Background. Current advances in transplantation practices may influence the development of cytomegalovirus (CMV) disease after renal transplantation.
Methods. From September 2003 through February 2005, 1470 renal transplant recipients (55 of whom were kidney-pancreas transplant recipients) were prospectively studied in the 16 transplant centers affiliated with the Spanish Network of Infection in Transplantation, with use of an ad hoc–designed online database. Univariate and multivariate analyses with logistic regression were performed to detect risk factors for CMV disease.
Results. A total of 105 episodes of CMV disease (37 with visceral involvement) developed in 99 (6.7%) of 1470 patients. Attributable mortality appeared in 1 (1.0%) of 105 cases. Multivariate analysis showed that, apart from CMV serological mismatch, presence of rejection episodes, and the use of antilymphocitic drugs, a simultaneous pancreas transplantation (odds ratio [OR], 3.7; 95% confidence interval [CI], 1.5–9), use of cyclosporine (OR, 1.7; 95% CI, 1.18–2.9), a donor >60 years of age (OR, 2.3; 95% CI, 1.5–3.7), and chronic graft malfunction (OR, 1.8; 95% CI, 1.14–2.9) were independently associated with CMV disease, whereas use of sirolimus had a protective effect (OR, 0.27; 95% CI, 0.1–0.78).
Conclusions. Additional risk factors related to current transplantation practices influence the epidemiology of CMV after renal transplantation and should be taken into account in the design of prophylactic strategies in this population of kidney or kidney-pancreas recipients.
Cytomegalovirus (CMV) is still is the single most important infectious agent affecting kidney allograft recipients (KTRs) . CMV infection is a major source of morbidity  and has recently been related to increased mortality in these patients . The classic risk groups for CMV disease are CMV-seronegative transplant recipients from a seropositive donor (D+/R−) and those patients receiving antilymphocytic antibodies [4, 5]. Despite preventive measures against CMV (mostly long-term prophylaxis for patients in high-risk groups), the incidence of CMV disease has not changed over the past 10 years [6, 7]. Current advances in surgical procedures and immunosuppressive therapy could influence the incidence and characteristics of CMV disease in KTRs as has been suggested in some reports [8–11], but a precise analysis of the incidence and risk factors of CMV disease in a current large cohort of KTRs was lacking. The aim of the present study was to describe the influence of recent advances in renal transplantation practices on the incidence, clinical characteristics, and risk factors associated with CMV disease through data obtained from a large, current, prospective cohort of KTRs.
Spanish Network of Infection in Transplantation (RESITRA) network. RESITRA is a research network financed by the Spanish Ministry of Health, formed by members of 16 transplant centers and 3 reference laboratories included in the Spanish Study Group of Infection in Transplantation that prospectively includes all solid-organ transplantations performed at these hospitals. RESITRA was created in April 2003. Each hospital has a research team (i.e., physician, nurse, and technician) that manages the research tools. Pretransplant, peritransplant, and follow-up (days 0, 7, 14, 30, 60, 90, 180, 270, 360, and 720 after transplantation) data are prospectively included in the online database, as are all infections (diagnostic evaluation, clinical presentation, therapy, and outcome) and rejection episodes, always according to the predefined definitions (based on universally accepted criteria) published on the RESITRA Web site . Data collection is made through forms available on PDF, which are sent to an SQL-server database hosted on the Web site. Data generation is made through managerial and statistical databases generated from the SQL-server database, after a validating process managed by each hospital coordinator. For the present study, we included all transplant recipients (i.e., KTRs and kidney-pancreas transplant recipients [KPTRs]) from in the RESITRA prospective cohort from September 2003 through February 2005 for whom there was a minimum follow-up of 1 month.
CMV infection and disease definition. CMV disease was defined according to predefined RESITRA criteria  and included CMV viral syndrome (temperature of >38°C with a positive antigenemia test with no other cause to account for it and 1 of the following findings: leukocyte count, <4000 cells/mm3; atypical lymphocytes of ⩾3%; and platelet count, <100,000 platelets/mm3). Definite tissue-invasive disease required histopathological evidence of CMV with or without culture of the virus from a tissue sample. For patients without histopathological evidence of CMV, a diagnosis of probable end-organ CMV disease required a compatible clinical presentation, evidence of CMV antigenemia, and clinical and/or virological response to specific treatment with ganciclovir or valganciclovir. Prophylaxis against CMV in solid-organ transplant recipients was performed according to Spanish Study Group of Infection in Transplantation guidelines . Long-term prophylaxis was restricted to high-risk groups, whereas preemptive therapy was applied to patients with moderate risk. We also included episodes of asymptomatic CMV viremia that required antiviral treatment (preemptive therapy). We expressed the incidence of CMV disease in cases of infection per 100 KTRs and in cases per 1000 transplantation-days.
Analyzed variables as possible risk factors for CMV disease. Variables were divided into groups by patient characteristics and by pretransplantation, intraoperative, and postoperative variables. The preoperative period referred to the month immediately before transplantation. More than 100 variables were included. In summary, the main variables that were included were as follows. Pretransplantation variables (1 month before transplantation) included age; sex; diabetes; previous transplantation; type and severity of underlying disease; renal failure (serum creatinine, >2.5 mg/dL for >7 days) and hemodialysis; leukopenia (<3000 leukocytes/mm3); admission to the intensive care unit; CMV, hepatits B virus, and hepatits C virus serological condition; donor-recipient matching; corticosteroid, antibiotic, or antifungal therapy; donor age; and presence of donor infection. Intraoperative factors included length of transplant surgery, cold ischemia at time of graft, total number of blood-product units given during surgery, specific technical procedures (single or double renal transplant and/or pancreas transplant), and type of antibiotic prophylaxis. The following postoperative factors were analyzed: length of stay in the intensive care unit; mechanical ventilation and vascular amine requirements; reoperation (new surgical procedure); urgent retransplantation (in the first 7 days after transplantation); repeated intensive care unit stay with or without mechanical ventilation; renal failure; hemodialysis; recrudescence of the underlying disease; fungal, bacterial, and other viral infections; CMV prophylaxis and preemptive treatment; fungal prophylaxis; and Pneumocystis jirovecii prophylaxis. The immune-related factors included crossmatching for HLA compatibility, leukopenia, acute and chronic rejection, type of immunosuppressive treatment (including that administered for acute rejection episodes), and chronic graft malfunction.
Data analysis. Continuous variables were expressed as the mean (±SD) for those values with a normal distribution and as the median (mostly for those with a skewed distribution). Discrete variables were expressed as percentages.
Student's unpaired t test was used to compare continuous variables, the Mann-Whitney U test was used to compare continuous variables with nonnormal distribution, and the Χ2 or Fisher's exact test was used to compare proportions. All statistical tests were 2-tailed, and the threshold of statistical significance was P<.05.
ORs were calculated for variables with statistically significant differences between patients with or without CMV disease. Binary logistic regression was applied individually to each variable to obtain the OR in the univariate analysis. Quantitative variables were previously converted into qualitative variables for that task.
Statistically significant variables (P<.05) in the univariate analysis were introduced in a multivariate model by use of forward stepwise logistic regression, to identify the independent risk factors for CMV disease. In addition, clinically relevant factors with P values <.1 that were considered to be potential confounders (on the basis of experience and the literature) were forced into the multivariate model to investigate their effect. To test the effect of sirolimus use (during ⩾3 months) in the development of CMV disease during the first year after transplantation, censored for death or graft loss, Kaplan-Meier analysis was used, and statistical significance was estimated using the log-rank test. The statistical software SPSS, version 13.0 (SPSS), was used to perform the calculations.
Data on 1470 KTRs were prospectively included in the RESITRA cohort during September 2003 through February 2005 and were analyzed in the present study, with a median follow-up of 350 days (range, 31–785 days). Global characteristics of the study cohort are presented in table 1.
CMV prophylaxis. Prophylaxis was globally administered to 369 patients (25%): 84 (67%) of 126 patients receiving anti-CD3 monoclonal antibody (OKT3) or antithymocite globulin, 114 (86%) of 132 CMV-seronegative recipients with a seropositive donor (D+/R−), and 49 (89%) of 55 KPTRs (table 1).
Incidence and clinical characteristics of CMV disease. A total of 99 patients developed 105 episodes of CMV disease (95 in KTRs and 10 episodes in KPTRs). The adjusted incidence of CMV disease was 0.2 episodes per 1000 transplant-days (0.5 for KPTRs and 0.2 for KTRs). At least 1 episode of CMV disease occurred in 92 (7%) of 1415 KTRs and in 7 (13%) of 55 KPTRs.
Characteristics of CMV disease episodes are summarized in table 2. The median interval between transplanation and CMV disease development was 62 days (range, 21–384 days). Late-onset CMV disease—defined as occurrence >100 days after transplantation—was reported in 25 (23.8%) of 105 cases.
The majority of episodes (68 [65%] of 105 cases) were reported as viral syndrome without end-organ disease. Probable or definite end-organ disease was reported in 37 cases; gastrointestinal disease was the most common (29 [78%] of 37 cases): CMV enterocolitis (19), CMV gastritis (7), and CMV esophagitis (3). Other less common organ CMV diseases were pneumonitis (5 cases) and interstitial nephritis (3). A definite diagnosis of CMV end-organ disease that was based on histopathological findings was reported only in 7 patients (table 2).
The majority of the episodes of CMV disease (93 [89%] of 105 cases) were managed with intravenous ganciclovir. In only 12 episodes was oral valganciclovir used as primary treatment, mainly in cases of CMV viral syndrome (8 cases), but also in 3 cases of CMV gastritis and 1 case of CMV colitis. The outcome was favorable in all the patients, except that there was 1 death related to CMV pneumonitis, which implied an attributable mortality of 1% of the patients and 0.9% of the episodes. End-organ disease was more common among patients with late-onset CMV disease (11 [44%] of 25 cases) than in patients with early disease (26 [33%] of 80 cases), although differences were not statistically significant (P=.2). Outcome was favorable in all the patients with late-onset CMV disease.
Risk factors for CMV disease among KTRs. For the risk factor analysis, we first performed a univariate analysis comparing patients who developed CMV disease with the remaining cohort of KTRs who did not develop disease and with a minimum follow-up of 30 days. Therefore, a total of 1441 KTRs from the RESITRA cohort were included in that analysis (median duration of follow-up, 354 days; range, 31–785 days). Clinical variables that were significantly related to the development of CMV disease are shown in table 3. The variables that were independently related to the development of CMV disease in the final multivariate model were donor age >60 years (OR, 2.3; 95% CI, 1.5–3.7), cyclosporine use (OR, 1.7; 95% CI, 1.1–2.9), kidney-pancreas transplantation (OR, 3.7; 95% CI, 1.5–9.1), CMV (D+/R−) condition (OR, 7.3; 95% CI, 4.4–12), use of OKT3 or antithymocite globulin as induction treatment (OR, 2.14; 95% CI, 1.1–4.4), at least 1 acute rejection episode (OR, 2.7; 95% CI, 1.6–4.4), and chronic graft malfunction (OR, 1.8; 95% CI, 1.1–2.9). Sirolimus use was related to a protective effect against CMV disease (OR, 0.27; 95% CI, 0.1–0.78).
We performed a further analysis of the group of patients receiving sirolimus as primary immunosuppressive regimens over the first 3 months after transplantation. As represented in the Kaplan-Meier survival curve (figure 1), the cumulative probability of being free of CMV disease 500 days after transplantation was 97% in the group of KTRs who had received ⩾3 months of treatment with sirolimus, compared with 91% in the remaining KTR cohort (log rank; P=.027). In table 4, other comparative data between KTRs receiving and KTRs not receiving sirolimus are presented. Sirolimus was administered to KTRs who received a graft from donors who were older than donors of the comparative group (mean age, 51.2 vs. 47.7 years; P=.05), and the frequency of double transplantation was also significantly higher among KTRs who received treatment with sirolimus (7.2% vs. 1.4%; P<.001). The frequency of KTRs who experienced ⩾2 episodes of acute rejection during the follow-up period was significantly higher among patients receiving sirolimus (5.9% vs 1.8%; P=.003).
This study analyzes the problem of CMV disease in the current kidney transplantation era, in which there have been several advances in surgical techniques, available immunosuppressive regimens, and preventive measures against CMV disease. On the other hand, expanded criteria for assigning kidney-transplant candidates and their potential donors in the most recent years [14, 15] have conditioned an increasingly higher number of elder donors and recipients. The implications of those new circumstances in the incidence, clinical characteristics, and risk factors of CMV disease in KTRs are unknown. The strength of the present study resides in the large current cohort of KTRs prospectively studied, with >100 CMV disease episodes evaluated through a database specifically designed to analyze infectious complications. In addition, we were able to examine >100 clinical variables, which conferred high sensitivity to the detection of risk factors and provided enough control variables to adjust for confounding factors.
Despite the fact that a high percentage of patients from the RESITRA cohort received CMV prophylaxis for high-risk groups (i.e., D+/R− CMV status, patients receiving antilymphocytic antibodies, and KPTRs), CMV disease occurred in 6.5% of KTRs and 12.7% of KPTRs, with a global adjusted incidence of 0.2 episodes per 1000 transplant-days (0.5 episodes per 1000 transplant-days for KPTRs and 0.2 episodes per 1000 transplant-days for KTRs). Sixty-five percent of the episodes comprised CMV viral syndrome and gastrointestinal CMV disease, which was by far the most frequent clinical form of end-organ CMV disease among KTRs. The outcome was favorable in the great majority of patients. With regard to late-onset disease, although there was a trend toward a higher incidence of end-organ disease compared with early episodes, in contrast to what was suggested in some preliminary reports [16, 17], the outcome was similar in both groups, as has been discussed in other recent studies [6, 18]. The present study confirms some classic risk factors for CMV disease: D+/R− CMV status and the use of antilymphocitic antibodies [4, 19] are still major risk factors for CMV disease, even in the presence of targeted antiviral prophylaxis.
The analysis of the RESITRA cohort has also provided the identification of emerging risk groups. Grafts from donors >60 years of age are associated with a 2-fold increased risk of developing CMV disease. The higher percentage of CMV seropositivity in transplanted grafts from older donors  led to reinfections or primary infections in seropositive and seronegative KTRs, respectively , which could partially explain a higher incidence of CMV disease in these subgroups of KTRs.
This prospective analysis of the RESITRA cohort provides an accurate comparison of the effects of different immunosuppressive regimens and avoids some biases inherent in retrospective studies—for example, the influence of changes occurring over time in global transplantation practices. The results of the present study confirm that the use of cyclosporine—but not Tacrolimus—was independently related to an increased risk of CMV disease. Use of cyclosporine has been associated elsewhere with an increased risk of CMV infection, compared with Tacrolimus, in a retrospective study of patients who had received liver transplants . The results of our study confirm that not only CMV infection but also CMV disease is independently related to the use of cyclosporine by KTRs. The explanation for such differences between both anticalcineurinic drugs remains uncertain. Tacrolimus is known to be 30–100 times more potent than cyclosporine in vitro, but peak in vivo calcineurin activity inhibition has been shown to be greater with cyclosporine , which could lead to a higher effect in T cell function.
Another immunosuppressive regimen that has been related to an increased risk of CMV disease in some retrospective studies is mycophenolate [8, 10]. In our prospective study, we could not find any significant relationship between the drug and CMV disease, in accordance with another previous report .
On the other hand, the use of sirolimus showed an independent protective effect in the present study, which led to a reduction of as much as 3 times the risk of developing CMV disease. Only 1 recent retrospective case-control study, which was specifically designed to evaluate the effect of sirolimus on the development of CMV infection, has pointed out an incidence of CMV infections that was lower among KTRs treated with sirolimus than those treated with tacrolimus . In 2 randomized studies designed to evaluate efficacy and tolerance of different immunosuppressive regimens, a lower incidence of CMV infection was reported among patients receiving sirolimus [26, 27], although an adjustment of this effect with other clinical variables was not done. Nevertheless, other similar studies could not demonstrate such finding [28, 29]. The basis of this probable protective effect is not clear. Mammalian target of rapamycin (mTOR) is a kinase that is a key regulator for protein synthesis in cells . Intracellular viruses such as CMV are heavily dependent on the cellular protein synthesis machinery to support synthesis of their constituent proteins and genomic replication, so it is conceivable that inactivation of this kinase by rapamycin can impair CMV viral replication . This fact is supported by a recently published anecdotal study reporting the efficacy of adding sirolimus to ganciclovir in treatment of 8 KTRs with ganciclovir-resistant CMV infection . Conversely, the fact that KTRs from the present study who received sirolimus experienced multiple acute rejection episodes more frequently than did KTRs who received other immunosuppressant therapies (table 4) suggests that sirolimus has a less potent immunosuppressive effect that could lead to a lower risk of opportunistic infection, including CMV disease.
The outcome of the graft also seems to be relevant in the development of CMV disease. Acute rejection is a well-known risk factor of CMV disease among KTRs [19, 33]. Patients with chronic graft malfunction also have an increased risk of developing CMV disease, in view of our results. Renal failure is a recognized risk factor for opportunistic invasive fungal infection in solid-organ transplant recipients , although there are also recent reports that suggest that renal failure could also be related to an increased risk of CMV disease . The basis for this relationship has not been completely explained. Some recent studies have determined that formation of functional CMV-specific CD8+ effector-memory CD4+ T cells are crucial for controlling CMV viral replication [36, 37], and low T cell proliferation and cytokine profile alterations are known to be influenced by both uremia and renal replacement therapy [38, 39]. Conversely, the presence of renal failure increases the difficulty of controlling serum levels of immunosuppressive regimens, which could lead to an increased net state of immunodepression.
Some drawbacks of the present study deserve specific consideration. End-organ CMV disease was considered to be definite in a minority of patients, because the diagnosis was rarely based on histopathological criteria. On the other hand, the design of the study makes it is difficult to demonstrate causality in some of the detected risk factors for CMV disease; therefore, such variables should be used exclusively to identify a high-risk population of solid-organ transplant recipients.
In conclusion, CMV disease is still a major infectious complication for KTRs, despite prophylaxis treatment for classic high-risk patients, although the great majority of the episodes—including late-onset cases—have a favorable prognosis. Our results suggest that directed prophylaxis should be broadened to patients receiving a graft from elderly donors and those receiving cyclosporine. KTRs with acute rejection complications or those with chronic graft dysfunction should be carefully followed up with sequential CMV viremia detection, to prevent CMV disease. Finally, new strategies are needed to prevent late-onset CMV disease in patients receiving prophylaxis, especially CMV D+/R− pairs and KPTRs.
Financial support. Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III- FEDER, Spanish Network of Infection in Transplantation (RESITRA G03/075), Spanish Network for the Research in Infectious Diseases (REIPI RD06/0008), and Spanish Study Group of Infection in Transplantation.
Potential conflicts of interest. All authors: no conflicts.