Article Text

Download PDFPDF

Clinicopathological features of paediatric deaths due to myocarditis: an autopsy series
  1. M A Weber1,
  2. M T Ashworth1,
  3. R A Risdon1,
  4. M Malone1,
  5. M Burch2,
  6. N J Sebire1
  1. 1
    Department of Paediatric Pathology, Great Ormond Street Hospital for Children, and UCL Institute of Child Health, London, UK
  2. 2
    Paediatric Cardiology, Great Ormond Street Hospital for Children, and UCL Institute of Child Health, London, UK
  1. Dr N J Sebire, Department of Paediatric Pathology, Great Ormond Street Hospital for Children, Great Ormond Street, London, WC1N 3JH; SebirN{at}gosh.nhs.uk

Abstract

Introduction: Myocarditis is a recognised cause of cardiac failure in childhood but the frequency of myocarditis as a cause of sudden unexpected death across the paediatric age range is uncertain.

Methods: A structured review of the results of all autopsies carried out in a single paediatric centre over a 10-year period, including the results of all investigations performed as part of the centre’s policy for the post-mortem investigation of paediatric deaths.

Results: During the study period there were 1516 autopsies of children aged 0–18 years. Histologically proven myocarditis was present in 28 cases (1.8%, age range 10 days to 16 years, median age 10 months), of which 16 (57%) presented as sudden death. More than half of all cases (54%) occurred in infants less than 1 year of age, accounting for 2% of infant deaths referred for autopsy, compared with around 5% of childhood deaths over the age of 5 years. In almost 40% of cases there were no macroscopic cardiac abnormalities, the diagnosis being entirely dependent on routine histological examination of the heart, and post-mortem heart weight was normal in the majority of cases. Virus was detected in nine (36%) of the 25 cases in whom virological analyses were performed. The histological features were similar in all cases, with an interstitial inflammatory cell infiltrate, predominantly lymphocytic, with focal myocyte necrosis and interstitial oedema.

Conclusions: Myocarditis is a rare cause of death in infancy and childhood, and the majority of cases present as sudden unexpected deaths, which require routine histological sampling of the heart for its detection.

View Full Text

Statistics from Altmetric.com

Myocarditis is a recognised cause of cardiac failure in childhood and may present with non-specific clinical features of progressive cardiac dysfunction or with those of dilated cardiomyopathy.13 It is also a recognised cause of sudden unexpected death in both children and adults.46 The reported incidence of fatal myocarditis based on death certification ranges from 1.59 to 0.24 to 0.12 per 100 000 person-years for children aged <1 year, 1–4 years and 5–14 years, respectively,7 but the incidence of histologically proven cases of fatal myocarditis seems to be much lower, and the frequency of myocarditis as a cause of sudden death across the paediatric age range is uncertain. The aim of this study is to determine the frequency of myocarditis in a large series of paediatric autopsies from a single centre and, in particular, to review the epidemiological, clinical and pathological features of all infants and children in whom the cause of death was given as myocarditis.

METHODS

As part of a large retrospective study examining the clinicopathological findings in an unselected series of paediatric autopsies from a single specialist centre during a 10-year period, 1996 to 2005 inclusive, all available clinical and pathological data were entered into a structured database using strictly defined criteria. The data included results of all investigations performed as part of the centre’s policy for the post-mortem investigation of paediatric deaths. Cases were identified by study number only, and the study was approved by the local research ethics committee.

A computerised search of the database was carried out to identify all cases of myocarditis, and the details of these were reviewed. For the purposes of this study, the diagnosis of myocarditis was based on the pathologists’ findings at the time of the post-mortem examination, and was limited to those cases in which there was histological evidence of inflammation of the myocardium with associated interstitial oedema, myocyte separation and/or focal myocyte damage. Cases in which there was a clinical suspicion of myocarditis in the absence of confirmatory histological features were not classified as myocarditis in this series, although it is recognised that by limiting this series to cases which meet the Dallas criteria one may underestimate the true incidence of myocarditis. However, in the absence of other well-defined, universally accepted criteria for the post-mortem diagnosis of myocarditis, it was felt that this remained justified, since this most probably reflects current pathologist practice in the United Kingdom. In addition to reviewing the clinical and pathological findings of all cases of myocarditis, heart weights of affected infants (<1 year of age) were compared with heart weights of the entire unselected population of infant deaths referred for autopsy at the centre over the 10-year study period using a linear regression model and calculation of delta values (the number of standard deviations by which the actual weight differed from the expected weight for age) to account for age-related weight changes. Delta values were compared between the groups using the Mann–Whitney U test. The significance of differences in proportions was determined using the χ2 test.

RESULTS

During the study period, a total of 1516 autopsies were carried out in children from birth to 18 years. Histologically proven myocarditis was identified in 28 cases (1.8%), of which five (18%) presented as sudden unexpected death with no apparent prodromal symptoms and 12 (43%) as sudden death with non-specific preceding symptoms that were not recognised to be life-threatening (table 1). The latter group included six children with varying degrees of dyspnoea and/or tachypnoea and three children with non-specific diarrhoea and vomiting; one child presented with a 4-day history of pyrexia and another with “non-specific viral symptoms” and abdominal pain. It also included one 15-year-old girl who presented with headache and vomiting owing to raised intracranial pressure caused by an undiagnosed cerebellar medulloblastoma in whom myocarditis was an incidental finding at autopsy. Thus, excluding the latter case, there were 16 cases overall (57%) that presented with sudden death owing to myocarditis. The remaining 11 children (39%) presented as non-sudden deaths: seven of these were hospitalised before death with worsening respiratory symptoms (at least three of whom were initially treated for a presumed chest infection) and two for suspected “sepsis”; the other two presented with clinical features of dilated cardiomyopathy.

Table 1 Clinicopathological features of 28 cases of histologically proven myocarditis

The age range of all cases in whom myocarditis was identified was 10 days to 16 years, with a median age of 294 days (10 months). Of the 28 cases, 15 (54%) were in infants under 1 year, five (18%) in children aged 1–4 years, four (14%) in children aged 5–9 years, two (7%) in children aged 10–14 years, and two (7%) in children ⩾15 years of age. Thus, as a proportion of all deaths referred for autopsy to the specialist centre over the 10-year study period, myocarditis was identified in 1.6% of all 965 infant deaths, in 1.9% of the 261 childhood deaths at 1–4 years of age, and in 5.1% of deaths in children aged 5–9 years, 4.8% of children aged 10–14 years and 7.7% of children ⩾15 years of age (χ2 10.59, df  = 4, p = 0.03, fig 1). Overall, there were 17 girls and 11 boys.

Figure 1 Prevalence of myocarditis across the paediatric age range. Fatal myocarditis is relatively more common in older children, accounting for 5% of childhood deaths over the age of 5 years (95% CI 2.4% to 10.5%) compared with <2% of all infant deaths referred for autopsy (95% CI 0.9% to 2.6%).

Macroscopic abnormalities of the heart were identified at post-mortem examination in only 61% of children with histologically proven myocarditis, including a dilated and/or subjectively enlarged heart in seven (25%), myocardial pallor/mottling in five (18%) and a pericardial effusion in six (21%). In contrast, in 11 children (39%) the heart appeared entirely normal upon post-mortem macroscopic examination, and myocarditis was only revealed following routine histological examination of the heart (figs 2 and 3). Of the cases that presented during the first year of life (in whom sufficient data were available to reliably determine normal ranges for heart weight in this population), the heart weight adjusted for age was greater in the group of infants with myocarditis compared with all infants referred for autopsy (Mann-Whitney U test, p = 0.002). However, the majority of cases with histologically proven myocarditis demonstrated heart weight within the normal range, with only three heart weights above the 95% prediction interval (fig 4). Heart weight is therefore a poor screening test for the detection of myocarditis at autopsy.

Figure 2 Macroscopic appearances of the heart in myocarditis. The gross findings are usually non-specific, and in almost 40% of cases the heart may appear entirely normal.
Figure 3 Pathological features of myocarditis at post-mortem examination. (A) In almost 40% of cases, the heart appeared entirely normal on macroscopic examination. (B) In 18% of cases the heart was described as showing areas of pallor or having a mottled appearance. (C, D) In this series, the diagnosis of myocarditis was made using the Dallas criteria, which require an inflammatory infiltrate (C) associated with myocyte necrosis or degeneration (D).
Figure 4 Heart weights of cases of myocarditis at autopsy plotted on the normal range of infant post-mortem heart weights. Only three (20%) of the weights were above the 95th centile, demonstrating that macroscopic heart weight at post-mortem is not a useful marker of myocarditis in this group.

Virological analyses were carried out in 25 (89%) of the 28 patients with myocarditis, including all 16 sudden death cases, and the virus was detected in nine (36%). Virological detection methods included immunofluorescence, polymerase chain reaction (PCR) and tissue culture, and these were performed upon ante-mortem (in two cases) and/or post-mortem specimens (in 24 cases), the latter comprising varying combinations of lung and heart samples, as well as bowel contents, brain, CSF, liver and others as indicated. Of the 16 sudden deaths, only four revealed the virus, which included parvovirus in one (aged 7.6 years), picornavirus in another (aged 18 days, detected on antemortem CSF sample, post-mortem tissue cultures were negative), and enterovirus in the remaining two (aged 180 and 194 days, respectively). Virus was identified in a further five of the 11 non-sudden deaths, including four cases of enterovirus (one with dual isolation of enterovirus and HSV-1), and one parvovirus. The histological features were similar in all cases, with a significant — though often patchy — interstitial inflammatory cell infiltrate, predominantly lymphocytic, with focal myocyte necrosis and interstitial oedema. There was no subjective difference in morphological appearances between cases with and without viral particles, and between the cases known to be caused by parvovirus or enterovirus infection.

DISCUSSION

The findings of this study have demonstrated that histologically proven acute myocarditis is an uncommon but distinct and recognisable cause of childhood death, representing about 2% of paediatric deaths referred for autopsy to a single specialist centre over a 10-year period. While the majority of cases (54%) occur in children less than 1 year of age, with a median age of 10 months; fatal myocarditis is relatively more common in older children, accounting for around 5% of all childhood deaths over the age of 5 years (95% confidence interval 2.4% to 10.5%) compared with 2% of infant deaths referred for autopsy (95% confidence interval 0.9% to 2.6%). At autopsy, around 60% of cases demonstrate non-specific findings such as cardiac dilatation, myocardial abnormalities or pericardial effusion, but the remainder are macroscopically unremarkable and heart weight is normal in the majority of cases, the diagnosis being entirely dependent on routine histological examination. In this series, more than half of the cases presented with sudden death, of which around one third demonstrated no apparent preceding symptoms and the remainder were associated with a variety of non-specific prodromal symptoms. Of the 11 non-sudden myocarditis deaths, two presented with features of classic dilated cardiomyopathy (DCM).

The clinical presentation of acute myocarditis, which in children carries an overall survival rate of 70–100%,811 can vary from asymptomatic electrocardiographic (ECG) abnormalities, to features of congestive cardiac failure, to sudden death. Previous studies have suggested that viral myocarditis represents a significant cause of sudden cardiac death in previously healthy young persons, including children,36 and that in children between 15–60% of acute-onset DCM may be caused by myocarditis.1013 The frequency of prodromal symptoms, such as fever, upper respiratory tract symptoms or gastroenteritis, are very variable, and are reported to occur in up to 70–90% of all children with histologically proven myocarditis.8 14

Histological sampling remains the gold standard for the diagnosis of myocarditis, which is based on standardised histopathological diagnostic criteria15 and includes histological evidence of myocardial cell injury with an associated interstitial inflammatory infiltrate. Although widely used, the clinical usefulness of these Dallas criteria may be limited by their low sensitivity and specificity, which in part may be due to variations in tissue sampling and in inter-observer variability in the interpretation of the microscopic findings,2 16 and it has been suggested that additional use of immunohistochemical staining for the identification and quantification of T-lymphocyte subtypes and major histocompatibility complex (MHC) antigens, as well as PCR identification of viruses, may improve myocarditis ‘detection’ rates,2 16 17 but the clinical significance of these criteria currently remains uncertain. For the purposes of this study, the diagnosis of acute myocarditis was based on the histological findings of a myocardial interstitial inflammatory infiltrate with associated interstitial oedema and focal myocyte damage. The interstitial infiltrate was composed predominantly of lymphocytes, with no significant eosinophilic or giant cell component, but in all cases, although multifocal in nature, the extent of the infiltrate was subjectively patchy, which may have implications for the diagnosis of myocarditis in living patients in whom the diagnosis is based on limited endomyocardial biopsies. The advantage of using the Dallas criteria in this study is that only unequivocal cases of myocarditis were assigned the diagnosis, but it is accepted that some cases with only a mild patchy lymphocytic infiltrate and no other characteristic histological features may not have been classified as myocarditis.

Several viruses have been implicated in the pathogenesis of myocarditis, the prototype being enterovirus Coxsackie B, in which disease is believed to be mediated via a Coxsackie viral protease that can cleave dystrophin to damage myocyte cytoskeletal structural integrity.18 Adenovirus, which shares a common cellular receptor with Coxsackie virus,18 has been identified in 25–40% of cardiac samples of children with acute myocarditis.13 14 A number of other viruses have also been implicated, including cytomegalovirus (CMV), parvovirus B19 virus, herpes simplex virus (HSV), influenza virus, Epstein–Barr virus (EBV), hepatitis C virus (HCV), respiratory syncytial virus (RSV) and human herpesvirus 6 (HHV6),1 2 13 1924 and dual infections with more than one virus have been described.13 In addition, the human immunodeficiency virus (HIV) has been associated with myocarditis.2 In this study, direct virological assessment of post-mortem tissue, usually by means of immunofluorescence, PCR and/or cell culture, was carried out in the majority of cases, and virus was detected in around one third. Previous studies of virological analysis in myocarditis have reported similar detection rates, ranging from 38%13 to around 45%,19 although Martin et al14 detected virus in 68% of myocardial samples. Interestingly, in the present series, of the nine patients in whom a viral aetiology could be demonstrated, parvovirus was detected in two (22%). Parvovirus B19 is now increasingly recognised as a cause of myocarditis in childhood.2023 The virus typically infects erythroid precursors via the B19 receptor (P antigen) and is usually associated with hydrops fetalis, generally thought to be due to the anaemia caused by the destruction of parvovirus-infected red cell precursors.22 However, the B19 receptor has also been demonstrated on fetal myocytes, and it has been suggested that the pathogenesis of the fetal hydrops may partly be related to an intrauterine myocarditis.21 22 The mechanism of parvovirus-associated myocarditis in older children remains controversial, and it is uncertain whether the myocarditis is due to direct viral infection of cardiac myocytes or an autoimmune-mediated secondary event.21 22

In viral myocarditis, the pathogenesis is thought to include three different phases: an initial virus-induced cytopathic myocardial injury, followed by autoimmune-mediated myocardial damage and the subsequent development of dilated cardiomyopathy in a proportion of cases.2 25 Myocarditis may also affect the cardiac conduction system, potentially leading to various arrhythmias.26 The mechanism of myocarditis-related death is likely to be via one of two possible pathways. Firstly, direct or indirect myocardial damage as a consequence of viral cytopathic effects or the effects of inflammatory cell mediators, resulting in myocyte dysfunction and the development of acute or subacute cardiac failure. This is the probable mechanism leading to chronic cardiac failure in cases of presumed myocarditis progressing to cardiac transplantation. Second, and more likely in the clinical setting of sudden and unexpected death of a patient with no preceding history of cardiac failure, death may be due to an inflammatory-mediated cardiac arrhythmia. Indeed, in one study of infants who died suddenly and unexpectedly of natural causes, a focal inflammatory infiltrate was found to involve the upper part of the interventricular septum, including parts of the conduction system, in 50% of cases with myocarditis.6

The total population affected by viral myocarditis is clearly greater than reported in this autopsy series, since most affected children will survive, and up to two thirds will make a complete recovery.810 27 Nevertheless, the present study has determined the frequency of fatal myocarditis at post-mortem examination, which represents about 2% of such paediatric deaths. All ages are affected, although myocarditis is relatively more common as a cause of death in older children, accounting for around 5% of childhood deaths over the age of 5 years. In the majority, heart weight is normal, and in almost 40% of cases there are no macroscopic cardiac abnormalities; the diagnosis requires routine histological sampling of the myocardium.

What is already known on this topic

  • Myocarditis can present with sudden unexpected death in infancy and childhood.

  • The majority of cases of myocarditis are caused by viral infections, with enterovirus and adenovirus being the most common viruses implicated in childhood myocarditis.

What this study adds

  • Myocarditis is a rare cause of death, representing around 2% of paediatric deaths referred for autopsy.

  • The heart may be entirely normal on macroscopic examination; therefore, routine histological sampling of the heart is required for its detection.

  • Heart weight is a poor screening test for myocarditis.

REFERENCES

View Abstract

Footnotes

  • Funding: This study was supported by a grant from The Foundation for the Study of Infant Deaths (FSID)

  • Competing interests: None.

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.