Summary of findings
Description of the condition
Cytomegalovirus (CMV) infection is an important cause of morbidity and mortality in solid organ transplant recipients, with up to 75% of patients developing or reactivating infection after transplantation (Fishman 2007). The risk of developing CMV is well established. CMV infection and disease are seen most commonly in CMV negative recipients of CMV seropositive organs and in recipients treated with anti-lymphocyte therapy (EBPG 2000). Recipients who have a positive serostatus for CMV, regardless of donor status, are at intermediate risk of developing or re-developing CMV; donor negative organs and negative serostatus recipients are at the lowest risk of infection (Fishman 2007).
CMV can lead to direct or indirect effects of infection. Direct effects include CMV syndrome (encompassing fever, myelosuppression, myalgia and arthralgia), tissue invasive CMV disease leading to hepatitis, gastroenteritis, pneumonitis and potentially any other organ or body tissue, and finally, mortality (Eid 2010). CMV infection may also be associated with indirect effects including acute and chronic organ rejection, graft loss, opportunistic infections other than CMV, and new onset diabetes mellitus (Eid 2010; Kotton 2010).
Description of the intervention
Pre-emptive therapy involves the detection of active CMV replication through routine surveillance using an appropriately sensitive and specific test (such as real-time polymerase chain reaction (PCR) or pp65 antigenaemia assay) with established thresholds for CMV infection. Once the CMV infection threshold has been reached, treatment with antiviral medications (valganciclovir or ganciclovir) is initiated with the aim of preventing progression to symptomatic CMV disease (Humar 2009).
How the intervention might work
The intervention aims to direct treatment to patients who are most at risk of developing CMV disease, rather than administering universal prophylaxis that would include some patients who would never develop CMV disease. By targeting higher risk patients it should be possible to direct therapy to those in need, to reduce the risk of adverse effects associated with long-term antiviral therapy, and possibly, to lower the cost of therapy. Pre-emptive therapy may also reduce the numbers of patients who develop late onset CMV disease after prophylaxis is ceased (Eid 2010).
Why it is important to do this review
Pre-emptive treatment of recipients with CMV viraemia detected on surveillance has been advocated as an alternative to antiviral prophylaxis because only recipients who develop CMV infection, and are thus at high risk of CMV disease, are treated. This approach exposes patients to a lower risk of adverse effects of medications and may reduce the risk of the emergence of resistant strains of CMV and of late onset CMV disease, which has been reported as increasing in frequency in solid organ transplant recipients (Arthurs 2007; Arthurs 2008; Eid 2010; Kotton 2010). Late onset CMV disease is seen in about 10% of patients given 12 weeks of oral ganciclovir prophylaxis but is rare following pre-emptive regimens.
It has been argued that pre-emptive treatment in CMV negative recipients of CMV positive organs allows controlled viral replication to occur before antiviral medications are administered (Limaye 2000). This could result in the development of specific immune responses which are important in the prevention of future episodes of CMV disease.
Pre-emptive therapy appears particularly attractive in recipients at low risk of CMV disease such as CMV positive recipients of kidney transplants who are not receiving antibody immunosuppression because their risk of CMV disease is only about 7% (Waiser 1998). However, pre-emptive therapy relies on the timely availability of sensitive and reliable methods for detecting CMV viraemia so that treatment based on identifying CMV infection can be initiated before CMV disease develops. Furthermore, the cost of screening may be higher than prophylaxis even though the cost of antiviral medication is lower (Eid 2010). Pre-emptive therapy potentially exposes patients to periods of CMV infection and possibly indirect effects of CMV infection. Moreover, treatment method relies on the patient's compliance with regular screening and easy access to a transplant centre if treatment is required.
Prophylaxis is currently the predominant intervention used to prevent CMV disease. Routine antiviral prophylaxis is associated with an increased risk of adverse effects of medications, the development of CMV resistance and late onset CMV disease when prophylaxis is discontinued and is costly (Arthurs 2008; Emery 2001; Hart 2001). Prophylaxis for all organ transplant recipients (except CMV negative recipients of CMV negative donors) means that a proportion of patients will receive potentially harmful medications when they were not destined to develop CMV disease.
In view of the actual and potential advantages and disadvantages of pre-emptive therapy and of prophylaxis to prevent CMV disease, a systematic review was warranted to ascertain the efficacy of pre-emptive therapy and the relative benefits and harms of these therapies for preventing CMV disease in solid organ transplant recipients.
This Cochrane systematic review should be considered in conjunction with the Cochrane review Antiviral therapy for preventing CMV disease in solid organ transplants recipients (Hodson 2013) which looked at the benefits and harms of prophylaxis with antiviral medications to prevent CMV disease. Hodson 2013 includes 37 studies (4342 participants). Prophylaxis with ganciclovir, valaciclovir or aciclovir (19 studies, 1981 participants) significantly reduced the risk of CMV disease, CMV infection and all-cause mortality compared with placebo or no specific therapy. In direct comparison studies, ganciclovir was significantly more effective than aciclovir (7 studies, 1113 participants). Valganciclovir was as effective as ganciclovir (1 study, 364 participants) and extended duration valganciclovir significantly reduced the risk of CMV disease compared with three months (2 studies, 454 participants).
The aim of this review was to evaluate the benefits and harms of pre-emptive treatment of CMV viraemia to prevent CMV disease, all-cause mortality and the indirect effects of CMV infection (acute rejection, graft loss, opportunistic infections) in solid organ transplant recipients and to assess the comparative effects of pre-emptive treatment regimens and routine CMV prophylaxis with antiviral medications.
Criteria for considering studies for this review
Types of studies
We included randomised controlled trials (RCTs) and quasi-RCTs (trials in which allocation was obtained by alternation, alternate medical records, date of birth or other predictable methods) only.
Types of participants
Studies enrolling any type of solid organ transplant recipient, adult or paediatric, were included.
Types of interventions
We included RCTs of pre-emptive treatment compared with placebo or standard care, pre-emptive treatment compared with antiviral prophylaxis, and different pre-emptive treatment regimens (different antiviral agents used for pre-emptive treatment, different doses, different routes of administration) in solid organ transplant recipients.
Types of outcome measures
The effects of these interventions were tested on the following outcomes:
- all-cause mortality
- death due to CMV disease
- CMV disease
- time to development of CMV disease
- graft loss
- acute rejection
- other infections
- adverse effects of medications.
For the purpose of analysis, pre-emptive treatment was defined as routine testing for CMV viraemia using any test and commencing antiviral treatment if viraemia was detected. Prophylaxis was defined as the regular administration of an appropriate antiviral medication for a period of time immediately post transplantation.
Other definitions for the meta-analysis were comparable with those reported previously by Ljungman 2002. CMV infection was defined as isolation of CMV in any tissue or body fluid by an appropriately sensitive and specific test. CMV DNAemia was defined as detection of CMV DNA in whole blood, plasma, peripheral blood leukocytes or buffy coat specimens by an appropriately sensitive and specific technique such as real time polymerase chain reaction (RT-PCR). CMV disease was defined as CMV infection together with CMV syndrome (fever and bone marrow suppression) and/or tissue invasive CMV confirmed on histopathology.
Search methods for identification of studies
Original review (2005)
We searched the following resources without language restriction.
- The Cochrane Renal Group's Specialised Register
- The Cochrane Central Register of Controlled Trials (CENTRAL in The Cochrane Database of Systematic Reviews issue 2, 2005).
- MEDLINE (1966 to February 2005).
- EMBASE (1980 to February 2005).
The Trials Search Co-ordinator ensured that all relevant studies had been identified. Additional studies were located through article reference lists and from abstracts from international meetings.
Review update (2013)
For this update, we searched the Cochrane Renal Group's Specialised Register (to 16 January 2013) through contact with the Trials' Search Co-ordinator using search terms relevant to this review.
The Cochrane Renal Group’s Specialised Register contains studies identified from:
- Quarterly searches of the Cochrane Central Register of Controlled Trials (CENTRAL)
- Weekly searches of MEDLINE OVID SP
- Handsearching of renal-related journals and the proceedings of major renal conferences
- Searching of the current year of EMBASE OVID SP
- Weekly current awareness alerts for selected renal journals
- Searches of the International Clinical Trials Register (ICTRP) Search Portal and ClinicalTrials.gov.
Studies contained in the specialised register are identified through search strategies for CENTRAL, MEDLINE, and EMBASE based on the scope of the Cochrane Renal Group. Details of these strategies, as well as a list of handsearched journals, conference proceedings and current awareness alerts, are available from the Specialised Register section of information about the Cochrane Renal Group.
See Appendix 1 for search terms used in strategies for this review.
Data collection and analysis
This systematic review was originally undertaken by four authors (GS, JC, EH, CJ) and was published in The Cochrane Database of Systematic Reviews in 2005 (Strippoli 2005). This update was undertaken by five authors (DO, AW, GS, KK, EH).
Selection of studies
Two authors independently screened titles and abstracts retrieved from the searches and identified those studies that met the inclusion criteria. This process favoured over-selection in order to include all relevant studies. The full article was retrieved if uncertainty existed or when the abstract was not available. Any disagreement with article selection was resolved through discussion and consultation.
Data extraction and management
Two authors independently extracted data from eligible studies using standardised data extraction forms. Studies reported in foreign language journals were translated before data extraction. Participant characteristics (number, age, sex, co-morbidities), interventions (type of treatment, dose, duration, co-interventions) and primary and secondary outcome measures were recorded. Authors were contacted to obtain missing information on allocation concealment. Any discrepancies in data extraction were resolved in discussion. Where results of a study were published more than once, the most complete data were extracted from all sources and used in the analysis only once.
Assessment of risk of bias in included studies
- Was there adequate sequence generation (selection bias)?
- Was allocation adequately concealed (selection bias)?
- Was knowledge of the allocated interventions adequately prevented during the study (detection bias)?
- Participants and personnel
- Outcome assessors
- Were incomplete outcome data adequately addressed (attrition bias)?
- Are reports of the study free of suggestion of selective outcome reporting (reporting bias)?
- Was the study apparently free of other problems that could put it at a risk of bias?
Measures of treatment effect
Dichotomous outcomes were expressed as risk ratios (RR) with 95% confidence intervals (CI). Risk differences (RD) with 95% CI were calculated for adverse effects. Continuous outcomes were calculated as mean differences (MD) with 95% CI.
Unit of analysis issues
If available, data for the first period of cross-over studies were to be included in meta-analyses; otherwise, cross-over studies were reported in the text only.
Dealing with missing data
Study authors were contacted for information on sequence generation, allocation concealments and for missing data. Where missing data were few and not thought likely to influence results, the available data were analysed.
Assessment of heterogeneity
Heterogeneity was analysed using a Chi² test on N-1 degrees of freedom, with an alpha of 0.05 used for statistical significance and with the I² test (Higgins 2003). I² values of 25%, 50% and 75% correspond to low, medium and high levels of heterogeneity.
Assessment of reporting biases
This updated review included all studies identified in the Cochrane Renal Group's Specialised Register, which is revised regularly with published and unpublished reports identified in congress proceedings. This reduces the risk of publication bias. All reports of a single study were reviewed to ensure that all outcomes were reported to reduce the risk of selection bias.
Data were pooled using a random-effects model to calculate a summary estimate of effect.
Subgroup analysis and investigation of heterogeneity
To determine whether there was any difference between study results due to plausible effect modifiers, subgroup analysis was planned provided that sufficient numbers of studies for analysis were identified. The analysis aimed to explore the effects of patient characteristics such as type of solid organ transplanted, type of intervention, dose and duration of intervention, level of pre-existent risk, timing and methods used for diagnosis of CMV infection, time to graft loss, HHV6/7 status, and quality of study on treatment effect.
Where a study's results differed considerably from other studies in a meta-analysis, exclusion of the study was investigated to determine whether this altered the result of the meta-analysis.
Description of studies
Results of the search
The literature search for the original review consisted of a combined search of MEDLINE, EMBASE, the Cochrane Central Registry of Controlled Trials (CENTRAL) and the Cochrane Renal Group's Specialised Register. There were 1930 articles identified. Of these, 1737 were excluded because they were not RCTs or were RCTs that evaluated ineligible interventions. Full-text assessment of 193 potentially eligible articles identified 10 studies (14 reports; 476 participants) (Brennan 1997a Kidney; Jung 2001 Kidney; Koetz 2001 Kidney; Paya 2002 Liver; Queiroga 2003 Kidney; Rayes 2001 Liver; Sagedal 2003 Kidney; Singh 2000 Liver; Singh 1994 Liver; Yang 1998 Kidney). Four study authors responded to queries about uncertainties on study methods (Brennan 1997a Kidney; Jung 2001 Kidney; Singh 2000 Liver; Singh 1994 Liver).
For this update, only the Cochrane Renal Group's Specialised Register was searched. This is updated continuously and contains all new kidney and related studies, reports and articles. The search identified 20 reports of nine studies. Of the nine studies, five were new studies of pre-emptive therapy versus prophylaxis (Gerna 2008 Liver; Khoury 2006 Kidney; Kliem 2008 Kidney; Reischig 2008 Kidney; Witzke 2012 Kidney). Two additional reports were identified in the search but had already been included in the original review (Singh 1998; Yang 1996). A further two studies (Qiu 2008 Kidney; Tian 2005 Kidney) were identified from a systematic review of RCTs conducted by Zhang 2011. Further assessment of these studies, including contact with the authors facilitated through the Chinese Cochrane Centre, indicated that participants were not randomised, hence these studies were excluded. One study author responded to queries about uncertainties on study methods (Khoury 2006 Kidney).
There are three ongoing studies that are potentially relevant to this review (NCT00372229; NCT00966836; NCT01552369). When concluded these studies will be assessed for inclusion in a future update of this review. Another study relevant to this review was also identified for inclusion (Scott 2011 Liver), however, more information has been requested from the authors before the study can be included in meta analysis.
The combined search results are presented in Figure 1.
|Figure 1. Combined search results for the original and updated reviews. Reasons for exclusions are provided in text|
The combined updated study data included 1098 participants from 15 studies. Three intervention rationales were investigated by these studies.
One rationale involved randomising participants at transplant to receive viral surveillance and pre-emptive treatment on development of CMV viraemia versus prophylaxis with antiviral medications for 30 to 168 days (8 studies, 785 participants: Gerna 2008 Liver; Jung 2001 Kidney; Khoury 2006 Kidney; Kliem 2008 Kidney; Queiroga 2003 Kidney; Reischig 2008 Kidney; Singh 1994 Liver; Witzke 2012 Kidney). (This was the only rationale in previous iterations of this review to which new study data were added for this update).
A second rationale involved randomising participants to pre-emptive treatment or to placebo or no specific therapy (6 studies, 291 participants: Brennan 1997a Kidney; Koetz 2001 Kidney; Paya 2002 Liver; Rayes 2001 Liver; Sagedal 2003 Kidney; Yang 1998 Kidney). In five of these studies (Koetz 2001 Kidney; Paya 2002 Liver; Rayes 2001 Liver; Sagedal 2003 Kidney; Yang 1998 Kidney), all transplant recipients were screened for CMV viraemia. Study participants who developed viraemia were then randomised to receive pre-emptive treatment or placebo/standard care. In the study by Brennan 1997a Kidney, transplant recipients were randomised at transplant to receive pre-emptive treatment (screening and treatment of those with positive viraemia) or standard care (treatment of symptomatic CMV infection when it developed). Screening occurred at the time of transplantation and at weekly intervals thereafter.
The third rational type was applied in a study by Singh 2000 Liver. It investigated 22 patients who developed CMV viraemia, and were randomised to receive oral or intravenous ganciclovir as pre-emptive treatment.
All studies investigated ganciclovir or valganciclovir. Six studies compared ganciclovir with placebo or delayed treatment (Brennan 1997a Kidney; Koetz 2001 Kidney; Paya 2002 Liver; Rayes 2001 Liver; Sagedal 2003 Kidney; Yang 1998 Kidney); five compared pre-emptive ganciclovir with prophylactic ganciclovir (Gerna 2008 Liver; Jung 2001 Kidney; Kliem 2008 Kidney; Queiroga 2003 Kidney; Singh 2000 Liver); two compared pre-emptive valganciclovir with prophylactic valganciclovir (Khoury 2006 Kidney; Witzke 2012 Kidney); one study compared pre-emptive valganciclovir with prophylactic valaciclovir (Reischig 2008 Kidney); and one compared pre-emptive ganciclovir with prophylactic aciclovir (Singh 1994 Liver).
Follow-up duration of all studies ranged from three to 18 months.
Overall, we excluded 1976 studies. The 2005 review excluded 1916 reports, 33 reports were assessed and excluded in 2010, and 27 reports in this update. Most exclusions were made following assessment of title and abstract appraisal. The most common reasons for exclusion were for non-randomisation of participants or were for RCTs that investigated interventions outside the inclusion criteria for this review.
Risk of bias in included studies
Assessing risk of bias was problematic because many details were difficult to ascertain or not provided (see Characteristics of included studies). Study authors were contacted for clarification of study outcomes where reporting was inadequate (see Figure 2; Figure 3).
|Figure 2. Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies|
|Figure 3. Risk of bias summary: review authors' judgements about each risk of bias item for each included study|
Allocation concealment was considered to be at low risk of bias in four studies (Khoury 2006 Kidney; Kliem 2008 Kidney; Paya 2002 Liver; Witzke 2012 Kidney); while the remaining studies did not report methods used to conceal the randomisation process.
Risk of bias in random sequence generation bias was considered low in five studies (Khoury 2006 Kidney; Kliem 2008 Kidney; Paya 2002 Liver; Reischig 2008 Kidney; Singh 1994 Liver). While the majority of other papers reported randomisation of patients, the method used to generate a random sequence was not reported. The risk of bias in random sequence generation was high in one study (Brennan 1997a Kidney) (last digit of medical number).
Blinding of participants and personnel was reported in Koetz 2001 Kidney and Paya 2002 Liver. However, only Paya 2002 Liver provided detailed information to indicate that the study was at low risk of performance bias. Neither study reported blinding or provided information on whether outcome assessors were blinded to intervention groups. Of the 15 included studies, five (Gerna 2008 Liver; Kliem 2008 Kidney; Reischig 2008 Kidney; Sagedal 2003 Kidney; Witzke 2012 Kidney) were reported to be open-label. These, and the remaining eight studies, were assessed as being at high risk of bias for performance and detection bias because interpretation of the clinical outcome of CMV disease by clinical features could be affected by lack of blinding of participants, investigators and outcome assessors.
Incomplete outcome data
Brennan 1997a Kidney had missing outcome data due to losses to follow-up. However, this was considered unlikely to influence the results of this study. Bias due to incomplete outcome data was considered low for all other studies.
Bias through selective reporting was considered a low risk in eight studies (Brennan 1997a Kidney; Jung 2001 Kidney; Khoury 2006 Kidney; Kliem 2008 Kidney; Rayes 2001 Liver; Reischig 2008 Kidney; Sagedal 2003 Kidney; Witzke 2012 Kidney). In the remaining seven, selective reporting bias was considered high as they did not report outcomes of importance for this review. In the study of Gerna 2008 Liver graft loss, adverse effects and opportunistic infections were not reported. In the studies by Koetz 2001 Kidney and Yang 1998 Kidney all-cause mortality, graft loss and acute rejection were not reported. In Paya 2002 Liver, all-cause mortality and graft loss were not reported. Acute rejection was not reported in either Queiroga 2003 Kidney or Singh 1994 Liver, and Singh 2000 Liver did not report graft loss or acute rejection.
Other potential sources of bias
Five studies reported pharmaceutical sponsorship (Khoury 2006 Kidney; Kliem 2008 Kidney; Paya 2002 Liver; Sagedal 2003 Kidney; Witzke 2012 Kidney) and were judged as high risk of bias. Several other studies reported sponsorship from educational and government organisations (Brennan 1997a Kidney; Gerna 2008 Liver; Reischig 2008 Kidney; Yang 1998 Kidney). These studies were considered to be at low risk of bias. The remaining studies did not state any form of sponsorship and their risk of bias was unclear (Jung 2001 Kidney; Koetz 2001 Kidney; Queiroga 2003 Kidney; Rayes 2001 Liver; Singh 1994 Liver; Singh 2000 Liver).
Effects of interventions
See: Summary of findings for the main comparison Pre-emptive medication for cytomegalovirus (CMV) viraemia compared to placebo/no treatment to prevent CMV disease in solid organ transplant recipients; Summary of findings 2 Pre-emptive medication compared to prophylaxis for cytomegalovirus (CMV) viraemia to prevent CMV disease in solid organ transplant recipients
Pre-emptive treatment for CMV infection versus placebo or standard care
There were 1393 patients screened for entry to the six studies of pre-emptive treatment versus placebo or standard care. Of these, 1035 patients were excluded (no CMV viraemia in 597, CMV viraemia below threshold for study entry in 116, other reasons in 322) so that 358 patients were eligible for study entry. However, 64 patients (17.8%) developed CMV disease within the 0 to 10 day time gap between testing for viraemia and randomisation, and thus were excluded. The proportion of patients in each study who developed CMV disease before CMV viraemia was detected varied from 0% to 32%. In addition, six (1.7%) other patients were excluded after randomisation for protocol violations, so that 288 patients were evaluated.
Compared with placebo or standard care, pre-emptive treatment significantly reduced the risk of CMV disease ( Analysis 1.1 (6 studies, 288 participants): RR 0.29, 95% CI 0.11 to 0.80). The heterogeneity (I² = 54%, P = 0.06) was explained by the study undertaken by Brennan 1997a Kidney, which was the only study in which transplant recipients were randomised to screen (and treat) for CMV viraemia versus no screening, rather than randomising participants with CMV viraemia to treatment or not. Removing this study resulted in homogenous results (I² = 0%, P = 0.54). There was no significant difference for the outcome of CMV disease in studies that used oral ganciclovir compared with studies that used intravenous ganciclovir preparations (P = 0.93 for interaction).
For the outcomes of CMV organ involvement ( Analysis 1.2 (5 studies, 217 participants): RR 0.41, 95% CI 0.06 to 2.63) or CMV associated symptoms ( Analysis 1.3 (5 studies, 217 participants): RR 0.28, 95% CI 0.06 to 1.21), the summary estimates favoured treatment, but the CIs were wide. There was also no significant difference in the risks of acute rejection ( Analysis 1.4 (3 studies, 185 participants): RR 1.21, 95% CI 0.69 to 2.12), all-cause mortality ( Analysis 1.5.1 (3 studies, 176 participants): RR 1.23, 95% CI 0.35 to 4.30), graft loss ( Analysis 1.5.2 (1 study, 36 participants): 0.28, 95% CI 0.01 to 5.35), leucopenia ( Analysis 1.7.1 (2 studies, 114 participants): RR 1.54, 95% CI 0.16 to 15.36), or kidney dysfunction ( Analysis 1.7.2 (1 study, 36 participants): RR 0.93, 95% CI 0.18 to 4.92). There was no significant variation in treatment effect for any these outcomes when studies that used oral and intravenous ganciclovir regimens were considered separately.
Pre-emptive treatment for CMV viraemia versus anti-viral prophylaxis
Symptomatic CMV disease occurred at rates between 0% and 28.7% in the pre-emptive group and 0% and 29.2% in the prophylaxis group. The risk of developing CMV disease was not significantly different among groups ( Analysis 2.1 (7 studies, 753 participants): RR 1.02; 95% CI 0.43 to 2.44).
There was considerable heterogeneity among the seven studies (I² = 67%). Of these, two favoured pre-emptive therapy and two favoured prophylaxis. Removal of studies by Khoury 2006 Kidney and Singh 1994 Liver from the analysis, which both favoured pre-emptive therapy, decreased heterogeneity considerably (I² = 26%) and a clinically significant result favouring prophylaxis was observed (5 studies, 608 participants: RR 2.19, 95% CI 1.3 to 4.23). Investigation of the Singh 1994 Liver study indicated that an inferior agent (aciclovir) was used in the prophylaxis arm compared with a superior agent (ganciclovir) in the pre-emptive arm (Hodson 2013) which may have exerted a favourable influence on pre-emptive treatment. Investigation of the Khoury 2006 Kidney study did not yield any substantial differences to account for the reduction in heterogeneity.
It was found that alternately excluding the two studies (Kliem 2008 Kidney; Witzke 2012 Kidney) that favoured prophylaxis, the relative risk supported pre-emptive therapy, although the result was not significant (5 studies, 319 participants: RR 0.42; 95% CI 0.16 to 1.11).
CMV infection was significantly more common in the pre-emptive group compared with the prophylaxis group ( Analysis 2.2 (7 studies, 727 participants): RR 2.06, 95% CI 1.44 to 2.96). However, there was considerable heterogeneity (1² = 71%).
There was no significant difference observed between the pre-emptive and prophylaxis groups for all-cause mortality ( Analysis 2.3.1 (7 studies, 753 participants): RR 1.19, 95% CI 0.56 to 2.51), graft loss ( Analysis 2.3.2 (7 studies, 753 participants): RR 1.07, 95% CI 0.41 to 2.82), and acute rejection ( Analysis 2.4 (6 studies, 693 participants): RR 1.23, 95% CI 0.75 to 2.03). No heterogeneity existed among the studies for all-cause mortality. Some degree of heterogeneity existed in the graft loss analysis (I² = 41%) and acute rejection (I² = 44%) analyses. Heterogeneity was considerably diminished (I² = 14%) in the graft loss analysis when Jung 2001 Kidney was excluded. The only differences observed between Jung 2001 Kidney and all other studies were a later start for prophylaxis and a higher dose of oral ganciclovir (3000 mg/d compared with 1500 mg to 2000 mg/d).
Heterogeneity was abolished entirely in the acute rejection analysis when the study by Witzke 2012 Kidney was excluded from the analysis. The only difference that existed between the Witzke 2012 Kidney study and others was the non-inclusion of high risk transplant recipients (D+/R-).
No significant difference was identified between pre-emptive therapy and prophylaxis for infections other than CMV including bacterial ( Analysis 2.5.1 (2 studies, 168 participants): RR 0.89, 95% CI 0.55 to 1.43), viral ( Analysis 2.5.2 (1 study, 70 participants): RR 1.57, 95% CI 0.92 to 2.70), and fungal infections ( Analysis 2.5.3 (1 study, 70 participants): RR 1.89, 95% CI 0.18 to 19.89).
Leucopenia was significantly less common in patients who underwent pre-emptive therapy compared with prophylaxis ( Analysis 2.6.1 (6 studies, 729 participants): RR 0.42, 95% CI 0.20 to 0.90). Some heterogeneity existed among the studies (I² = 45%) which was considerably diminished when the Kliem 2008 Kidney study was excluded from the analysis (I² = 15%). No identifiable differences in the Kliem 2008 Kidney study were observed in comparison with others in this subgroup analysis. No significant difference was observed for neurological dysfunction between the therapies ( Analysis 2.6.2 (3 studies, 187 participants): RR 0.58, 95% CI 0.17 to 1.96).
Serostatus stratification of CMV disease was reported in two studies (Khoury 2006 Kidney; Reischig 2008 Kidney). No significant differences in symptomatic CMV disease or CMV infection were observed between pre-emptive treatment and prophylaxis for high risk transplant recipients (D+/R-) ( Analysis 2.7). For lower risk transplant recipients (D+ or D-/R+), there was no significant difference in symptomatic CMV disease ( Analysis 2.8.1) but CMV infection was significantly less common in recipients receiving prophylaxis ( Analysis 2.8.2 (2 studies, 129 participants): RR 2.07, 95% CI 1.25 to 3.42).
Oral versus intravenous ganciclovir for pre-emptive treatment of CMV viraemia
We found that 22/72 (31%) liver transplant recipients undergoing surveillance developed CMV viraemia and entered a study comparing oral and intravenous ganciclovir for pre-emptive treatment (Singh 2000 Liver). There were no significant differences in the risk of CMV disease ( Analysis 3.1), all-cause mortality ( Analysis 3.2) or other infections ( Analysis 3.3) between the regimens.
No data were available on the outcomes of death due to CMV disease. The time to development of CMV disease outcome was reported in four studies (Gerna 2008 Liver; Khoury 2006 Kidney; Kliem 2008 Kidney; Reischig 2008 Kidney). Moreover, the different methods of measurement used to report these outcomes (median, mean and frequencies) prevented meta-analysis.
Subgroup analyses according to organ transplanted, antiviral medication, duration of treatment, timing and methods used for diagnosis of CMV infection, time to graft loss, HHV6/7 status or methodological quality were not possible because of the small number of studies and enrolled patients.
Summary of main results
This review identified eight studies that compared pre-emptive with prophylactic therapy (784 patients) and six studies that compared pre-emptive therapy versus placebo or standard care (288 patients) where the primary outcome was CMV disease. Tests and thresholds used to detect CMV viraemia were different across the studies, but results were generally consistent. Confidence intervals were wide across most outcomes, indicating considerable imprecision ( Summary of findings for the main comparison; Summary of findings 2). Fewer patients were evaluated for other outcomes, and CIs around observed treatment effects were so wide that other benefits (or harms) of pre-emptive therapy could not be excluded.
Pre-emptive treatment using standard detection methods for CMV viraemia was 71% more effective than placebo or standard care in reducing the risk of CMV disease (RR 0.29, 95% CI 0.11 to 0.80). CMV organ involvement was 59% lower in the pre-emptive therapy regime compared with the placebo/standard care regime. However, the quality of evidence was low due to small patient numbers and lack of blinding across all studies. No significant differences in the risks of all-cause mortality, acute rejection, graft loss, other infections, leucopenia, or kidney dysfunction were demonstrated between pre-emptive therapy and placebo or standard care ( Summary of findings for the main comparison).
There was no significant difference in preventing CMV disease using pre-emptive versus prophylactic therapy. However, there was significant heterogeneity among studies so that superiority of one or other treatment could not be excluded. The major benefit observed for pre-emptive therapy was a 58% reduction in leucopenia compared with prophylaxis, while other adverse effects did not differ significantly, or were not reported. There were no significant differences in the risks of all-cause mortality, graft loss, acute rejection and infections other than CMV. There was also no significant difference observed between pre-emptive therapy and prophylaxis with regards to CMV disease in high risk solid organ transplant patients (D+/R-) ( Summary of findings 2).
Only one study of oral versus IV pre-emptive regimens was included in this review; it showed no significant difference in the risk of CMV disease between the groups.
It should be emphasised that oral ganciclovir, which was the basis for three of the included analyses, is no longer available and has been replaced by valganciclovir.
Overall completeness and applicability of evidence
Overall, the evidence from the 15 studies included in this review prevented confident conclusions to be made on the efficacy of pre-emptive therapy in preventing CMV disease, graft loss, acute rejection and all-cause mortality in comparison with placebo/no specific treatment or with prophylaxis. This was primarily due to a lack of adequately powered studies and low precision of the measured outcomes. There was also a lack of studies that compared pre-emptive therapy across a broad range of different transplanted organs; the highest representation was for kidney transplant recipients among the 15 studies included in this review.
Many studies did not address important outcomes, including adverse effects of medications. In contrast, the evidence base for routine prophylaxis compared with placebo/no specific therapy is substantial, arising from 19 studies of 1981 participants (Hodson 2005; Hodson 2013) which enabled more precise results to be derived.
Prophylaxis has been shown to reduce CMV disease by about 60% (RR = 0.42), with tight CIs (0.34 to 0.52), strong evidence of statistically significant benefit (P < 0.0001), and considerable homogeneity of results across all studies (I² = 13%, where only one of the point estimates of the 19 studies did not favour prophylaxis (Hodson 2005; Hodson 2013).
Antiviral prophylaxis has been shown to reduce CMV-related mortality, all-cause mortality, and clinically important disease caused by opportunistic infections.
Although both prophylaxis (Hodson 2013) and pre-emptive therapy significantly reduce CMV disease compared with placebo or no specific therapy in solid organ transplant recipients, this review has demonstrated that the available data evaluating pre-emptive therapy (6 studies; 288 participants) was of low quality (GRADE) compared with the high quality data (GRADE) evaluating antiviral prophylaxis (19 studies; 1981 participants).
There appeared to be little evidence that the use of medications now considered to be less effective than valganciclovir or ganciclovir (Humar 2009; Pescovitz 2007) influenced the results in studies comparing pre-emptive therapy with prophylaxis. Two studies (117 participants) of eight studies used aciclovir (Singh 1994 Liver) or valaciclovir (Reischig 2008 Kidney) in the prophylaxis arm of the studies. This would potentially favour pre-emptive therapy and could have contributed to the results of the Singh 1994 Liver study. All studies comparing pre-emptive therapy with placebo or no specific therapy used ganciclovir, which has been superseded by valganciclovir for prophylaxis or pre-emptive therapy. RCT data have shown no significant difference in efficacy for prophylaxis of CMV disease between ganciclovir and valganciclovir (Hodson 2005) so it can be presumed that valganciclovir would be more effective than placebo in prophylaxis or pre-emptive therapy studies, although this has not been formally tested in RCTs.
There was also substantial difference among studies with regard to CMV surveillance testing. The type of test used (pp65 antigenaemia assay or PCR DNA), the assays used, the frequency of surveillance (weekly to monthly), the cut-off values (DNA PCR > 400 to > 2000 copies/mL) to define CMV infection and the blood component used for testing were factors contributing to these differences. This large variability among studies further reduced confidence of drawing definite conclusions about the applicability of pre-emptive therapy for several reasons. Longer time intervals between surveillance tests potentially increases the chances of CMV infection not being detected so that the patient develops CMV disease before he or she can be entered in a study. This means that the proportion of preventable CMV disease patients in the pre-emptive arm may have erroneously increased. Lower cut-off values used for diagnosing CMV infection with more frequent testing could result in earlier detection of CMV infection and enable earlier initiation of therapy to prevent CMV disease and reduce the effects of indirect CMV infection. Alternatively, higher cut-off values with longer periods between tests could result in delayed therapy and potentially increase the number of preventable CMV disease cases. Finally, PCR testing needs consideration as to what component of blood is being tested; whole blood will often yield higher rates of CMV DNA compared to plasma (Humar 2009).
Quality of the evidence
This systematic review identified 15 studies of 1094 participants; of these, 784 participants were evaluated in studies that compared pre-emptive therapy with prophylaxis; and 288 were evaluated in studies that compared pre-emptive therapy with placebo/no specific treatment. The 15 studies failed to report all relevant outcomes, were frequently at high risk of bias, demonstrated imprecision and heterogeneity among studies which lowered the integrity of the meta-analysed results.
The poorly reported methods of randomisation and allocation concealment have prevented an accurate assessment of the risk of selection bias: only three studies adequately performed both. The lack of allocation concealment in these studies introduced the potential to overestimate outcome benefits (Hewitt 2005). The blinding of outcome assessors and personnel was poorly performed in most studies; only one reported blinding of participants and investigators. The blinding of outcome assessors was not conducted in any study because all study authors were judged to also be outcome assessors. Therefore, most included studies were at high risk of both performance and detection bias. The risk of attrition bias was considered to be low in all studies. Half of the studies did not include outcomes that were assessed to be important by the review authors.
Pre-emptive therapy compared with placebo/no specific therapy significantly reduced CMV disease (RR 0.29, 95% CI 0.11 to 0.80). However, CIs were wide (0.11 to 0.80), statistical significance was moderate (P = 0.02), and there was evidence of significant heterogeneity among studies (I² = 54%): two studies favoured pre-emptive therapy and two large studies favoured prophylaxis. Similarly, no significant difference was identified in the risk for CMV disease between pre-emptive therapy and prophylaxis (RR 1.02, 95% CI 0.43 to 2.44); CIs were wide, indicating considerable imprecision, and there was significant heterogeneity (I² = 63%).
Data for outcomes other than CMV disease were more limited. Particularly, we found that there was failure to report outcomes relating to adverse effects. We also found that low participant numbers and few events increased imprecision.
Only two studies stratified CMV disease and infection by the high risk serostatus D+/R- (Khoury 2006 Kidney; Reischig 2008 Kidney). Only 49 patients were available from these two studies which prevented comprehensive analysis. Such subset analysis required a larger sample size to enable accurate assessment of this high risk group.
Overall, the quality of the evidence was considered to be low or very low because of small numbers of enrolled participants, few events, significant heterogeneity, lack of blinding, and inadequate allocation concealment ( Summary of findings for the main comparison; Summary of findings 2).
Potential biases in the review process
The search criteria, analysis, and data extraction were performed to a high standard (see Methods). Four reviewers independently performed the analysis and data extraction; any conflict was resolved by an independent party. This further decreased the risk of bias for the review. The search strategy for the review was thorough. The Cochrane Renal Group's Specialised Register includes complete listings of current studies and is regularly updated to include new studies and additional reports of existing studies from journals and conference reports.
It is possible that some studies may have been missed where the study report has only been published in abstract form in conference proceedings, which have not yet been handsearched.
Overall, this review and review search represent an up-to-date assessment of current pre-emptive therapy studies.
Agreements and disagreements with other studies or reviews
Zhang 2011 conducted a systematic review comparing pre-emptive therapy and prophylaxis in solid organ transplants, and consistent with our review findings, also found no significant difference in the risk of CMV disease, graft loss, acute rejection and mortality between pre-emptive therapy and prophylaxis. However, our review identified a 58% reduced risk of leucopenia in the pre-emptive therapy rationale which was not identified by Zhang 2011. The studies included in this review for pre-emptive treatment versus prophylaxis were mostly the same as those analysed by Zhang 2011; the exceptions were two Chinese studies included by Zhang 2011 that were excluded from our review on the basis that they were not RCTs (Qiu 2008 Kidney; Tian 2005 Kidney). Zhang 2011 did not include either the Gerna 2008 Liver transplant study in children or the liver transplant study by Singh 1994 Liver which were included in this review. This would account for numerical variations between these reviews, although much of the statistical outcomes remain the same.
A recent commentary on CMV in solid organ transplant recipients also suggested the need for more studies comparing pre-emptive therapy with prophylaxis to identify if pre-emptive therapy may increase the risk of the indirect effects of CMV infection (Humar 2009). This is in agreement with our findings where several studies did not report on outcomes relating to indirect effects of CMV infection. Our findings are consistent with recent guidelines from the British Transplantation Society (BTS 2011), Kidney Disease Improving Global Outcomes (KDIGO 2009) and Caring for Australians with Renal Impairment (CARI 2010). A summary of these guidelines are presented in Table 1.
Implications for practice
This systematic review identified no significant differences in efficacy between pre-emptive therapy and prophylaxis for preventing CMV disease, graft loss, and death in solid organ transplant recipients. However, study data remain sparse. The ability to ascertain true benefits and harms remains problematic when compared with prophylaxis due to heterogeneity between study results together with low sample sizes and event rates leading to imprecision.
Pre-emptive therapy has shown benefits over placebo and standard care. This review has identified an increased risk of CMV infection with pre-emptive therapy, which may have deleterious impacts on the indirect effects of CMV infection including graft loss, acute rejection, and infections other than CMV compared with prophylaxis. While no significant differences in meta-analyses could be demonstrated for these outcomes, heterogeneity around the results indicate that differences cannot be excluded. Furthermore, four year follow-up data reported by Reischig 2008 Kidney suggest that graft loss may be higher in patients treated with prophylaxis compared with pre-emptive therapy. Four year follow-up data for the much larger Witzke 2012 Kidney study are awaited.
This review highlights the need to standardise testing frequency and cut off levels for positive tests used in the surveillance testing for CMV infection which could reduce variability between institutions and in studies. A standardised method should test patients frequently (especially within the first six months post-transplant) and should have a low threshold for detection. The results should be immediately available to enable therapy to be commenced as soon as possible after the surveillance threshold is exceeded.
Implications for research
In view of the heterogeneity of results in currently available studies, further well designed and adequately powered studies are required to compare three or more months of prophylactic therapy using an anti-CMV viral medication with pre-emptive therapy for patients who develop viraemia. Such studies should be powered to enable assessment of direct and indirect effects of CMV infection, adverse effects of medications and resource utilisation have sufficient follow-up to assess graft loss and mortality, and include standardised thresholds and frequency of testing for CMV viraemia.
The original review was co-published in Transplantation (Strippoli 2006b).
We are indebted to Narelle Willis, Managing Editor of the Cochrane Renal Group, Linda Heslop, Gail Higgins and Ruth Mitchell, Trial Search Co-ordinators of the Cochrane Renal Group, and Sandra Puckeridge of the Centre for Kidney Research for their assistance in the conduct of this review.
The authors would like to thank the referees for their editorial advice during the preparation of this review.
We also gratefully acknowledge the contributions of authors of previous version of this review: Cheryl Jones and Jonathan Craig,
Data and analyses
- Top of page
- Summary of findings [Explanations]
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Sources of support
- Index terms
Appendix 1. Electronic search strategies
Appendix 2. Risk of bias assessment tool
Last assessed as up-to-date: 16 January 2013.
Protocol first published: Issue 1, 2005
Review first published: Issue 1, 2006
Contributions of authors
2013 review update
- DO, AW, GS, KK and EH contributed to the data extraction, quality assessment, data analysis and rewriting of the review update.
- EMH identified and extracted data from included studies, contacted authors, analysed and interpreted the results and wrote the manuscript.
- CAJ conceived, designed and developed the protocol and search strategy for the review, identified and extracted data from included studies and participated in revision of the manuscript.
- GFMS checked the analysis and interpretation of the results and participated in the revision of the manuscript.
- KK identified and extracted data from included studies and participated in revision of the manuscript.
- JCC conceived, designed and developed the protocol, analysed and interpreted the results and edited the drafting and revision of the manuscript.
Declarations of interest
Sources of support
- Australia-Europe Endeavour Scholarship, 2005, Australia.
- University of Sydney Program Grant PhD Scholarship, Australia.
- No sources of support supplied
Medical Subject Headings (MeSH)
*Organ Transplantation; Acyclovir [therapeutic use]; Antiviral Agents [*therapeutic use]; Cytomegalovirus Infections [*prevention & control]; Ganciclovir [therapeutic use]; Immunocompromised Host; Opportunistic Infections [prevention & control]; Randomized Controlled Trials as Topic; Viremia [*prevention & control; virology]
MeSH check words
* Indicates the major publication for the study