with substantially improved patient survival but had a negligible impact on functional graft survival. Because of a more intact immune system, younger recipients may experience higher graft rejection rates after CMV infection, yet respond well to CMV prophylaxis, be able to control infection, and reduce progression to disease. By contrast, a weaker immune system in older recipients may cause less graft rejection but increase susceptibility to progressive CMV disease. Our findings in pediatric renal transplant recipients are similar to results from a registry study of pediatric patients hospitalized for CMV infection, in whom CMV prophylaxis led to significantly improved 3-year graft survival, whereas patient survival did not differ from the at-large renal transplant population (30).
Considerable clinical and experimental evidence indicates that CMV infection accelerates the development of acute and chronic rejection in cardiac transplants (31–33). Our data show significantly improved graft survival after CMV prophylaxis of seronegative cardiac transplant recipients receiving a heart from a seropositive donor. In these patients, as in kidney recipients, significantly fewer patients who received CMV prophylaxis were treated with antirejection therapy, again suggesting a reduction of rejection episodes with CMV prophylaxis. In lung or heart/lung transplants, CMV prophylaxis was associated with a graft survival advantage exceeding that observed in renal and heart transplants. The lung is rich in cells that harbor latent or replicating CMV and may provide the donor with a higher initial CMV load, which may then be subject to reactivation (34). Cytomegalovirus prophylaxis may also attenuate the risk of obliterative bronchiolitis, a manifestation of chronic rejection after lung transplantation that has been associated with CMV infection (35).
Graft and patient outcome after liver transplantation remained unaffected by CMV prophylaxis although the prevalence and manifestation of CMV disease are similar after liver and other organ transplants. Although rejection treatment was reduced by 11% in liver recipients who received CMV prophylaxis, this difference did not reach statistical significance. The lack of demonstrable benefit of CMV prophylaxis on graft survival in liver transplants may perhaps be explained by the regenerative capacity of the liver that allows it to recover from the damage caused by rejection without significant long-term consequences. We also found that liver recipients were the only transplant group in which transplant outcome without CMV prophylaxis was similar in all CMV serogroups. In a recent study, prophylactic use of anti-CMV immunoglobulin in liver recipients protected against severe CMV disease in all serologic groups except the donor-positive/recipient-negative group; graft and patient survival were not influenced by prophylaxis (23).
Clinical manifestations of CMV infection are exacerbated by increased immunosuppression, particularly the use of antilymphocyte antibodies for induction therapy; CMV-seropositive transplant recipients treated with these antibodies are considered to be a high-risk population requiring CMV prophylaxis (36). Additionally, MMF-containing immunosuppressive regimens have been linked with a higher incidence of CMV disease (20,21). In this study, the use of antibodies or MMF did not appear to confer an added risk for graft rejection or death, as our analysis showed that the graft survival advantage associated with CMV prophylaxis in CMV-seronegative patients transplanted with a CMV-seropositive kidney was similar in patients who received ATG/OKT3 or MMF, compared with patients who were not treated with these immunosuppressive agents.
This is the first study to show prolongation of graft and patient survival after CMV prophylaxis. Acyclovir, gancyclovir, and anti-CMV immunoglobulin were associated approximately to the same degree with improved survival in seronegative recipients of renal and cardiac transplants from a seropositive donor. Inclusion of the year of transplantation as a covariable in the multivariate analysis and the use of more recent (1995–2002) data when all three anti-CMV agents were in use allowed us to compare these agents and prevented a bias of poorer survival rates in patients transplanted in earlier years. However, we cannot exclude the possibility that center preference for the administration of a certain prophylactic agent for certain patients considered at risk of CMV infection may have evened out any possible advantages of a particular agent. Previous studies showed that prophylaxis with high-dose acyclovir (37), gancyclovir (38), CMV immunoglobulin (23), or valcyclovir (17) did not improve 1-year graft or patient survival, although these agents reduced symptomatic CMV disease, and valacyclovir also reduced acute rejection by 50%. A recent meta-analysis of CMV prophylaxis also demonstrated that use of acyclovir or gancyclovir was associated with a significant decrease of CMV infection and disease but not of the risk of acute rejection, graft loss, or transplant-related death (12).
Some limitations of our study must be recognized. Because this was a registry analysis, we were unable to document the actual occurrence of CMV disease and unable to demonstrate direct causality between CMV infection or disease and transplant outcome. As a corollary, the data also did not allow us to determine the extent to which CMV prophylaxis reduced CMV disease. On the other hand, the use of seromatching as the basis of comparison in this study removed potential bias arising from differences in sensitivity and specificity of post-transplant diagnostic methods, timing, periodicity of monitoring, and definition of infection and disease.
In conclusion, our retrospective analysis showed that CMV prophylaxis with antiviral agents or anti-CMV immunoglobulin was significantly associated with improved graft and patient survival in CMV-negative kidney, heart, or lung transplant recipients who received a graft from a CMV-positive donor. Our data suggest that CMV serology influences the efficacy of CMV prophylaxis in mitigating transplant outcome. Together with data demonstrating the significant economic impact of CMV serostatus in transplant recipients (14), our results highlight the importance of minimizing the detrimental consequences of CMV infection in organ transplantation, especially in CMV-seronegative recipients of a CMV-seropositive organ transplant.
The support of staff members at 435 transplant centers in 44 countries participating in the Collaborative Transplant Study who provided information on the CMV serostatus of transplant recipients and donors is gratefully acknowledged.