OUP user menu

An overview of the evolution of enterovirus 71 and its clinical and public health significance

Peter C. McMinn
DOI: http://dx.doi.org/10.1111/j.1574-6976.2002.tb00601.x 91-107 First published online: 1 March 2002


Since its discovery in 1969, enterovirus 71 (EV71) has been recognised as a frequent cause of epidemics of hand-foot-and-mouth disease (HFMD) associated with severe neurological sequelae in a small proportion of cases. There has been a significant increase in EV71 epidemic activity throughout the Asia–Pacific region since 1997. Recent HFMD epidemics in this region have been associated with a severe form of brainstem encephalitis associated with pulmonary oedema and high case-fatality rates. The emergence of large-scale epidemic activity in the Asia–Pacific region has been associated with the circulation of three genetic lineages that appear to be undergoing rapid evolutionary change. Two of these lineages (B3 and B4) have not been described previously and appear to have arisen from an endemic focus in equatorial Asia, which has served as a source of virus for HFMD epidemics in Malaysia, Singapore and Australia. The third lineage (C2) has previously been identified [Brown, B.A. et al. (1999) J. Virol. 73, 9969–9975] and was primarily responsible for the large HFMD epidemic in Taiwan during 1998. As EV71 appears not to be susceptible to newly developed antiviral agents and a vaccine is not currently available, control of EV71 epidemics through high-level surveillance and public health intervention needs to be maintained and extended throughout the Asia–Pacific region. Future research should focus on (1) understanding the molecular genetics of EV71 virulence, (2) identification of the receptor(s) for EV71, (3) development of antiviral agents to ameliorate the severity of neurological disease and (4) vaccine development to control epidemics. Following the successful experience of the poliomyelitis control programme, it may be possible to control EV71 epidemics if an effective live-attenuated vaccine is developed.

  • Enterovirus 71
  • Molecular epidemiology
  • Molecular genetics of virulence
  • Clinical significance

1 Introduction

Enterovirus 71 (EV71) belongs to the human Enterovirus A species of the Enterovirus genus within the family Picornaviridae [1]. Virions consist of a non-enveloped capsid surrounding a core of single-stranded, positive-polarity RNA approximately 7.5 kb in size. The viral capsid is icosahedral (T=1) in symmetry and is composed of 60 identical units (protomers) each consisting of the four structural proteins VP1–VP4. The complete nucleotide sequence of the EV71 prototype strain BrCr has been determined [2]. The single open reading frame (ORF) encodes a polyprotein of 2194 amino acids and is flanked by untranslated regions (UTRs) at the 5′ and 3′ ends; a variable length poly-A tract is located at the terminus of the 3′ UTR. The polyprotein is subdivided into three regions, P1, P2 and P3. P1 encodes four viral structural proteins 1A–1D (VP1–VP4); P2 and P3 encode seven non-structural proteins 2A–C and 3A–D. The functions of the 11 individual EV71 proteins are thought to be identical to that determined for poliovirus and other non-polio enteroviruses [3]. The genome structure of EV71 is shown in Fig. 1.

Figure 1

Genome structure of EV71. The single ORF is flanked by UTRs at the 5′ and 3′ ends, a variable length poly-A tail is found at the 3′ UTR. The ORF is divided into three regions: the P1 region encodes four structural proteins VP1–VP4, the P2 and P3 regions encode seven non-structural proteins 2A–2C and 3A–3D, respectively. Adapted from Brown and Pallansch [2].

Since the initial description of EV71 in 1974 [4], outbreaks of infection with this virus have occurred periodically throughout the world [512]. EV71 infection manifests most frequently as the childhood exanthem known as hand-foot-and-mouth disease (HFMD) and is considered to be clinically indistinguishable from HFMD caused by Coxsackievirus A16 (CA16). Molecular studies of the evolution of human enteroviruses have shown that EV71 and CA16 share a close genetic relationship, and, together with CA7 and CA14, form a distinct genetic sub-group within the human Enterovirus A species [1]. Despite the close genetic relationship between EV71 and CA16, EV71 has a propensity to cause neurological disease during acute infection [5, 6], a feature not observed in CA16 infections. Children under 5 years of age are particularly susceptible to the most severe forms of EV71-associated neurological disease, including aseptic meningitis, brainstem and/or cerebellar encephalitis, and acute flaccid paralysis (AFP). The neurological complications of EV71 infection may occasionally cause permanent paralysis or death. Several large epidemics of severe EV71 infection in young children, including numerous cases of fatal brainstem encephalitis, have recently been reported from Southeast Asia [1315], raising concern that both the prevalence and virulence of EV71 may be increasing.

In this review, I will present some perspectives on the history, epidemiology and evolution of EV71. I will also review the spectrum of neurological and other diseases associated with EV71 epidemics, with particular emphasis on AFP and on the severe forms of brainstem encephalitis that have occurred during recent large EV71 epidemics in Asia. The review will conclude with an analysis of the current state of knowledge on the molecular genetics of EV71 virulence and a discussion of the prospects for control of EV71 infections through public health surveillance, new antiviral agents and research into the development of vaccines.

2 History of EV71

2.1 Early epidemic activity

EV71 was first identified in 1969 in California, when it was isolated from the stool of an infant suffering from encephalitis [4]. Over the next 3 years, EV71 was isolated from a further 23 cases of severe neurological disease in California [16]. EV71 was first isolated outside of California in 1972, 28 cases of EV71 infection were identified in New York between 1972 and 1977 [16, 17], 19 of which were associated with severe neurological disease, including meningitis, encephalitis and AFP. The first isolation of EV71 outside of the USA was made in Australia, during an epidemic of aseptic meningitis in Melbourne between 1972 and 1973 [10]. The first linkage of EV71 with HFMD was made during small epidemics in both Sweden [18] and Japan [19, 20] during 1973. A second EV71 epidemic in Japan during 1978 [9] resulted in a large number of HFMD cases in addition to cases of aseptic meningitis, acute cerebellar ataxia and AFP.

2.2 Endemic circulation of EV71

Many of the early reports of EV71 epidemiology were studies of outbreaks and may thus have presented a distorted picture of the disease associations, in the particular neurovirulence, of EV71. Two small prospective studies have attempted to assess the importance of endemic circulation of EV71 as a cause of serious enterovirus disease [12, 21]. A study of endemic circulation of EV71 in two provinces in Brazil over a 3-year period (1988–1990) [12] identified EV71 as the cause of a significant proportion of acute neurological disease cases, including poliomyelitis (6.2%), Bell's palsy (7.1%), acute cerebellar ataxia (20%) and Guillain-Barré syndrome (GBS; 5.2%). A Canadian study [21] showed that endemic circulation of EV71 (during 1998) was associated with a broader range of disease than seen in Brazil [12], including undifferentiated febrile illness, HFMD and mild respiratory infections, in addition to acute neurological disease (aseptic meningitis, encephalitis, acute cerebellar ataxia). Thus, it is clear that EV71 may undergo both endemic and epidemic circulation and that this may be associated with differing distributions of EV71-associated disease.

2.3 The European epidemics, 1975–78

The first reports of large and severe epidemics of encephalitis and AFP due to EV71 came from Bulgaria in 1975 [22] and Hungary in 1978 [23]. The Bulgarian epidemic occurred between May and September 1975 [22]. Altogether, 705 cases of febrile illness attributable to EV71 infection were recorded, of which 77.3% (545 cases) were identified as aseptic meningitis and 21.1% (149 cases) as AFP. Of the paralytic cases, bulbar disease was documented in 9.6% (68 cases) and a fatal outcome was reported in 6.2% (44 cases) [22], almost exclusively in cases of bulbar disease. The majority of clinical cases occurred in children under 5 years old, with 93% of the paralytic cases and 83.8% of the total cases occurring in this age group [22]. Many of the cases of bulbar disease were reported as being rapidly fatal, with death occurring between 10 and 30 h after illness onset. Post-mortem examination of the fatal cases revealed lesions typical of bulbar poliomyelitis, predominantly within the medulla oblongata and anterior horns of the spinal cord [24]. It is of interest that, despite the size and severity of the Bulgarian epidemic, no cases of HFMD were reported. A total of 92 EV71 strains were isolated during the Bulgarian epidemic, including 37 strains from the brainstem and spinal cord of post-mortem cases [22]. EV71 was isolated from 25.3% of clinically diagnosed cases and from 100% of fatal cases. Serological diagnosis was achieved by neutralisation assay with a total of 71.9% of clinically diagnosed cases having either a single high titre antibody level (38.8%) or a four-fold rise in antibody (33.2%). The initial identification of EV71 (strain 258) in the Bulgarian epidemic was made by Melnick et al. [25] and a formalin-inactivated vaccine was prepared from this strain [22]. However, the vaccine was never put into use, as the epidemic declined before the vaccine was ready for use. The widespread use of Sabin oral polio vaccine (OPV) was instituted in Bulgaria in response to the epidemic (i.e. prior to the realisation that EV71 was the cause of the epidemic), as poliovirus was considered to be the cause of the outbreak. The widespread use of OPV was considered to have contributed to the rapid decline in the EV71 epidemic [26], although evidence for this effect was not provided.

A second large EV71 epidemic occurred in Hungary between May and September 1978 [23]. A total of 826 cases of aseptic meningitis and 724 cases of encephalitis were reported, with the latter group including cases of cerebellar ataxia and AFP. EV71 infection was confirmed in 323 cases, either by virus isolation from central nervous system (CNS) or peripheral samples (44 cases), or by serum neutralisation assay. As was found in the Bulgarian epidemic [22], only a minority of cases demonstrated a four-fold or greater rise in EV71-specific antibody, with the majority of cases showing high titre EV71 antibody in the first serum sample. Cases of severe neurological disease in which EV71 infection was proven included 13 cases of poliomyelitis-like paralysis (identified at post-mortem), 145 cases of encephalitis and 161 cases of aseptic meningitis [23]. In contrast to the Bulgarian epidemic, four cases of HFMD associated with EV71 infection were documented in Hungary [23].

2.4 EV71 in the Asia–Pacific region

As noted above, EV71-associated HFMD epidemics were identified in Japan in 1973 [19, 27] and 1978 [9, 28, 29]. A characteristic feature of both epidemics was the low prevalence of acute neurological disease linked to EV71 infection. A small EV71 outbreak, in which several cases of AFP were identified, was recorded in Hong Kong in 1985 [7]. The first report of EV71 activity in China was during a HFMD epidemic in Hubei Province during the winter of 1987 [30]. No cases of AFP or aseptic meningitis were identified during this epidemic. A small HFMD and aseptic meningitis outbreak due to EV71 was also identified in Singapore in 1987 [31]. Epidemics of EV71 associated with both HFMD and acute neurological disease occurred in Melbourne, Australia, in 1973 [10] and 1986 [11]. Furthermore, there appears to have been continuous low-level endemic circulation of EV71 in eastern Australia, associated with febrile illnesses, HFMD and acute neurological disease, from 1972 until the present (D. Dwyer, personal communication).

Since 1997, several large epidemics and high-level endemic circulation of EV71 have been reported in the Asia–Pacific region. The first epidemic occurred in 1997 in Sarawak [32, 33], followed by smaller outbreaks in Japan [13], peninsular Malaysia [15] and Singapore in 1998 [34]. These outbreaks were associated with numerous cases of HFMD and herpangina in young children, and neurological complications such as aseptic meningitis, AFP and cerebellar ataxia occurred in a proportion of cases. An alarming feature of these epidemics was the appearance of a syndrome of rapidly fatal neurogenic pulmonary oedema associated with severe brainstem encephalitis [35, 36]. Thirty-four deaths from this syndrome occurred in Sarawak [33] and four fatal cases were reported in peninsular Malaysia [15].

In 1998, the largest EV71 epidemic reported to date occurred in Taiwan [14, 37, 38]. The epidemic occurred in two waves, the first (March–July) involved the whole island, and the second (September–November) was mainly confined to the south. A total of ∼130×103 cases of HFMD and herpangina were reported to the national sentinel surveillance programme [39] over a period of 8 months. As the national sentinel surveillance network involves ∼9% of Taiwan's primary physicians, it has been estimated that ∼1.5 million cases of HFMD occurred during the epidemic [40]. Furthermore, it has been estimated that ∼29% of EV71 infections in Taiwan were symptomatic [40]. Thus, the estimate of asymptomatic infections is 2.1 million and the total number of infections 3.6 million, or 43% of the population at risk [40]. Risk factors for acquisition of EV71 infection in Taiwan were identified as (1) having an older sibling with positive EV71 serology, (2) age between 6 months and 3 years of age, (3) number of children in the family, and (4) history of contact with a case of HFMD or herpangina [41]. These risk factors are consistent with a predominantly faecal–oral mode of transmission of EV71 during the Taiwanese epidemic, although other forms of transmission, such as respiratory droplet transmission, were considered possible on epidemiological grounds [42].

Virological studies identified the cause of the HFMD and herpangina epidemic in Taiwan as due to both EV71 and CA16 [14, 42], with EV71 isolated from ∼2/3 of the cases. A total of 405 cases of severe neurological disease due to EV71 infection were identified, of which 78 were fatal mainly due to the development of neurogenic pulmonary oedema [14, 38, 43]. It has been estimated that the proportion of severe and fatal cases was 0.1 per thousand EV71 infections [42]. Radiological [43, 44], histopathological [45, 46], virological [4547] and serological [37, 45] studies clearly established that the cases presenting with acute pulmonary oedema and other neurological diseases in Taiwan were due to brainstem encephalitis resulting from acute EV71 infection.

A large outbreak of EV71 infection occurred in Perth, Western Australia, during 1999 [48]. Approximately 6000 cases of HFMD occurred over a 6-month period (March–August) and 29 cases of severe neurological disease were identified. The rate of neurological disease in this epidemic was estimated to be ∼1 per 1000 EV71 infections [48]. Similar to the experience in Taiwan, virus isolation studies established that HFMD was caused by both EV71 and CA16 (in roughly equal proportion) but that cases of neurological disease were associated exclusively with EV71 infection. The spectrum of neurological disease seen in Perth included aseptic meningitis, acute cerebellar ataxia, brainstem encephalitis and AFP. Interestingly, six cases of post-infectious neurological disease following acute EV71 infection were identified, including two cases of GBS, two cases of transverse myelitis, one case of opso-myoclonus syndrome and one case of benign intracranial hypertension [48]. Importantly, no cases of fatal neurogenic pulmonary oedema were observed during the Perth epidemic.

High-level EV71 epidemic activity continued in the Asia–Pacific region during 2000–01. EV71 continues to circulate endemically in Sarawak (M.J. Cardosa, personal communication), with reports of sporadic cases of HFMD and encephalomyelitis. A large outbreak of HFMD was reported in Singapore and southern peninsular Malaysia in the latter half of 2000, during which several fatal cases of encephalomyelitis were reported in both countries [49]. EV71 was identified as the sole cause of the epidemic in Singapore, whereas both EV71 and echovirus 7 were isolated from cases of HFMD and encephalomyelitis in peninsular Malaysia [49, 50]. Epidemic activity has also been reported in southeastern Australia during 2001, with numerous cases of HFMD and several cases of severe encephalomyelitis reported to date.

3 Clinical features of EV71 infection

3.1 HFMD and herpangina

Epidemics of HFMD have been associated most frequently with CA16 [51], although EV71 has been increasingly recognised as a cause of this disease, other enterovirus serotypes known to cause HFMD include CA5, CA9 and CA10 [51]. The illness is characterised by 3–4 days of fever and the development of a vesicular enanthem on the buccal mucosa, tongue, gums and palate and a papulovesicular exanthem on the hands, feet and buttocks. Clinical observations from EV71 epidemics in Japan [13], Malaysia (K.B. Chua, personal communication) and Western Australia [48] indicate that the HFMD rashes due to CA16 and EV71 may differ. The rash associated with CA16 infection is characterised by the formation of larger vesicles than in EV71 infection, in which the rash is more frequently papular and/or petechial, often with areas of diffuse erythema on the trunk and limbs.

There also appear to be marked differences in the prevalence of HFMD and CNS disease during specific EV71 epidemics. For example, HFMD was identified very rarely during the EV71-associated encephalitis and poliomyelitis epidemics in Europe during the 1970s [22, 23], but was the predominant disease associated with EV71 infection in Japan over the same time period [9, 19, 2729], in which only small numbers of mild neurological disease cases were identified. Both HFMD and neurological disease were prevalent during the recent EV71 epidemics in the Asia–Pacific region [14, 33, 48], however the proportion of acute neurological disease cases presenting concurrently with HFMD varied widely. For example, HFMD was present at the onset of CNS disease in only 7% of cases during the Perth epidemic [48] whereas the majority (80%) of CNS disease cases in Taiwan were noted to have concurrent HFMD [52]. Furthermore, it is of interest that occasional EV71 epidemics have been associated predominantly with HFMD, with few reports of associated neurological disease [19], suggesting that EV71 strains circulating during specific epidemics vary widely in their dermatotropism and neurotropism, and that these two pathogenetic characteristics of EV71 infection are not rigidly linked.

In addition to HFMD, EV71 was identified as a cause of herpangina during epidemics in Hong Kong [7], Japan [13, 19] and Taiwan [14, 52]. Herpangina is an illness characterised by an abrupt onset of fever and sore throat, associated with the development of raised papular lesions on the mucosa of the anterior pillars of Fauces, soft palate and uvula [51]. Herpangina is a common manifestation of acute Coxsackievirus A infections in young children [51], with CA8, CA10 and CA16 being implicated most frequently. EV71-associated herpangina was very prevalent during the 1998 Taiwanese epidemic [14, 47, 52] and was the second most common diagnosis after HFMD. Approximately 10% of children with EV71-associated neurological disease in Taiwan had an initial or concurrent diagnosis of herpangina [52].

3.2 Neurological disease

Since its first identification, EV71 has been recognised as highly neurotropic and associated with a diverse range of neurological diseases such as aseptic meningitis, brainstem and/or cerebellar encephalitis, AFP and several post-infectious neurological syndromes (reviewed in [6]). The link between EV71 infection and AFP has been established by numerous studies over many years [5, 7, 8, 14, 16, 22, 23, 48]. However, only a small proportion of these studies have provided rigorous radiological [5, 52, 53] or histopathological [22, 24] evidence for the induction of paralysis through infection and destruction of anterior horn motor neurones of the spinal cord – a process identical to that of poliovirus [54]. While it is clear that the pathogenesis of AFP in the Bulgarian EV71 epidemic [22, 24] and in some cases of paralysis in Taiwan [55] was very similar to that of poliomyelitis, many of the other studies failed to rigorously demonstrate a mechanism for induction of paralysis. As was demonstrated conclusively by Albert Sabin in the case of clinically diagnosed poliomyelitis [56], it is likely that many clinically diagnosed cases of EV71-associated AFP are not due to cytopathic damage to anterior horn motor neurones (i.e. poliomyelitis), but are due to other neuropathological mechanisms [57].

AFP associated with EV71 infection appears to be milder and associated with higher rates of complete recovery than occurs after poliovirus infection [5, 7, 48]. As noted above, the more varied presentation and milder clinical manifestations of AFP due to EV71 (and other non-polio enterovirus) infection are likely to reflect causation though more than one neuropathological mechanism [56, 57], in particular through immunopathological processes. For example, prospective studies of the causes of AFP in Brazil (1988–1990) showed that EV71 infection (defined by virus isolation or a four-fold rise in EV71-specific neutralising antibody titres) was the cause of 6% of acute poliomyelitis-like disease (11/175) and 5% of GBS (3/58 cases) [12, 58]. In the 1999 Perth EV71 epidemic, only one of five cases of AFP was due to poliomyelitis-like disease, with two cases identified as transverse myelitis and two as GBS [48]. Two other cases of GBS linked to EV71 infection have been identified: one in Australia [10] and the other in the USA [6]. Furthermore, a clinical and MR imaging study of EV71-associated AFP showed the presence of transient MR changes in the anterior horns and ventral roots of the spinal cord in six of seven cases [55]. These changes are almost identical to those seen in cases of clinically and electrophysiologically diagnosed GBS [59, 60]. Four of the children in this study recovered completely and three had mild residual weakness 1 year after onset [55]. However, as these children were not diagnosed with GBS and nerve conduction studies were not reported, the causes of the AFP in this study remain unclear. Nevertheless, it is clear that EV71 may induce AFP by several mechanisms in addition to virus-mediated destruction of anterior horn motor neurones, and this is reflected in the more varied clinical presentation of EV71-associated AFP than seen during poliovirus infection.

The most severe neurological manifestation of EV71 infection is brainstem encephalitis. This disease occurs most frequently as an extension of spinal cord disease [15, 24, 46, 48] but may occur in isolation [44, 52]. MR imaging and post-mortem studies indicate that the medulla oblongata, reticular formation, pons and midbrain structures are most frequently involved [15, 24, 44, 48]. Children with this infection usually present with myoclonus, tremor, ataxia, nystagmus and cranial nerve palsies [44, 48, 52]. During the 1998 Taiwanese epidemic, brainstem encephalitis was classified into three grades (I, II, III) on clinical criteria [44]. Grade I (mild) brainstem encephalitis was characterised by generalised myoclonus and ataxia, was associated with 100% recovery and only 5% of children developed permanent neurological sequelae. Grade II disease was associated with cranial nerve palsies in addition to myoclonus/ataxia, 100% of children with grade II disease recovered and 20% developed permanent neurological sequelae. Grade III disease was associated with a rapid onset of cardiopulmonary failure (‘neurogenic pulmonary oedema’), 80% of children with grade III disease died and all surviving children developed significant neurological sequelae. This classification appears to be a useful indicator of the prognosis of EV71-associated brainstem encephalitis. Examination of clinical data from the Malaysian HFMD epidemics indicates that children with illnesses consistent with grade I or II disease had a generally good outcome, whereas children whose illness was consistent with grade III disease had a high mortality [15]. Several children in the 1997 Japanese [13] and 1999 Perth [48] epidemics presented with grade I or II brainstem encephalitis, no children with grade III encephalitis were identified, and no fatal cases were recorded in either epidemic. The clinical features and pathogenesis of grade III brainstem encephalitis/neurogenic pulmonary oedema are presented in the following section.

Acute cerebellar ataxia has been linked to EV71 infection in many previous epidemics [9, 12, 13, 21, 23, 48]. This disease usually presents as truncal ataxia with or without nystagmus and is typically accompanied by full recovery. An MR imaging study of two cases of EV71-associated cerebellar ataxia [48] indicated that the acute disease was associated with inflammation of grey matter in one or both cerebellar hemispheres and that variable degrees of cerebellar cortical atrophy developed upon recovery, despite the complete recovery of function.

3.3 Neurogenic pulmonary oedema

Prior to the large epidemics in the Asia–Pacific region, only one case of EV71-associated brainstem encephalitis presenting with neurogenic pulmonary oedema had been described [61]. This case occurred in a 3-year-old girl from Connecticut, USA, in 1995 and was fatal; EV71 was isolated from the CSF and spinal cord at post-mortem [61]. In retrospect, it is possible that the rapidly fatal cases identified in the Bulgarian epidemic [22, 24] may have been due to a similar syndrome. It is of interest that a similar syndrome of pulmonary oedema and brainstem encephalitis was frequently described in association with cases of acute bulbar poliomyelitis as early as 1957 [62]. The poliovirus-associated cases had a high mortality and post-mortem studies revealed specific inflammatory lesions in the dorsal nucleus of the vagus nerve and in the reticular formation of the brainstem in all fatal cases [62].

The most striking clinical characteristic of type III brainstem encephalitis is its rapid progression and high mortality [44, 6365]. Typically, children develop tachycardia, tachypnoea and cyanosis between 2 and 5 days of the onset of fever, HFMD or herpangina [15, 42]. The mortality from this condition is 80–90%[44, 52, 64], with most children dying within 12–18 h of the onset of the syndrome [44, 6365]. In Taiwan, the presence of hyperglycaemia and AFP at the time of admission to hospital correlated strongly with the development of neurogenic pulmonary oedema [41, 64]. The authors suggested that these risk factors indicated CNS invasion by EV71. In the case of hyperglycaemia, they argued that dysregulation of glucose homeostasis resulted from autonomic nervous system dysfunction and that this condition ultimately led to pulmonary oedema and shock [41, 42, 64].

Several post-mortem studies of neurogenic pulmonary oedema due to EV71 infection have been published [15, 33, 44, 52, 65]. In each case, disease appears confined to the brainstem, with histological evidence of acute inflammatory encephalitis, isolation of EV71 or identification of EV71 antigen within neurones [15, 52, 63, 65, 66]. These studies support the hypothesis that the pulmonary oedema is of neurogenic origin and is secondary to autonomic dysfunction resulting from infection of specific regulatory structures within the brainstem. These findings are supported by neuroradiological evidence of brainstem pathology in many cases of neurogenic pulmonary oedema [44, 52]. The post-mortem studies also consistently show that brainstem lesions are accompanied by intense neutrophil and mononuclear cell inflammatory infiltrates [15, 33, 44, 45, 52, 65]. Furthermore, experimental inoculation of Cynomolgus monkeys with EV71 results in the development of paralytic disease associated with intense inflammatory lesions in the spinal cord and medulla oblongata [67, 68]. Thus, although EV71 is known to be cytopathic in cell culture [69], it is possible that inflammation contributes to the pathogenesis of encephalitis in vivo. In support of this hypothesis, a Taiwanese group has shown that children with severe EV71 encephalitis were significantly more likely to possess a certain cytotoxic T lymphocyte antigen haplotype (CTLA-4) than children who developed mild EV71 infections [70]. The authors suggested that children whose T cells expressed the CTLA-4 antigen developed cell-mediated immune responses to EV71 infection that predisposed them to the development of severe disease [70].

Despite radiological and histological evidence of brainstem encephalitis in cases of fatal neurogenic pulmonary oedema and immunohistochemical evidence of the direct involvement of EV71 in brainstem encephalitis, the cause of death among children during the 1997 outbreak in Sarawak remains controversial. Although many children died from a syndrome of rapidly progressive pulmonary oedema [33, 63] similar to that observed in peninsular Malaysia and Taiwan, a clinical diagnosis of acute myocarditis was made in many cases. In addition, both EV71 and a novel group B adenovirus were isolated from sterile (including brain and heart) and non-sterile sites from both pre- and post-mortem specimens of several fatal cases [33]. The authors suggest that the novel adenovirus may have played a causative role in the fatal cases, either as the primary pathogen or by way of interaction with EV71. Unfortunately, data currently available in the published literature do not allow a rigorous assessment of the role of adenovirus infection in the pulmonary oedema syndrome. However, it is clear from post-mortem studies of fatal cases in Taiwan [44, 52], peninsular Malaysia [15] and Hong Kong [65] that brainstem encephalitis due to EV71 infection is solely sufficient to cause neurogenic pulmonary oedema.

3.4 Other EV71-associated diseases

Non-specific febrile illnesses are a common clinical manifestation of enterovirus infection in young children [71] and have been described in association with EV71 [14, 21, 52]. This syndrome is particularly common in infants less than 6 months of age and is often of sufficient severity to warrant admission to hospital for investigation and empirical therapy of suspected bacterial sepsis or meningitis [71].

Acute respiratory disease (other than neurogenic pulmonary oedema) has been linked to EV71 infection in Australia [11], Canada [21] and during the 1998 Taiwanese epidemic [72]. Respiratory diseases associated with EV71 infection include pharyngitis, croup, bronchiolitis and pneumonia [11, 21, 72]. Most respiratory infections occur in children 1–3 years of age and are often of sufficient severity to require hospitalisation.

A single case of intrauterine infection with EV71 leading to symptomatic foetal infection (hydrocephalus, hepatosplenomagaly) and stillbirth has been reported [73]. In this case, EV71 was isolated from amniotic fluid, EV71 antigen was detected within neurones of foetal brainstem structures by immunohistochemistry and EV71 RNA was detected in several maternal and foetal tissues by reverse transcription-PCR, providing direct evidence of foetal infection. However, the causal link between intrauterine EV71 infection and foetal abnormality remains unproven.

A summary of the acute neurological diseases and other clinical syndromes commonly, uncommonly or rarely associated with EV71 infection is presented in Table 1.

View this table:
Table 1

Clinical syndromes associated with EV71 infection

Clinical syndromeAssociation with EV71 infectionReferences
Neurological disease:
Aseptic meningitisVery commona[6, 914, 16, 18, 21, 24, 43, 44, 47, 48, 52, 105]
Poliomyelitis-like paralysisCommon[5, 6, 9, 16, 22, 24, 33, 43, 52, 53, 55, 105]
Brainstem encephalitisCommon[6, 13, 15, 24, 35, 43, 44, 48, 52, 105]
‘Neurogenic pulmonary oedema’Commonb[15, 33, 4144, 52, 61, 63, 64, 66, 105]
Cerebellar ataxiaUncommon[9, 12, 13, 21, 23, 48]
GBSUncommon[6, 10, 12, 48, 55, 58]
Transverse myelitisRarec[48]
Opso-myoclonus syndromeRarec[48]
Benign intracranial hypertensionRarec[48]
Febrile rash illnesses:
HFMDVery commond[6, 9, 13, 14, 16, 19, 2729, 33, 43, 47, 48, 52]
HerpanginaCommon[7, 13, 14, 19, 47, 52]
Other exanthemsUncommon[6, 14, 21, 52, 71]
Other syndromes:
Acute respiratory diseaseUncommon[11, 16, 21, 72]
Intrauterine infectionRarec[73]
  • aAseptic meningitis has been described in all reported epidemics of EV71 infection.

  • bNeurogenic pulmonary oedema was first described in association with EV71 infection in 1995 [61] and has been frequently associated with EV71 epidemics in the Asia–Pacific region since 1997.

  • cOnly one example reported in the literature.

  • dHFMD has been described in all reported epidemics of EV71 infection, with the sole exception of the 1975 outbreak in Bulgaria [22].

4 Molecular epidemiology and evolution

A small study of the molecular epidemiology of EV71 strains isolated from cases of AFP, encephalitis and HFMD in Japan, Taiwan, Bulgaria, Hungary and the USA (BrCr) was undertaken in 1984 using capsid polypeptide analysis and RNase T1 oligonucleotide mapping [68]. This study showed that there were marked molecular differences between strains isolated from different geographic regions. However, the methods used in the study were insufficiently sensitive to identify specific EV71 genetic lineages or molecular markers of virulence [68]. In another small study, the 5′ UTRs of an EV71 strain isolated in China in 1987 and the prototype BrCr (California, 1969) were compared by nucleotide sequencing [30]. This study provided further evidence of significant genetic variation between EV71 strains separated in time and by geographic location.

Our understanding of the molecular epidemiology and evolution of EV71 has increased enormously since the publication of a seminal study by Brown et al. [74]. This study compared the complete VP1 gene sequences of 113 EV71 isolates from around the world. However, it is unfortunate that EV71 strains from the Bulgarian and Hungarian epidemics were not included in this study. The VP1 gene is considered to be the most informative region of the enterovirus genome for studying evolutionary relationships because (1) the VP1 protein is the major viral neutralisation determinant and thus has a high degree of antigenic and genetic diversity that correlates with viral serotype [75], and (2) homologous recombination has not been shown to occur within the VP1 gene [75]. Consequently, the VP1 gene has been found most useful in distinguishing within and between enterovirus serotypes [75, 76]. Brown et al. [74] showed the development of three independent genetic lineages of EV71 (A, B, C) over a 30-year period. Virus isolates within these three genetic lineages (genogroups) share >92% nucleotide sequence identity, whereas the nucleotide sequence identity between the genogroups is 78–83%. Genogroup A includes a single virus, the prototype strain BrCr [2, 4]. All other EV71 strains examined in the study belonged to either genogroup B or C, both of which were divided into two sub-lineages B1/B2 and C1/C2, respectively.

Numerous reports on the molecular epidemiology of recent EV71 strains from the Asia–Pacific region have been published [34, 7782]. Several different regions of the EV71 genome have been used for analysis, including the 5′ UTR [77, 78], VP4 gene [78, 79, 81] and VP1 gene [34, 80, 82]. The studies indicate that at least four genetic lineages have circulated widely in the Asia–Pacific region since 1997 – four in Southeast Asia/Australia and two in Taiwan. In addition, there does not appear to be a single ‘neurovirulent’ genotype associated with the severe and fatal cases, as at least three distinct genotypes have been isolated from fatal cases in Sarawak, peninsular Malaysia, Japan and Taiwan.

Shih et al. [80] compared the VP1 gene sequences of 18 strains from the 1998 Taiwanese epidemic and showed that the majority of viruses, including all strains isolated from fatal cases, belong to a new lineage within genogroup C2. Interestingly, a single isolate (TW/1743/98) belonged to genogroup B [80]. The findings of this study correlate closely with those of McMinn et al. [82], who confirmed that the majority of Taiwanese epidemic strains belong to genogroup C2 with a smaller contribution from viruses belonging to genogroup B. Furthermore, Chu et al. [81] studied 23 Taiwanese EV71 isolates (based on VP4 sequencing) and also showed that the 1998 Taiwanese epidemic isolates belong to two genetic clusters. The smaller cluster was consistent with ‘VP1-based’ genogroup B and the larger cluster with ‘VP1-based’ genogroup C. The authors argued that the large cluster belonged to a new lineage (C3) within genogroup C [81]. However, as indicated above, VP1-based phylogenetic studies using several of the same virus isolates indicate that these strains belong to genogroup C2 [80, 82].

McMinn et al. [82] compared the complete VP1 gene sequences of 66 EV71 strains isolated in Malaysia, Singapore, Taiwan and Western Australia between 1997 and 2001. This study showed that two previously unidentified lineages within genogroup B (B3 and B4) circulated in Southeast Asia between 1997 and 2001 and confirmed that viruses belonging to genogroup C2 were the primary cause of the Taiwanese epidemic. A fourth genogroup (C1), previously prevalent in North America and eastern Australia [74], appears to have undergone low-level endemic circulation within Southeast Asia and Western Australia between 1997 and 2001 [82]. Between 1997 and 1999, genogroup B3 viruses were the predominant strain in Southeast Asia and were identified as the major cause of epidemics in Sarawak in 1997, Singapore in 1998 and Western Australia in 1999 [82], but have not been isolated in the region since 1999. Viruses belonging to genogroup B4 were identified in a small proportion of cases in Singapore in 1997 [82], peninsular Malaysia in 1997–98 (L. Herrero and P. McMinn, unpublished data) and Taiwan in 1998 [80, 82], indicating that this genogroup was widespread, although not predominant, throughout the Asia–Pacific region until it became the focus of large epidemic activity in Malaysia (peninsular and Sarawak) and Singapore in 2000 [82], apparently replacing genogroup B3 viruses. The current VP1-based genetic classification of EV71 (1970–2001), based on Brown et al. [74] and McMinn et al. [82], is shown in Fig. 2.

Figure 2

Phylogenetic tree showing genetic relationships among 25 EV71 field isolates based on alignment of the complete VP1 sequence (nucleotide positions 2442–3332) [74, 82]. Branch lengths are proportional to the number of nucleotide differences. Strain names indicate a unique number/country or U.S. state of isolation/year of isolation: AUS – Australia; CA – California, USA; CT – Connecticut, USA; IA – Indiana, USA; MAA – peninsular Malaysia; OR – Oregon, USA; SAR – Sarawak, Malaysia; SIN – Singapore; TW – Taiwan; TX – Texas, USA. The trees were constructed by neighbour-joining using the Kimura two-parameter distance method [83]. The bootstrap values in 1000 pseudo-replicates for major lineages within the tree are shown as percentages. The marker denotes a measurement of relative phylogenetic distance. The VP1 nucleotide sequence of CA16 [85] was used as an outgroup in the analysis.

The molecular epidemiology of the 1999 Western Australian epidemic is of particular interest. Virus strains belonging to single lineages within genogroups B3 and C2 were identified during this epidemic [82], suggesting that virus was introduced into Western Australia from endemic sources within Asia. In support of this hypothesis, several expatriate Australian children returning from Sarawak during the 1997 HFMD epidemic had evidence of recent EV71 infection and one child had EV71 cultured from her stool [84]. The genogroup B3 strains isolated in Western Australia share >99% nucleotide sequence identity with viruses isolated in Sarawak during 1997 and the Western Australian genogroup C2 strains share >98% nucleotide sequence identity with strains isolated in Taiwan during 1998 [80, 82].

A Japanese phylogenetic study of EV71 based on the VP4 gene sequence included one strain each from the Bulgarian (strain 258) and Hungarian epidemics [78]. The two isolates were closely related to one another and to viruses isolated in Japan in 1973 [19, 20] and Taiwan in 1980 [78]. Unfortunately, because these strains have not undergone phylogenetic analysis based on their VP1 sequence, their evolutionary relationship to recent Asia–Pacific isolates is unclear. However, the Bulgarian and Hungarian strains and the United States strain 7423-MS-87 were classified together within ‘VP4-based’ genogroup A [78]. Analysis of the VP1 sequence of 7423-MS-87 has placed it within ‘VP1-based’ genogroup B2 [74]. Thus, it appears that viruses belonging to ‘VP1-based’ genogroup B caused the Bulgarian and Hungarian epidemics, although their relationship to lineages B1 and B2 is unclear. Furthermore, the Shimizu et al. study [78] showed that the European strains are distantly related to recent ‘VP1-based’ genogroup B3 and B4 isolates from the Asia–Pacific region. VP1-based phylogenetic analysis of the existing European strains will be of enormous value in identifying their origin and relationship to other genogroup B viruses, and, given the apparent neurovirulence of these viruses, may help to elucidate the location of virulence determinants within the EV71 genome.

5 Molecular genetics of virulence

Investigations of the molecular genetics of poliovirus virulence over many years have shown that minor sequence variations in restricted areas of the genome (in particular the 5′ UTR and VP1 gene) are sufficient to account for large differences in neurovirulence between strains, suggesting that the neurovirulence phenotype of enteroviruses may be encoded by a small number of critical genetic determinants (reviewed in [3]). Although numerous clinical and epidemiological studies have established that EV71 possesses an intrinsic neurovirulence that distinguishes it from its closest relative CA16, the genetic determinants of EV71 neurovirulence have remained elusive. The nucleotide sequence and amino acid identities of the prototype EV71 and CA16 strains are 77% and 89%, respectively [2]. Given the large number of nucleotide and amino acid differences between the two viruses, it is not surprising that genomic comparison of these two viruses has been unable to provide useful information on EV71 neurovirulence determinants. However, careful observation has established that EV71 epidemics vary significantly in the prevalence of associated neurological disease, suggesting that strains of higher neurovirulence may occasionally arise and circulate for varying periods of time and in specific geographic locations. For example, a phylogenetic study has shown that EV71 strains isolated in Japan and Bulgaria during the mid-1970s clearly belong to the same genogroup and appear to be very closely related [78]. However, the 1973 Japanese epidemic was associated almost exclusively with HFMD [9, 27] while the Bulgarian epidemic was associated with numerous cases of poliomyelitis and without any apparent cases of HFMD [22]. Furthermore, recently circulating C1 genogroup viruses have been isolated only from HFMD cases and appear to have low neurovirulence compared to C2 genogroup viruses [82], which were responsible for numerous cases of AFP and brainstem encephalitis in Taiwan and Western Australia. As these two genogroups are quite closely related (the nucleotide sequence identity of the VP1 gene of genogroups C1 and C2 is ∼92%), comparison of the complete genomic sequences of representative isolates from these genogroups may prove to be informative about the location of neurovirulence determinants. However, the complete genomic sequence of a C1 genogroup virus has not been reported to date.

Advances in our understanding of the molecular genetics of EV71 virulence have also been severely hampered by the species specificity of EV71 and, thus, the lack of a convenient small animal model. Although EV71 is able to infect infant mice and causes myositis and flaccid paralysis typical of human Enterovirus A species viruses [16], newborn mice provide a very poor model of EV71-associated CNS disease. By contrast, intraspinal inoculation of Cynomolgus monkeys with EV71 causes poliomyelitis-like paralysis with histological features similar to those seen in human disease [67, 68]. Hagiwara et al. [68] also showed that low passage EV71 field isolates could cause poliomyelitis and brainstem encephalitis in Cynomolgus monkeys, and could thus serve as a useful animal model for defining EV71 neurovirulence determinants. Unfortunately, the crude RNA genetic analysis available at the time of the study (RNase T1 oligonucleotide mapping) was insufficiently sensitive to identify genetic markers for monkey neurovirulence [68]. However, this study provided the first demonstration of the intrinsic neurovirulence of EV71, in that strains obtained from cases of HFMD and neurological disease were capable of causing poliomyelitis after intraspinal inoculation of monkeys [68]. Unfortunately, there have been no further reports on the use of Cynomolgus monkeys to examine the pathogenesis of EV71-associated neurological disease.

In an attempt to identify EV71 neurovirulence determinants, several studies have compared genomic regions known to influence the neurovirulence of poliovirus, in particular the 5′ UTR [77, 80] and VP1 gene [74, 80], of EV71 strains isolated from well characterised mild and severe clinical cases. These studies have all failed to identify neurovirulence determinants. This is not surprising, because (1) enterovirus neurovirulence is a complex phenotypic characteristic that is likely to be determined by more than one region of the virus genome, and (2) host factors, such as age-related resistance to enterovirus infection, the presence of cross-protective immunity or specific major histocompatibility complex haplotypes, are likely to play a significant role in limiting the severity of most EV71 infections.

In order for EV71 strain comparison to lead to the identification of virulence determinants, rigorous strain selection based on the observed frequency of neurological disease occurring within particular epidemics or in patient cohorts within epidemics needs to be undertaken. For example, this approach has resulted in the identification of an EV71 lineage of high neurovirulence [82]. During the Perth EV71 epidemic in 1999, it was noted that genogroup C2 viruses were exclusively isolated from cases of severe neurological disease and that only one case of uncomplicated HFMD was caused by a C2 genogroup virus [82]. Phylogenetic analysis of these viruses showed that they belonged to a single genetic lineage (Fig. 3, lineage 1). Furthermore, comparison of the VP1 deduced amino acid sequences of the Perth genogroup C2 viruses with VP1 consensus amino acid sequences for EV71 [74], genogroups A, B and C [74] and CA16 [85] revealed the presence of an alanine to valine mutation at position 170 of VP1 in all five genogroup C2 viruses isolated from children with severe neurological disease (Fig. 4). By contrast, the earliest virus isolate in this lineage, obtained from a case of uncomplicated HFMD, had alanine (wild-type) at position 170 in VP1. All other EV71 isolates examined in this study, including the other two Perth genogroup C2 isolates (Fig. 3, lineage 2), had alanine at VP1-170 (Fig. 4). Amino acid position 170 is part of a highly conserved region of the enterovirus VP1 protein and has an alanine residue in CA16 and in all of the EV71 consensus sequences. The A→V substitution at position 170 in VP1 increases the hydrophobicity at this site and may result in a critical change in protein conformation (see below). These data suggest that the VP1-170 (A→V) substitution may have been associated with increased neurovirulence of EV71 during the Perth epidemic.

Figure 3

Phylogenetic tree showing genetic relationships among eight EV71 strains isolated in Perth, Western Australia, during the 1999 epidemic [48] and belonging to genogroup C2 [82]. Two EV71 strains isolated in Victoria, Australia (2642-AUS-95, 2644-AUS-95) in 1995 are included for comparison. The dendrogram was constructed from an alignment of the complete VP1 sequences by neighbour-joining using the Kimura two-parameter distance method [83]. The bootstrap values in 1000 pseudo-replicates for major lineages within the tree are shown as percentages. Branch lengths are proportional to the number of nucleotide differences, the marker denotes a measurement of relative phylogenetic distance.

Figure 4

Partial alignment of VP1 deduced amino acid sequences (residues 150–200) of the EV71 isolates from Western Australia belonging to genogroup C2. The deduced amino acid sequence in the same region of VP1 is also shown for CA16 [85], the EV71 consensus sequence [74], genogroup A (BrCr), consensus sequences for genogroups B and C [74] and the 1998 Taiwanese isolate NCKU9822 [46]. Amino acid residues that are identical to the EV71 consensus sequence are denoted with hyphens.

The enterovirus VP1 protein has a highly conserved tertiary structure and has been identified as a source of virulence determinants for several enteroviruses [86, 87]. A deep cleft on the virion surface has been identified at the junction of VP1, VP2 and VP3 at the five-fold axis of symmetry by X-ray crystallography [88] and is thought to function as the site of virion attachment to the cellular receptor. Amino acid position 170 is located in an α-helical structure within the E–F loop of VP1, at the interface between protomer subunits and on the rim of the canyon, a region thought to play a role in uncoating following receptor binding. Mutations within the E–F loop of the poliovirus VP1 protein have been shown to alter receptor specificity [89, 90] by a mechanism that is thought to involve conformational change within the cleft. Secondary structure analysis using the Garnier–Osguthorpe–Robson equation [91] predicts that the alanine to valine substitution is likely to alter protein conformation in the E–F loop from α-helix to β-sheet. Although the receptor for EV71 is unknown, it is possible that the VP-170 (A→V) substitution alters virus binding to the receptor by inducing conformational change within the cleft. Such a change in receptor binding may account for the observed change in neurovirulence of the Perth genogroup C2 viruses.

The enterovirus 5′ UTR contains a group of conserved secondary structural elements that collectively form the internal ribosome entry site (IRES). The IRES regulates enterovirus replication through control of cap-independent translation of the polyprotein [3]. The enterovirus IRES is composed of a variable number of stem-loop structures (depending on the virus serotype) but is remarkably uniform in function. IRES-like stem-loop structures have been identified within the 5′ UTR of EV71 ([77], R. Hurrelbrink and P. McMinn, unpublished data). Single nucleotide changes within the poliovirus IRES have been found to result in large alterations in neurovirulence [92]. Furthermore, substitution of the IRES from wild-type poliovirus type 1 (Mahoney) (PV1(M)) with the IRES of the (non-neurovirulent) human rhinovirus 2 results in a viable heterotypic chimera that has dramatically attenuated neurovirulence and reduced replication in neuroblastoma cells compared to PV1(M) [93]. Although it seems likely that the EV71 IRES is an important neurovirulence determinant, nucleotide variation linked to neurovirulence in the EV71 IRES has not been identified to date, and so the role of this structure in the control of EV71 neurovirulence remains unclear. Furthermore, EV71/PV1(M) chimeric viruses constructed by replacement of the PV1(M) IRES sequence with EV71 IRES cDNA derived from several recent Malaysian strains were found to have identical neurovirulence in CD155 Tg mice compared to PV1(M) (M. Gromeier, personal communication), suggesting that the EV71 IRES may have a lesser influence on neurovirulence than the poliovirus IRES.

The role of enterovirus non-structural proteins in neurovirulence remains poorly defined [3]. However, mutations within the RNA-dependent RNA polymerase 3D of the Sabin poliovirus type 1 vaccine strain have been implicated in the low neurovirulence phenotype of the virus [94]. Unfortunately, there are no data currently available on the role of EV71 non-structural proteins in the neurovirulence phenotype.

6 Prospects for control of EV71 infections

6.1 Public health surveillance

There is no effective antiviral treatment for severe EV71 infections and no vaccine is available. Thus, the only current means to prevent EV71 infection is through avoidance of contact between infected and susceptible individuals. Realistically, this can only be achieved through actions of limited efficacy, such as handwashing and reducing contact between infected and susceptible people during epidemics. Indeed, if these actions are to have any effect, it is imperative that adequate surveillance of EV71 activity is maintained in the community to provide early warning of impending epidemics. Surveillance systems include clinical surveillance for HFMD and laboratory surveillance to identify EV71 and other neurotropic enteroviruses. In response to the increased prevalence of EV71 in the Asia–Pacific region, several countries have implemented surveillance programmes to monitor EV71 activity. In some instances, it is claimed that these programmes have provided information that, when acted upon by public health intervention, resulted in early control of EV71 epidemics and reduced the total number of cases of neurological disease.

A clinical sentinel surveillance network for viral infections of public health significance was commenced in Taiwan in 1989 [39]. The network is based on weekly reporting of specific childhood infections seen by 850 sentinel physicians located throughout the country. Surveillance for HFMD and herpangina was commenced in 1997 [39] in response to the HFMD/encephalitis epidemic in Malaysia [15, 32, 33]. This network is supported by several public health virology laboratories that are capable of identifying enteroviruses to the serotype level. Thus, clinical reporting of HFMD was commenced 1 year before the EV71 epidemic in 1998 [39]. By March 1998, the surveillance system indicated the presence of large numbers of HFMD cases, two times above the background level observed during the previous year. This, together with laboratory reports of isolation of EV71 from cases of HFMD and encephalitis, triggered immediate public health warnings via print and electronic media (encouraging hand washing, confining infected children at home, avoiding contact between uninfected and infected children etc.). Further increases in HFMD prevalence over the next 2 weeks, coupled with reports of severe and fatal cases of EV71-associated encephalitis, triggered more draconian public health actions, in particular the closure of schools and preschools. It has been claimed, although not proven, that these interventions curtailed the epidemic and resulted in a smaller number of severe and fatal cases than would have occurred without public health intervention [39].

A clinical surveillance system for identification of HFMD within child-care centres was initiated in Singapore during 1987 [95] in response to an outbreak of EV71-associated HFMD and aseptic meningitis [31]. Since that time, outbreaks of HFMD in child-care centres have been administratively notifiable to the Ministry of the Environment. The data provided by this surveillance indicated considerable variation in HFMD prevalence from year to year until 1998, when a large epidemic occurred with case numbers 50% larger than previous reports. Concurrent laboratory data indicated that CA16 had been the predominant cause of HFMD in Singapore prior to 1998 but that EV71 was predominant during the 1998 epidemic [34, 95]. In response to this information, the Singapore government issued orders for the closure of child-care centres and schools. The surveillance system also predicted the HFMD epidemic in 2000 and provided the trigger for similar public health intervention by the Singapore government.

In Western Australia, a trial of community surveillance for the detection of neurotropic enteroviruses circulating among preschool-aged children attending child-care centres is currently being undertaken. This system is based on the collection of throat swab and stool samples from healthy children, culture of the samples and identification of the viral genotype of culture isolates by PCR using degenerate primers designed to amplify the VP1 gene of all 66 human enterovirus serotypes [76] and also using primers that specifically identify EV71 [96]. Healthy preschool-aged children were chosen for this project because they are known to be the major reservoir for enteroviruses in the community [97] and the majority of enterovirus infections within this age group are sub-clinical [51]. This surveillance is supplemented by isolation and rapid identification of enteroviruses in children admitted with febrile illnesses or acute neurological disease to the local children's hospital. Preliminary data indicate that 10% of the children surveyed are asymptomatically infected with an enterovirus (S. Szefczyk and P.C. McMinn, unpublished data), a finding similar to that of earlier enterovirus surveillance studies in the USA [98, 99]. It is anticipated that this surveillance will provide a sensitive early warning system for identification of EV71 activity in the community and that this information will be of use for initiation of public health interventions. These studies also indicate that the high standards of hygiene practiced in developed countries are not sufficient to prevent the transmission of enteroviruses between children and that human populations within industrialised countries remain vulnerable to epidemics of virulent neurotropic enteroviruses such as EV71.

6.2 Antiviral agents

Although a number of promising antiviral agents with activity against enteroviruses are currently being developed, none is close to release, and their effect on EV71 is unknown. The ‘WIN’ group of compounds is the most promising of these agents, several of which have undergone clinical trial [100]. WIN compounds bind within the hydrophobic pocket that lies beneath the floor of the picornavirus receptor binding canyon. Their antiviral activity is thought to be mediated by stabilisation of the viral capsid and the prevention of virus uncoating after receptor binding. The WIN compound pleconaril (3-{3,5-methyl-4-[[3-methyl-(5-isoxazolyl)propyl]phenyl]-5-(trifluoromethyl)}-1,2,4-oxadiazole) [101] has been found to provide significant therapeutic benefit in aseptic meningitis, AFP and encephalitis due to many enterovirus serotypes [100, 102, 103] and is currently undergoing phase III clinical trial in the USA. Pleconaril has an excellent side effect profile that has been attributed to its metabolic stability and the virus-specific nature of its mechanism of action [104]. Unfortunately, pleconaril has been found to have limited activity against EV71 at concentrations tested in vitro [D.C. Pevear, personal communication] and anecdotal information on its use in EV71-associated brainstem encephalitis has indicated poor responses to therapy.

As indicated previously, there is some evidence that inflammation may contribute to the pathogenesis of EV71 encephalomyelitis [45, 52, 70], suggesting that anti-inflammatory agents or intravenous immunoglobulin (IVIG) may be of use in the management of this disease. However, no reports demonstrating the efficacy of IVIG, steroid or non-steroid anti-inflammatory agents in EV71 neurological disease have been published to date. Furthermore, although IVIG was used extensively in the management of neurological disease cases during the Taiwanese [105] and Western Australian [48] epidemics, it was not associated with any objective evidence of improvement in clinical outcome.

A major obstacle to successful use of antiviral agents in EV71-associated encephalitis is that a large proportion of children have already suffered irreversible brain damage by the time that they present to hospital (especially grade III brainstem encephalitis and poliomyelitis-like paralysis). Thus, even if an antiviral agent has high-level activity against EV71, it is likely that little clinical improvement can be expected from therapy. Consequently, our major research effort needs to focus on prevention of EV71 infection through the development of effective vaccines.

6.3 Vaccine development

The success of both the formalin-inactivated and live-attenuated vaccines in controlling epidemic poliomyelitis and in the eradication of poliovirus highlights the potential for control of EV71 epidemics by mass vaccination. As noted previously, a formalin-inactivated EV71 vaccine was developed in response to the Bulgarian epidemic in 1975 [22, 24] but was not used to control the epidemic and has not been used since. Furthermore, no data on the efficacy of the Bulgarian vaccine are available. A group from Taiwan has recently reported on the development of two candidate EV71 vaccines: (1) a formalin-inactivated whole virus vaccine, and (2) a DNA vaccine expressing VP1 [106]. The efficacy of both vaccine constructs is currently being tested in animal models [106, 107]. In addition, the expression of recombinant EV71 VP1 capsid protein has been reported [108]. This approach may also prove valuable in the development of a subunit EV71 vaccine.

One of the major barriers to development of an EV71 vaccine is the lack of a suitable animal model for the testing of vaccine immunogenicity and efficacy. Laboratory mice are only susceptible to EV71 infection in the first 4 days of life and become completely resistant by 6 days of age [107, 109]. Although Cynomolgus monkeys are susceptible to infection with low passage clinical isolates of EV71 and develop encephalomyelitis and poliomyelitis-like paralysis after subcutaneous [67] or intraspinal [68] inoculation, the high cost of purchase and maintenance of the animals is a major barrier to their use in large-scale pathogenesis studies. A promising approach to understanding the pathogenesis of enterovirus encephalitis and for testing of candidate vaccines is through the development of transgenic mice expressing the appropriate human receptor molecule(s) for the virus under study. This approach has been extremely successful for poliovirus [110, 111] and has recently been reported for echovirus 1 [112]. The receptor for EV71 is unknown. Thus, identification of the EV71 receptor is a major priority in vaccine research, as it will allow the development of a transgenic mouse model for studies of EV71 pathogenesis and vaccine efficacy. Furthermore, construction of an infectious cDNA clone of EV71 will be essential for studies of the molecular genetics of virulence, for identification of virulence determinants and for construction of genetically defined, live-attenuated candidate vaccine strains. It is likely that these powerful molecular biological approaches will lead to the development of an effective EV71 vaccine, a result keenly awaited by the medical and public health communities in the Asia–Pacific region.

7 Conclusions

There has been a significant increase in EV71 epidemic activity in the Asia–Pacific region during the past 4 years. In addition, a new clinical manifestation of EV71 infection, a rapidly fatal syndrome of neurogenic pulmonary oedema associated with brainstem encephalitis, has been identified. Molecular genetic studies of EV71 isolates indicate that four distinct viral genotypes circulated in Sarawak, peninsular Malaysia, Japan, Taiwan and Australia between 1997 and 2001. Unfortunately, these studies have not provided conclusive evidence for an association between particular viral genotypes and the development of brainstem encephalitis with pulmonary oedema.

The apparent increase in the prevalence of acute neurological disease and the emergence of neurogenic pulmonary oedema as a distinct clinical entity in the Asia–Pacific region is likely to be of multi-factorial origin, the causes of which need to be urgently addressed by epidemiological, clinical and molecular biological investigations. For example, demographic factors in the human population of the region, such as population growth, urbanisation and the increasing reliance on communal child-care by working parents with young children, may have contributed to the increased endemic circulation of EV71. Furthermore, the increased prevalence of acute neurological disease may either be due to greater virulence of recent EV71 strains and thus a reduced disease-to-infection ratio in the population, or may simply result from the higher magnitude of EV71 infections occurring in the region.

Changes in the mode of virus transmission may also lead to increased virus circulation. It has been suggested that the recent emergence of large epidemics and high-level endemic circulation of EV71 in the Asia–Pacific region may be due to a change in virus transmission from primarily faecal–oral spread to one of respiratory/aerosol spread [42]. However, preliminary risk analysis studies have shown that EV71 was transmitted primarily by faecal–oral spread during the Taiwanese epidemic [41, 64]. It is clear that the mode of transmission of recent EV71 strains needs to be determined rigorously and as a matter of public health urgency, as this knowledge will have major implications for the control of future EV71 epidemics.

At present, the reasons for the emergence of EV71 as a cause of large epidemics of acute neurological disease in young children in the Asia–Pacific region remain elusive. In many ways the emergence of EV71 as a cause of large-scale epidemics of AFP and other acute neurological diseases is reminiscent of the emergence of epidemic poliomyelitis in Europe and North America during the late 19th century. The fact that epidemic EV71 did not emerge concurrently with epidemic poliomyelitis suggests that the reasons for the emergence of these two highly neurotropic enteroviruses may differ. Despite this, it has been suggested that EV71 may become the major infectious cause of AFP in the world following the eradication of poliovirus [12, 56, 58]. Thus, it is imperative that the medical and scientific communities prepare for such an eventuality in order to avoid the large-scale loss of life and human potential that resulted from the poliomyelitis epidemics of the 20th century and that has already occurred as a result of EV71 epidemics. Exciting developments in the techniques of molecular biology over the past 20 years have provided us with powerful tools to combat epidemics of EV71. It behoves us to use these tools to prevent EV71 infection through the development of regional surveillance to predict impending epidemics and to develop vaccines to protect our children from the devastating neurological consequences of this disease.


I would like to express sincere thanks to Professor C.J. Burrell for critical appraisal of the manuscript. This study was supported by grants from the Princess Margaret Hospital for Children Research Fund, the Royal College of Pathologists of Australasia, the National Health and Medical Research Council of Australia and the Health Department of Western Australia.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
  52. [52].
  53. [53].
  54. [54].
  55. [55].
  56. [56].
  57. [57].
  58. [58].
  59. [59].
  60. [60].
  61. [61].
  62. [62].
  63. [63].
  64. [64].
  65. [65].
  66. [66].
  67. [67].
  68. [68].
  69. [69].
  70. [70].
  71. [71].
  72. [72].
  73. [73].
  74. [74].
  75. [75].
  76. [76].
  77. [77].
  78. [78].
  79. [79].
  80. [80].
  81. [81].
  82. [82].
  83. [83].
  84. [84].
  85. [85].
  86. [86].
  87. [87].
  88. [88].
  89. [89].
  90. [90].
  91. [91].
  92. [92].
  93. [93].
  94. [94].
  95. [95].
  96. [96].
  97. [97].
  98. [98].
  99. [99].
  100. [100].
  101. [101].
  102. [102].
  103. [103].
  104. [104].
  105. [105].
  106. [106].
  107. [107].
  108. [108].
  109. [109].
  110. [110].
  111. [111].
  112. [112].
View Abstract