Article Text


Human herpes virus 6 in archival cardiac tissues from children with idiopathic dilated cardiomyopathy or congenital heart disease
  1. M Comar1,4,
  2. P D’Agaro1,
  3. C Campello1,
  4. A Poli2,
  5. J P Breinholt III3,
  6. J A Towbin3,
  7. M Vatta3
  1. 1Department of Public Medicine Sciences, UCO Hygiene and Preventive Medicine, University of Trieste, and Institute of Child Health IRCCS Burlo Garofolo, Trieste, Italy
  2. 2Department of Medicine and Public Health, University of Verona, Italy
  3. 3Pediatric Cardiology, Baylor College of Medicine and Texas Children’s Hospital, Houston, Texas, USA
  4. 4Molecular Virology and Microbiology, Baylor College of Medicine, Houston, Texas, USA
  1. Professor C Campello, Department of Public Medicine Sciences, University of Trieste, Via dell’Istria 65/1, 34137 Trieste, Italy; campello{at}


Objective: To explore the possible role of human herpes virus 6 (HHV-6) in cardiac disorders in childhood in a retrospective study on archival specimens of explanted hearts.

Methods: 16 children (median age at transplantation 11.0 years) with idiopathic dilated cardiomyopathy (DCM) and 19 children (median age at transplantation 1.0 year) with congenital heart disease (CHD), previously found to be negative for other cardiotropic viruses such as enteroviruses, adenovirus, parvovirus B19, cytomegalovirus and Epstein–Barr virus, were tested for HHV-6 by quantitative real-time PCR and by genotyping. In addition, HHV-7/8 infection was investigated by qualitative PCR.

Results: HHV-6 B variant was detected in 11 of 35 samples (31.4%) with a mean viral load of 3.1×102 copies/μg of DNA. When assessed by heart disorder, the prevalence was different in the two groups (43.7% in DCM and 21% in CHD) while the mean viral loads were similar. In a logistic multivariate analysis HHV-6 was independently associated with DCM, taking CHD as reference and adjusting for age (best estimate: OR = 6.94; 95% CI 1.00 to 49.85; p = 0.05).

Conclusions: Although the clinical significance of the results is unknown, HHV-6 B genome is frequently detected in explanted hearts from children with DCM and to a lesser extent with CHD, thus adding evidence for HHV-6 cardiac involvement.

View Full Text

Statistics from

Infections by cardiotropic viruses have been associated with acute and chronic myocarditis leading to the development of dilated cardiomyopathy (DCM). As a result of viral presence and activity in the myocardium, inflammatory and/or autoimmune responses are frequently promoted, affecting the myocardium and coronary blood vessels.1 The finding of viral genome in cardiac tissues, even in the absence of the histological hallmarks of myocarditis, has been associated with endothelial dysfunction and dilated cardiomyopathy in adults.2 3 In transplant recipients, the detection of viral genome in the myocardium is predictive of poor outcomes and graft loss.4 Enteroviruses and adenovirus have been linked to cardiac disease over the past decade.57 More recently, other viral aetiologies have been identified including parvovirus B19 (PV-B19)8 9 and human herpes virus 6 (HHV-6).10 In addition, the presence of multiple viral genomes and the evidence of synergism were notable.11 In children, the prevalence of viral myocardial infections and their role in cardiac dysfunction has been evaluated to a lesser degree. Three cases of acute PV-B19 and/or HHV-6 myocarditis with fatal outcome have been reported.1214 In a large series of endomyocardial biopsies from transplanted patients, adenovirus and enteroviruses turned out to be the lead agents,4 while adenoviruses were frequently involved in children with myocarditis.15 Further, in a previous report we described a series of congenital heart defects as well as idiopathic cardiomyopathy in children in which PV-B19 genome was detected.9

HHV-6, the agent of an early childhood disease (exanthema subitum), may persist in a silent status in different organs; its reactivation can ensue into severe disorders in immunologically compromised subjects and in patients having solid organ or bone marrow transplantation.16 HHV-6 infection of tonsillar lymphoid tissue and of endothelial cells has been linked to inflammatory response. Specifically, in endothelial vascular cells, a site of latent viral persistence, the RANTES production, the expression of molecules for leucocyte adhesion and the release of chemokines are prominent mechanisms for endothelial inflammatory processes. Moreover, the HHV-6 tropism for the endothelium is further supported by a report of an active replication in human endothelium primary cell cultures.17 18

Considering that in the natural history of HHV-6 infection the pathogenetic pathway may lead to the involvement of endothelial and myocardial tissues, we sought to investigate HHV-6 genome in archival explanted myocardial tissue of children who underwent heart transplantation owing to cardiomyopathy or congenital heart disease.


Patients and samples

This retrospective survey was carried out on archived snap frozen endomyocardial tissue samples obtained from 35 children who underwent orthotopic heart transplantation at Texas Children’s Hospital. All samples had been previously found negative for enteroviruses, adenovirus, parvovirus B19, cytomegalovirus (CMV) and Epstein–Barr virus (EBV) using reverse transcription PCR (RT-PCR) and PCR, as reported.9 Samples were obtained following informed consent collection and institutional board approval.

Two distinct cohorts of patients were examined. Sixteen children with end-stage cardiomyopathy (all DCM) had a median age of 11.0 years (range 1–19 years, interquartile range 1.75–13.0 years). Histological examination was performed to assess the inflammatory and myocyte necrosis pattern as described for the Dallas criteria of myocarditis.19 Nineteen patients with end-stage complex congenital heart disease (CHD) had a median age of 1.0 year (range 20 days to 17 years, interquartile range 0.08–4.25 years).

Virological analysis

DNA extraction, primer sequences, PCR reaction conditions and the expected size of amplified fragments have been previously described.9 20 DNA from the HHV-6A (U1102) strain, as well as DNA from the HHV-6B (CV) strain, was included as positive control. In addition, amplification of a 110 bp sequence of the human β-globin gene was used to exclude the presence of inhibitors. Identification of HHV-6 variants was performed by restriction analysis using HpaI enzyme digestion of a U41 sequence amplified with variant-specific primer sets, including U41s-5′-TTT CGG AAC ATT GTT GAG C3′ and U41r-5′-GTG AAA ACT ACG ATT CAG GC 3′.21 Quantification of HHV-6 positive samples was performed using quantitative real-time PCR (qPCR) with TaqMan technology following the protocol published previously.22 In brief, PCR was performed in a 50 μl volume of mix containing 10 μl of DNA (10 ng/μl), for 45 cycles in 96-well plates using ABI PRISM 7000 Sequence Detection System (Applied Biosystems, Foster City, California, USA). For each assay, clinical samples, positive controls and negative controls were included in triplicate. The sensitivity threshold was estimated as 10 viral copies. The RNase P gene amplification kit (Applied Biosystems) was used as additional reaction control testing the quantity of human cells, ranging from 8.5×102 to 4.2×103 copies/μl of tested samples. Analysis for HHV-7 and HHV-8 genome was performed according to published protocols.20 23

Statistical analysis

Data are presented as median values with interquartile range (IQR, ie the range between first and third quartile) or as a geometric mean (GM) with 95% CI. Geometric mean was computed only on the basis of experimental measures. Differences in continuous measures were tested by the Mann–Whitney U test or the Student t test, as requested by normal or not normal distribution. Frequencies were compared using Fisher’s exact test. Multivariate analysis was performed by a logistic regression model, adjusting the odds ratios for age. To assess the goodness of fit in the model, we used the c index which estimates the probability of concordance between predicted and observed values: the index range is from 0.5 (no concordance) to 1.0 (perfect concordance).24 Hypotheses supported by a p value ⩽0.05 were considered significant. The statistical analysis was performed using the SPSS (SPSS Chicago, Illinois, USA) or the Egret packages (Search, Seattle, Washington, USA).


Tissues from 35 explanted hearts, found negative for enteroviruses, adenovirus, parvovirus B19, CMV, EBV and HHV-7/8, were analysed for HHV-6 genome; 11/35 (31.4%) tested positive. Genotyping of PCR positive samples was consistent with variant B. The viral genome was quantified and expressed as copies/μg DNA, ranging from 1.5×102 to 8×102, with a geometric mean of 3.0×102 copies/μg (95% CI, 2.0×102 to 4.4×102). In DCM children, the mean viral load was 2.9×102 copies/μg (95% CI, 1.9×102 to 4.5×102), while in CHD the mean was 3.1×102 copies/μg (95% CI, 1.4×102 to 7.1×102) (p = 0.70). HHV-6 prevalence was evaluated in the two groups, showing distinct differences regarding age at transplantation time (expressed by the median age of cohorts; p = 0.006) and the disorder (table 1).

Table 1 Age, HHV-6 presence and HHV-6 load in children with dilated cardiomyopathy (DCM) and congenital heart disease (CHD)

While the mean viral load was similar between CHD and DCM, the prevalence rate in subjects with DCM was double than that of the CHD group; however, the difference was not statistically significant probably due to low power of the sample. Since age appeared to be a determinant of HHV-6 infection, multivariate logistic analysis of both HHV-6 infection and type of cardiac disorder was performed, adjusting for age. Two statistical models were investigated in which HHV-6 and the types of cardiac disorder were alternatively evaluated as outcome. As shown in table 2, the probability that HHV-6 infection is associated with DCM was higher than in CHD (OR = 6.94; 95% CI, 1.0 to 49.85, p = 0.05). In this model, a good concordance was found between predicted and observed values (c = 76.2%; 95% CI, 59.0 to 93.3, p = 0.012). In addition, the probability that patients with DCM are infected by HHV-6 was about 10-fold higher than in CHD patients (OR = 10.94; 95% CI, 1.14 to 104.7, p = 0.038), but with weak concordance between predicted and observed values (c = 67.9%; 95% CI, 48.1 to 87.6, p = 0.087).

Table 2 Multivariate analysis of relationship between human herpes virus 6 (HHV-6) and heart disorders adjusting the OR for age


In this study HHV-6 genome was detected in 11 of the 35 (31.4%) archival myocardial tissue samples from children who underwent heart transplantation. Quantification of HHV-6 DNA in these cardiac tissues using qPCR showed that the viral load was low, typically less than 103 copies/μg of DNA. Genotyping identified always the variant B which is known to affect more commonly immunocompetent children.

The pattern of viral persistence in cardiac tissues of children is potentially different from that discovered in adults suffering with myocarditis and/or DCM; this difference could have several explanations. Firstly, the viruses may display a different cardiotropism due to the expression of specific receptors. Enterovirus and adenovirus recognise a specific receptor protein, the coxsackie and adenoviral receptor (CAR), to gain entry into the myocyte.25 On the other hand, parvovirus B19 binds to the P receptor present in embryonic myocytes.26 Since the main receptor for HHV-6 is the ubiquitous CD46 protein, a tropism for muscle and endothelial cells could play a role.27 Accordingly, endothelial cells have been identified as a natural reservoir for this virus with a lower frequency of infection in large versus small vessels. For instance, in primary cell cultures, the proportion of infected aortic endothelial cells is high (10%), while micro-vessels of the heart show small numbers of infected endothelial cells leading to the expression of a greater amount of inflammatory chemokines.17 A second explanation is epidemiological as the aetiological differences may reflect distinct temporal or geographical scenarios in viral heart infection.28 Finally, the different patterns between children and adults could simply reflect age-related epidemiological dynamics where infections by HHV-6 (and probably by PV-B19) affect young people more commonly, thus explaining both the HHV-6 prevalence and the absence of other viruses.

In this study, the most striking finding was the HHV-6 different prevalence between CHD and DCM, with DCM presenting about twice the rate of infection than CHD despite a similar mean viral load. HHV-6 infection rate increases with age and most children have antibody to B variant by 2 years. Therefore, the described HHV-6 infection prevalence in DCM could be merely age-dependent, owing to the high mean age of this cohort. However, using a logistic regression model to adjust for age, an independent association emerged between HHV-6 infection and DCM in comparison to CHD. The clinical relevance of this is unclear and is beyond the scope of this retrospective analysis, but the finding may be consistent with the recently defined role of HHV-6 in chronic cardiac disorders.11

Does HHV-6 play a role in the development of cardiomyopathy or congenital heart defects? No relationship between HHV-6 infection in utero and congenital heart disorders has been reported to date, but the possibility is theoretically plausible due to the frequency of its precocious vertical transmission. In a recent large survey the prevalence of cord blood infection by HHV-6 was established to be 1%. However, no symptomatic subjects were identified, still in the presence of active HHV-6 replication.28 Thus, despite the fact that in utero HHV-6 infection is frequent and similar to that due to CMV, the HHV-6-related embryopathy appears to be a rare event. Moreover, no prospective survey has delineated the long-term consequence of HHV-6 congenital infections.29 Alternatively, the presence of HHV-6 in CHD cases could be linked to an intrapartum infection or early horizontal transmission.22 30

Collectively, the findings reported here raise the question of the significance of HHV-6 in paediatric heart disease: an innocent bystander expressing a latent infection or a pathogenic agent? Histopathological analysis fails to provide definitive information since only one sample showed active myocarditis. A firmer conclusion could be drawn when the virological evaluation couples the viral load with the expression of mRNA, either early or late.21 In this study, only the first marker could be studied, owing to the small amount of residual archival samples. Hence, a number of other limitations should be recognised in this study, such as the lack of multiple biopsies examined and the absence of histochemical and/or in situ PCR evaluation aimed to recognise and quantify the endothelial/myocardial cell involvement.17

In conclusion, this is the first evidence of HHV-6 infection in the explanted hearts of children combined with a low viral load suggestive of latent infection. Whether this virus is an important cause of CHD and DCM should be carefully evaluated in a prospective study.

Take-home messages

  • Human herpes virus 6 (HHV-6) variant B was frequently found in explanted heart tissues of children, while the viral loads were constantly low.

  • HHV-6 was more frequently detected in tissues from dilated cardiomyopathy patients than congenital heart disease patients.

  • Though the clinical significance of these findings remains obscure, the HHV-6 involvement in paediatric heart disorders is established, but requires further evaluation.


View Abstract


  • Funding: MV is funded by the NIH (grant numbers HL077706 and HL078807). MC is funded by grant R.C. 70/2005 from the IRCCS Burlo Garofolo, Trieste, Italy. JAT is funded by the NIH (grant numbers HL67155 and HL65652), the Abby Glaser Fund, the Vivian L Smith Foundation, and the Children’s Cardiomyopathy Foundation.

  • Competing interests: None.

  • Ethics approval: Ethics approval was obtained.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.