An Assessment of Herpesvirus Co-infections in Patients with CMV Disease: Correlation with Clinical and Virologic Outcomes


  • Other members of the VICTOR study group are listed in the Appendix.

* Corresponding author: Atul Humar,


The effect of herpesvirus co-infections (HHV-6, HHV-7) on cytomegalovirus (CMV) disease and its response to therapy is unknown. We prospectively analyzed herpesvirus co-infections in transplant recipients with CMV disease. All patients received 3 weeks of antiviral therapy. Samples were collected at baseline (day 0) and then day 3, 7, 14 and 21 poststart of therapy. Viral load testing for CMV, HHV-6 and HHV-7 was done using quantitative PCR assays in 302 patients of whom 256 had documented symptomatic CMV viremia. In this subset, day 0 HHV-6 co-infection was present in 23/253 (9.1%) and HHV-7 in 17/253 (6.7%). Including those positive at any time point raised the prevalence to 79/256 (30.9%) for HHV-6 and 75/256 (29.3%) for HHV-7. Viral co-infection did not influence the response of CMV disease to antiviral therapy. Baseline CMV viral loads, time to eradication and risk of recurrence were similar in patients with and without HHV-6 or HHV-7 co-infection. Ganciclovir and valganciclovir had no clear effect on HHV-6 and HHV-7 viremia. In conclusion, herpesvirus co-infections are common in patients with CMV disease but with standard antiviral therapy, no clear clinical effects are discernable. Routine monitoring for viral co-infection in patients with CMV disease is not indicated.


Cytomegalovirus (CMV) disease continues to be common in certain transplant populations. Recent data have demonstrated differing viral kinetics in response to antiviral therapy and high rates of recurrence with standard treatment courses (1,2). Two other common herpesviruses, Human Herpesvirus-6 (HHV-6) and Human Herpesvirus-7 (HHV-7) may co-infect patients with CMV disease and could potentially influence the response to antiviral therapy. These viruses are all classified in the β-herpesvirus family. Infection with both HHV-6 and HHV-7 commonly occurs in childhood and then subsequently results in life long latency such that the seroprevalence rate in adults is over 90% (3,4). Therefore in adult solid organ transplant recipients, while primary infection is uncommon, reactivation of endogenous latent viruses seems to occur very frequently with infection rates of 30–50% reported in many studies (5–12). Most patients appear to have asymptomatic viral reactivation, and clinical syndromes have commonly only been described with HHV-6. These include a CMV negative viral syndrome, hepatitis, encephalitis and pneumonitis (13–16). However, both HHV-6 and HHV-7 have been associated with indirect immunomodulatory effects, which may be more important in this population than the direct effects.

The best-described indirect effect of these viruses appears to be an interaction with CMV that may promote replication and persistence of the latter virus (17). The underlying pathophysiological basis of such interactions is largely unknown, but may be due to cytokine dysregulation induced by viral gene products or due to more direct effects of viral co-infection within a single cell (17). In one study, 17/19 patients (89%) with CMV viral syndrome were found to have a co-infection with HHV-6 and/or HHV-7 (18). It has been suggested that many viral syndromes observed in clinical practice are due to a combination of herpesvirus infections (17–19). However, the contribution of HHV-6 and HHV-7 to the overall disease state in patients with CMV disease is largely unknown. Similarly, the influence of these viral co-infections on the response of the CMV disease to antiviral therapy is also unknown since routine diagnostic assessment for these viruses in patients with CMV disease is not currently recommended (17).

The objectives of this study were to prospectively assess viral co-infections with HHV-6 and HHV-7 in a large cohort of patients with CMV disease. The incidence of such co-infections and its effects on CMV treatment response and recurrence rates was determined.


Patients, study design and definitions

Patients were recruited as part of a large international randomized CMV treatment trial (the VICTOR trial) comparing intravenous ganciclovir to oral valganciclovir for the treatment of CMV disease in solid organ transplant recipients (1). This HHV-6 and HHV-7 sub-study was designed a priori and all patients signed an informed consent for HHV-6 and HHV-7 viral load testing as part of this sub-study. For the purposes of the sub-study, all patients regardless of treatment randomization (intravenous vs. oral) were combined and analyzed together. The VICTOR trial is registered at (NCT00431353) and was approved by local regulatory agencies and ethical committees.

The inclusion and exclusion criteria have been described in detail before (1). Briefly, solid organ transplanted patients with symptomatic CMV disease were eligible for inclusion in the study. This included liver, kidney, heart, pancreas, lung and combined organ transplant recipients. Immunosuppression protocols differed according to the specific transplant programs at each center. CMV viral syndrome disease and CMV tissue invasive disease were defined according to standard criteria (20). All patients with CMV disease were treated with induction doses of either intravenous ganciclovir (10 mg/kg/day in divided doses) or oral valganciclovir (900 mg twice daily) for 21 days. After the induction treatment, all patients received an additional 28 days of secondary prophylaxis with valganciclovir 900 mg orally once daily. All doses were adjusted for renal function according to the labels.

Laboratory testing

CMV viral load testing:  CMV viral load testing was done centrally on plasma samples using the Roche Cobas Amplicor assay as per manufacturer's instructions. The detection limit of this assay is approximately 200 copies/mL and the quantification level is 600 copies/mL. CMV viral load testing was performed at baseline (day 0) and at days 3, 7, 10, 14, 17, 21, 28, 35, 42, 49 and at months 3 and 6 and for assessment of recurrent disease if clinically indicated.

HHV-6 and HHV-7 testing:  HHV-6-, HHV-7 viral load testing was carried out centrally by a technologist who was blinded to the outcomes of the clinical trial. Viral DNA was extracted from 200 μL of whole blood using the IsoQuick Method (ORCA Research, Bothell, WA) and was eluted in 50 μL of DNAse-free and RNAse-free water. Five microliters of the eluted DNA was added to 15 μL of Mastermix Solution (Roche Molecular Biochemicals, Indianapolis, IN) containing the primers and probes specific for each of the viruses under investigation, and viral load was quantified using LightCycler™ PCR (Roche Molecular Biochemicals). Each virus was detected and quantified independently of the others. For HHV-6 the RealArt HHV-6A/B LC PCR Kit and protocol from Qiagen Inc (Valencia, CA) was used according to manufacturer's instructions. Melting curve analysis was used to differentiate HHV-6A from HHV-6B as per manufacturer's instructions. For HHV-7, an in-house assay was utilized using primers: P1 5′ to 3′: ATG TAC CAA TAC GGT CCC ACT TG and P2 5′ to 3′: AGA GCT TGC GTT GTG CAT GTT (Invitrogen Custom Primers) and the DNA SYBR-Green kit from Qiagen. All results were in copies/mL. The copy number was determined using a standard curve with known copy numbers. The lower limit of detection is approximately 10 copies/mL, but quantification is most accurate at above approximately 100 copies/mL. Testing was performed on day 0 (baseline) samples and then on days 3, 7, 14 and 21 after the start of antiviral therapy.

Statistical analysis:  Descriptive statistics were used to calculate the incidence of viral co-infection. Dichotomous variables were analyzed using the chi-square test. Continuous variables were analyzed by Mann–Whitney U test. Times to eradication/recurrence were analyzed with the Kaplan–Meier survival analysis. Statistical analysis was done using SPSS version 15.0 (SPSS Inc., Chicago, IL).


Patient population

In the primary study a total of 321 patients with CMV disease were enrolled. Baseline demographic characteristics of the total patient population have been described in detail previously (1). Transplant types included kidney (n = 237), liver (n = 23), heart (n = 18), lung (n = 19) and other or combined (n = 24). Mean age was 45.3 years (range 18–72 years) and average time from transplantation was almost one year (range 10–9257 days). For the purposes of this sub-study, patients were analyzed as a single group regardless of antiviral treatment allocation (valganciclovir or intravenous ganciclovir). Of the total population, 302 patients had at least one sample (94.1%) available for analysis of HHV-6 and HHV-7 viral loads. In the remaining 19 patients, insufficient sample was available for HHV-6 and HHV-7 testing because of other testing as required by the parent study (CMV viral load, resistance testing, ganciclovir levels). Samples were available for 256 patients with CMV viremia confirmed at the central laboratory (out of the 259 ‘per-protocol-population’ of the parent study [98.8%]) and 46 patients who were CMV viremia negative at the central laboratory, but were included in the study based on a positive local assessment of CMV viremia and clinical signs of CMV disease (intention to treat population) (45/46 day 0 samples available).

HHV-6 and HHV-7 viremia

The results of day 0 (onset of antiviral therapy) PCR testing in 298 transplant patients are shown in Table 1 (4/302 (1.3%) patients had no day 0 sample available). No virus (CMV, HHV-6 or HHV-7) was detected in 33 (11.1%) patients despite a diagnosis of CMV disease by the local site investigator. An additional 12 (4.0%) patients had no detectable CMV, but were found to have either HHV-6 viremia (n = 4), HHV-7 viremia (n = 7) or a combination of the two (n = 1). In patients with detectable CMV at baseline (i.e. patients with centrally confirmed positive CMV viremia (n = 253 day 0 samples available out of 256 per-protocol patients with a sample available at any time point), a viral co-infection at baseline with either HHV-6 or HHV-7 was detectable in 38 (15.0%) patients (HHV-6 in 8.3%, HHV-7 in 5.9% and both in 0.8% of patients with CMV disease). In patients with a positive HHV-6 or HHV-7 viremia at baseline (day 0), the median viral load for HHV-6 was 372 copies/mL (range 21–186 209 copies/mL) and 35 copies/mL (range 11–2818 copies/mL) for HHV-7. Many of the viral loads were very low, especially for HHV-7. For example, for day 0 HHV-6 positive results, 9/28 (32.1%) were below 100 copies/mL. Similarly for day 0 HHV-7 positive results, 20/26 (76.9%) were below 100 copies/mL (Figure 1).

Table 1.  Results of PCR testing for CMV, Human Herpesvirus-6 (HHV-6) and Human Herpesvirus-7 (HHV-7) at baseline (day 0) blood samples from 298 transplant patients (intention to treat population)
VirusN (%)
Monitored298 (100%) 
No viruses detected 33 (11.1%)
CMV alone215 (67.0%)
CMV + HHV-621 (7.0%)
CMV + HHV-715 (5.0%)
CMV + HHV-6 + HHV-7 2 (0.7%)
HHV-7 alone 7 (2.3%)
HHV-6 alone 4 (1.3%)
HHV-6 + HHV-7 alone 1 (0.3%)
Figure 1.

Baseline viral loads of HHV-6 and HHV-7 for intention to treat patients. Solid line indicates median viral load (copies/mL).

All patients with HHV-6 viremia at baseline were found to have HHV-6 subtype B based on melting curve analysis. The proportion of patients with no detectable CMV at baseline but with either HHV-6 and/or HHV-7 was 12/45 (26.7%) compared to a rate of viral co-infection of 38/253 (15.0%) in the CMV positive per-protocol population (p = 0.08). The incidence and level of viremia was compared in patients who received intravenous ganciclovir versus oral valganciclovir therapy. The incidence of HHV-6 viremia at any time point was 31.1% for the ganciclovir group versus 31.2% for the valganciclovir group (p = NS). The median peak HHV-6 viral load was also similar in the two arms (281 vs. 224 copies/mL respectively; p = NS). Similarly, the incidence of HHV-7 viremia and peak viral loads were not different in patients receiving ganciclovir versus valganciclovir (34.5% vs. 31.2% and 27 vs. 50 copies/mL respectively; p = NS for all comparisons).

Response to antiviral therapy

The influence of antiviral therapy upon the HHV-6 and HHV-7 viral loads was assessed. In the patients with a baseline positive HHV-6/-7 viremia (day 0), the viral loads decreased or became negative in the majority of patients on subsequent measurements obtained after commencement of antiviral therapy. In HHV-6 viremic patients, by day 3 of antiviral therapy, 44.4% were negative and by day 7, 60.7% were negative. The median log10-fold decline in viral load by day 3 of antiviral therapy was 1.05 logs (range: increase of 0.30 to decrease of 3.98 logs) (where a 1-log decline corresponds to a 10-fold decline in viral load). By day 7 of antiviral therapy, the median log10-fold decline in viral load of HHV6 was 1.57 logs (range: increase of 0.16 to decrease of 3.98 logs). For HHV-7 viremic patients, by day 3 of antiviral therapy, 60.0% were negative and by day 7, 66.7% were negative.

In the per-protocol population many of the patients who were initially negative for both HHV-6 and HHV-7, subsequently became positive at some point between day 3 to day 21 despite full-dose antiviral therapy. Therefore, the overall viremia rate at any time point measured were 79/256 patients (30.9%) for HHV-6 and 75/256 patients (29.3%) for HHV-7. The majority of viral loads were low especially for HHV-7 (<100 copies/mL), but occasionally relatively high viral loads were detected for the first time while on antiviral therapy (Figure 2). Using a cut-off of 100 copies/mL for HHV-6 22/79 (27.8%) of positive patients had a viral load <100 copies/mL and for HHV-7 59/75 (78.7%) of positive patients had a viral load <100 copies/mL.

Figure 2.

(A) HHV-6: Prevalence of positive viral loads at different time points post antiviral therapy in the intention to treat population (n = 302) (only positive results shown). (B) HHV-7: Prevalence of Positive Viral Loads at different time points post antiviral therapy in the intention to treat population (n = 302) (only positive results shown).

Impact on CMV

Viral co-infection did not influence the response of CMV disease to antiviral therapy. The effect of HHV-6 and HHV-7 co-infections on CMV disease responses are shown in Tables 2 and 3. Baseline CMV viral loads were similar in patients with and without HHV-6 or HHV-7 co-infection at baseline. The median viral load in both groups by day 21 was undetectable, and time to CMV eradication was not any longer in co-infected patients (Tables 2 and 3). For example, if HHV-6 co-infection was present at baseline the median time to CMV eradication was 18 days versus 20 days if no co-infection was present (p = 0.51).

Table 2.  The effect HHV-6 and CMV co-infection on outcomes of CMV disease in the per-protocol population (n = 253)
 HHV-6 present at day 0 (n = 23)HHV-6 absent at day 0 (n = 230)p-Value
Baseline CMV VL (copies/mL plasma)15 400 (645–392 500)19 325 (645–750 000)0.73
CMV VL at Day 21 (copies/mL plasma)Undetectable (undetectable–6100)Undetectable (undetectable–80 000)0.55
Median days to CMV clearance (95% confidence interval)18 (12–24)20 (19–21)0.51
Clinical recurrence of CMV3/20 (15.0%)31/209 (14.8%)0.98
Table 3.  The effect of HHV-7 and CMV co-infection on outcomes of CMV disease in the per-protocol population (n = 253)
 HHV-7 present at day 0 (n = 17)HHV-7 present at day 0 (n = 236)p-Value
Baseline CMV VL (copies/mL plasma) (range)7150 (645–165 000)19 325 (645–750 000)0.43
CMV VL at Day 21 (copies/mL plasma) (range)Undetectable (undetectable–5450)Undetectable (undetectable–80 000)0.58
Median days to CMV clearance (95% confidence interval)18 days (10–26)20 days (19–21)0.95
Clinical recurrence of CMV2/17 (11.8%)32/212 (15.1%)0.71

Assessing co-infection at time points other than day 0 (day 3, 7, 14 and 21) there was still no apparent effect on viral co-infection and CMV disease response to antiviral therapy. For example, median time to CMV eradication was 18 days in patients with either HHV-6 or 7 co-infection at any time point (n = 130) versus 21 days in patients with no co-infection at any time point (n = 126; p = 0.08).

Patients were followed out to 12 months from enrolment to assess CMV recurrence with regular viral load sampling (at 3 and 6 months) and clinical follow-up with viral load testing at the time of suspected recurrence. A total of 233/259 patients (90%) had the required follow-up viral load measurements. Overall, clinical recurrence rates for CMV disease by month 12 were 14.6% and the virologic recurrence rate was 29.2%. The presence of HHV-6 or HHV-7 co-infection either at baseline or at any time point from day 0 to day 21 did not influence the rate of CMV recurrence. For example, in patients with detectable HHV-6 infection at anytime between day 0 and day 21, the rate of virologic and clinical recurrence of CMV was 32.9% and 13.7% respectively. This compared to virologic and clinical CMV recurrence rates of 28.3% and 15.7% in patients without HHV-6 co-infection (p = 0.45 and 0.67, respectively).

Latent versus lytic infection

In an effort to determine if some of the HHV-6/HHV-7 positive samples were in fact picking up latent virus, we examined the likelihood of persistent positives over multiple measurements. For HHV-6, only 7/302 patients were found to be positive on all 5 time points tested. In only four of these seven patients, a less than 1-log-fold variation was found over time suggesting very similar viral loads over multiple measurements and possibly detection of latent virus. In only one patient was viral load consistently greater than 4-log10 copies/mL suggesting the possibility of genomic integration of HHV-6. For HHV-7, only 5/302 patients were found to be positive on all 5 time points testing. In only two of these five patients, a less than 1-log-fold variation was found over time suggesting very similar viral loads over multiple measurements and possibly detection of latent virus.


We analyzed the effects of HHV-6 and HHV-7 co-infection on CMV disease in a very large cohort of patients with CMV disease. All patients received the same duration of antiviral therapy, and all outcomes were followed prospectively. We found that detection of viral co-infections was not uncommon both at baseline (start of antiviral therapy) and during subsequent measurements over the first 21 days of induction antiviral therapy. A significant range of viral loads were observed with viral loads tending to be higher for HHV-6 compared with HHV-7. However, no clear clinical significance could be assigned to the presence of HHV-6 and HHV-7 and the findings may represent a marker of intense immunosuppression in these patients. Specifically, viral co-infection had no discernable effect on the response of CMV disease to therapy. Therefore, the rate of CMV clearance in response to antiviral therapy was similar in patients who had co-infection with either HHV-6 or HHV-7 compared with those who were only infected with CMV. Also, no effect of HHV-6 or HHV-7 infection on the development of recurrent CMV disease was observed. Based on these data, routine assessment of HHV-6 and HHV-7 viremia in patients with CMV disease is not indicated. In addition, we conclude that the detection of either virus in patients with CMV disease is not a reason to alter the clinical management. A small number of patients in the intention to treat population had no CMV detected at the central lab but did have HHV-6 and/or HHV-7 detected (n = 12/298; 4.0%). The latter viruses may have accounted for the ‘viral syndrome’ presentation in these patients. This is supported by the higher proportion of CMV negative patients found to have HHV-6 or HHV-7 infection compared with CMV positive patients (26.7% vs. 15.0%; p = 0.08).

Previous studies of HHV-6 and HHV-7 have suggested that reactivation of these viruses is very common following transplantation. Rates of viremia ranging from 14–82% have been reported with most studies suggesting an incidence of approximately 30% for the two viruses, which concurs with our present data (5–12). Wide variations in rates may be due to different diagnostic tests and different patient populations. Several studies have suggested that HHV-6 and HHV-7 have significant immunomodulatory indirect effects that likely outweigh direct clinical effects (7–11,14,17). HHV-6 infection of T cells results in down-regulation of IL-2 mRNA and protein synthesis accompanied by a significant reduction in mitogen-driven proliferative responses resulting in a cell mediated immune defect (21). This immunosuppressive effect may result in further opportunistic infections, and specifically may promote CMV replication. HHV-6 has also been shown to modulate cytokine expression in infected mononuclear cells and acts as a potent inducer of TNF-α production (22). In turn, TNF-α has been shown to stimulate the CMV immediate early gene enhancer/promoter region in a dose-dependent manner resulting in CMV reactivation (23).

The proposed pathophysiological basis of this viral interaction is supported by a number of epidemiological studies. Many of these studies have found that HHV-6 and/or HHV-7 infection are important risk factors for the development of subsequent opportunistic infections, including CMV disease. For example, Desjardins et al. (6) evaluated HHV-6 seroconversion in 53 kidney transplant recipients and found it was a risk factor for CMV infection and disease. Studies using molecular diagnostic assays have also found significant, although occasionally conflicting, associations with viral infection and indirect effects. For example, in a study of 200 liver transplant recipients, HHV-6 viremia was associated with a 3.6-fold increased risk of subsequent CMV disease (7). In contrast, in a study by Razonable et al. (12), in 263 D+/R− organ transplant patient given 3 months of antiviral prophylaxis, no association between HHV-6/HHV-7 and CMV disease was found. Although these studies may be confounded by use of different diagnostic tests, sample size restrictions and difficulty in determining temporal relationships between viral infection and outcomes, most of the literature does support indirect effects of viral replication (24). By the nature of its design, our study did not assess if HHV-6 or HHV-7 predispose to the development of CMV disease. This is because, by definition, all patients included in the parent study already had investigator diagnosed CMV disease. Instead, we evaluated the effect of viral co-infections on CMV disease progression and response to anti-CMV therapy, which has not been previously assessed in a large prospective study assessing several time points. The present results do not support any effect of co-infection on clinical outcome once CMV disease is present and treated with standard antiviral drugs.

This specific question has been assessed in a number of studies with small sample sizes. In a study by Razonable et al. (18), HHV-6 and/or HHV-7 were detected in 18 of 20 (90%) episodes of presumed CMV infection. Clinical effects were difficult to discern due to the small sample size but certain trends in laboratory abnormalities suggested more significant disease in co-infected patients. Similarly, Lautenschlager et al. (25) found a high rate of co-infections in 12 liver transplant patients with CMV disease (10/12 had HHV-6 and 9/12 had HHV-7). Antiviral treatment resulted in resolution of HHV-6 and HHV-7 viremia but at a slower rate than CMV viremia. In another study, Kidd et al. (26) evaluated 52 kidney transplant patients with serial PCR testing. CMV and HHV-7 co-infections were found in 15/52 (29%) patients and co-infected patients were more likely to develop symptomatic CMV disease.

Our study had some limitations. First, the study was a sub-study of a larger CMV treatment study and therefore the assessment of viral co-infection was not the primary goal of the study. However, the decision to assess HHV-6 and HHV-7 co-infection as a secondary endpoint was made a priori (i.e. during the planning stages of the parent study). Therefore, frequent sampling was done with appropriate specimens saved for further viral testing. However, another potential limitation is that some subjects, albeit very few, did not have a sample for testing of HHV-6/7 at each of the time points. Another point is the ability of molecular diagnostic assays to distinguish latent versus actively replicating viral infection. Since the majority of patients had viremia detected only on one or two occasions (only 7/302 [2.3%] patients had HHV-6 viremia detected on all time points tested), and since viral loads especially for HHV-6, were relatively high for some patients, we do not believe this to be a major confounder of the results. Integrated genomic HHV-6 virus has also been described as a cause of persistently very high viral loads in otherwise asymptomatic people (27). However, viral loads in patients with genomic integration of HHV-6 are often greater than 107 copies/mL and persistently detectable at a high-copy number, something that was not observed in the current study. Another limitation of our study was that immunosuppression protocols were not mandated. Therefore, investigators were free to change immunosuppression as per their own clinical practice. This may have confounded the interpretation of viral load changes with the commencement of antiviral therapy. Strengths of our study include the prospective multicenter nature of the data collected, sequential samples and the large sample size. This is the largest study to report on viral co-infections in transplant patients with CMV disease, and as such many of the findings can be considered definitive.

In vitro, current antiviral agents, including ganciclovir, cidofovir and foscarnet appear to have some activity against HHV-6 and HHV-7 (28). However, ganciclovir appears less active against the HHV-6B variant compared to the HHV-6A variant, an important finding since all day 0 HHV-6 positive subjects in our study were infected with the HHV-6B variant (28). The in vivo efficacy of these agents is unclear. In a prospective study of HHV-6 and HHV-7 viremia in a large cohort of CMV D+/R- patients, antiviral prophylaxis seemed to be associated with a low rate of HHV-6 viremia but had no effect on HHV-7 viremia (12). In a study evaluating HHV-7 viremia in 92 kidney transplant patients, 89% were PCR positive on at least one occasion and oral ganciclovir prophylaxis or treatment with intravenous ganciclovir had no effect on the presence of HHV-7 (29). This is further supported by our data, in which antiviral therapy had no discernable effect on either HHV-6 or HHV-7 viremia. Also, no significant differences were observed between ganciclovir and valganciclovir. Many patients did have intermittent viremia after commencement of therapy. The initial decline in viral load observed in baseline positive patients may represent the natural history of intermittent viremia or be due to changes in immunosuppression. As per protocol, no other antiviral agents including immunoglobulin were used to treat the CMV disease. Overall, the detection of multiple circulating viruses in these patients may be a sentinel marker of over-immunosuppression. It should be noted our study was not specifically designed to assess the efficacy of antiviral agents against HHV-6/7 and multiple confounding factors may be present such as changes in immunosuppression, type of organ transplant, unknown interactions between herpesviruses and lack of a control group that did not receive antiviral therapy.

In conclusion, our study demonstrates that in patients with CMV disease, viral co-infection is not uncommon. Viral loads were generally low, especially for HHV-7 and no clear clinical relevance could be assigned to HHV-6 or HHV-7 viremia except possibly in the subset of patients without CMV viremia. In patients with CMV disease, viral co-infection had no discernable effect on the response of CMV to standard antiviral therapy or on the risk of subsequent CMV disease recurrence. Based on these data, routine monitoring for HHV-6 and/or HHV-7 co-infection in patients with CMV disease is not indicated and standard anti-CMV therapy does not normally need to be modified when viral co-infection is detected.


We are indebted to Angelo Bignamini for statistical support. The work was funded by F. Hoffmann-La Roche.

Conflict of Interest: Atul Humar, Anders Åsberg, Anders Hartmann, Alan Jardine and Mark D. Pescovitz have done consultancy work for F. Hoffmann-La Roche. Houria Mouas is a former employee of F. Hoffman-La Roche and Christoph G. Gahlemann is a current employee.


Appendix: Members of the VICTOR Study Group

North America: Sandra Cockfield. Latin America: Anabela Armino, Luis F.A. Carnargo, Carmen Garcia, Carlos Henriquez, Marilda Mazzali, Elias David Neto, Irene L. Noronha, Jose Osmar Medina Pestana, Rfael Reyes. Asia, Pacific Islands and Australia: Allan Glanville, George T. John, Vijay Kher, R.K. Sharma, C.M. Thiagarjan. Europe and Middle East: Jose Maria Aguado, Emel Akoglu, Hofman Blazenka, Felix Burkhalter, Magdalena Durlik, Antonio Franco Esteve, Marciej Glyda, Abdul Hammad, Rajko Hrvacevic, Madis Ilmoja, Marian Klinger, Dirk Kuypers, Phil Mason, Mai Ots, Patrik Peters, Rafails Rozentals, Boleslaw Rutkowski, Sabine Schmaldienst, Juerg Steiger, M. Tuncer, Zbigniew Wlodarczyk, Michael Zakliczynsk.