Address correspondence to Usha Kant Misra, Professor & Head, Department of Neurology, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Raebareily Road, Lucknow-226014. E-mail: email@example.com, firstname.lastname@example.org
Viral encephalitis presents with seizures not only in the acute stage but also increases the risk of late unprovoked seizures and epilepsy. Acute symptomatic and late unprovoked seizures in different viral encephalitides are reviewed here. Among the sporadic viral encephalitides, Herpes simplex encephalitis (HSE) is perhaps most frequently associated with epilepsy, which may often be severe. Seizures may be the presenting feature in 50% patients with HSE because of involvement of the highly epileptogenic frontotemporal cortex. The occurrence of seizures in HSE is associated with poor prognosis. In addition, chronic and relapsing forms of HSE have been described and these may be associated with antiepileptic drug-resistant seizures. Among the epidemic (usually due to flaviviruses) viral encephalitides, Japanese encephalitis (JE) is most common and is associated with acute symptomatic seizures, especially in children. The reported frequency of acute symptomatic seizures in JE is 7–46%. Encephalitis due to other flaviviruses such as equine, St. Louis, and West Nile viruses may also manifest with acute symptomatic seizures. In Nipah virus encephalitis, seizures are more common in relapsed and late-onset encephalitis in comparison to acute encephalitis (4% vs. 1.8%). Other viruses like measles, varicella, mumps, influenza, and entero-viruses may cause seizures depending on the area of brain involved. There is no comprehensive data regarding late unprovoked seizures in different viral encephalitides. Prospective studies are required to document the risk of late unprovoked seizures and epilepsy following viral encephalitis due to different viruses as well as to determine the clinical characteristics, course, and outcome of post-encephalitic epilepsy.
Viral encephalitis refers to an acute inflammatory process of the brain parenchyma due to direct viral infection. The clinical presentation of viral encephalitis is nonspecific and includes fever, varying degrees of alteration in sensorium with or without focal neurological deficits and/or seizures. These symptoms may as well be due to a variety of other infective and noninfective causes. In these circumstances, malaria, bacterial and fungal meningitis, and noninfectious causes of encephalopathy must be carefully excluded. Specific viral diagnosis can be achieved either by demonstration of viral nucleic acid or antibody in cerebrospinal fluid (CSF) or by isolation of the virus from CSF or brain tissue. Even with best efforts, 30–60% of patients with clinically suspected viral encephalitis remain undiagnosed. Isolation of the virus is seldom accomplished due to the short period of viremia, difficulty in obtaining brain tissue through biopsy, and lack of specialized facilities for viral cultures in many places. Serological (immunological) tests are simpler to perform and widely available for virological diagnosis. During active infection, the demonstration of a four-fold rise in antibody titers against the virus in serum or CSF reliably confirms a diagnosis of acute viral encephalitis. However, for past and remote viral infections, serological tests do not differentiate between clinical episodes of encephalitis and mere exposure to the virus. Polymerase chain reaction (PCR) technology has significantly improved virological diagnosis and finds widespread application in this area of clinical microbiology.
Epileptic seizures may not only occur during the acute phase of viral encephalitis but also later, following resolution of the acute phase of illness. The two types of seizures, that is, early seizures that are coincident with the acute episode and late seizures that are sequelae of viral encephalitis have different underlying mechanisms and prognostic implications. Seizures in acute stage have been described in sufficient detail with regard to their incidence, frequency, and prognosis. However, late unprovoked seizures remain poorly characterized in terms of their frequency, course, outcome, and prognosis.
Epidemiology of Viral Encephalitis
Over 100 viruses (some of which are listed in Table 1) can produce encephalitis. A number of these have specified geographical predilections. The largest burden of viral encephalitides occurs in less developed countries with poor health infrastructure. Population-based studies of the incidence of viral encephalitis are therefore scarce (Beghi et al., 1984). The nonspecific symptoms and signs of infection, variations in the biological and epidemiological behaviors of their causative agents, and limitations of viral diagnostic tests preclude elucidation of the specific aetiology of viral encephalitis. Hence, the aetiological agent of viral encephalitis remains unidentified in nearly 60% cases (Beghi et al., 1984; Khetsuriani et al., 2002; Davison et al., 2003).
Herpes simplex virus type 1 and 2, varicella zoster, Epstein Barr, and cytomegalovirus viruses and human herpes viruses type 6 and 7
Coxsackie, echo, enterovirus 70 and 71, parecho, and polio viruses
Measles and mumps viruses
Influenza, adenovirus, parvovirus, lymphocytic choriomeningitis, and rubella viruses
Geographically restricted causes of encephalitis
West Nile, La Crosse, St. Louis, Rocio, Powassan encephalitis, Venezuelan encephalitis, eastern and western equine encephalitis, Colorado tick fever, dengue, and rabies viruses
Tick-borne, West Nile, Tosana, rabies, dengue, and louping ill viruses
West Nile, rabies, Rift valley fever, Crimean-Congo hemorrhagic fever, dengue, and Chickungunya viruses
Japanese-B encephalitis, West Nile, Murray valley encephalitis, dengue, Nipah, Chikungunya, and rabies viruses
Murray valley encephalitis, Japanese encephalitis, dengue, and Kunjin viruses
The estimated incidence of clinically-diagnosed viral encephalitis is 3.5–7.4/100,000/year (Johnson, 1996). Most estimates are based on passive surveillance, which relies upon report of cases by physicians to a nodal registry and hence underestimates the actual incidence of disease. Surveys employing active case finding have reported higher incidence in comparison to passive surveillance (Beghi et al., 1984). The estimated incidence of viral encephalitis in Olmsted County, Minnesota in the United States (1950–1981) was 7.4/100,000/year (Beghi et al., 1984). In this survey and in others from Scandinavia, higher incidences have been observed in children (Rantala & Uhari, 1989; Koskiniemi et al., 1991; Ishikawa et al., 1993; Koskiniemi et al., 1997). The incidence in children below 1 year is about 22.5/100,000 population per year and in children below 10 years of age of age is 15.2/100,000/year. A second peak in old age has been suggested, particularly for Herpes simplex encephalitis (HSE), but has not been borne out by most studies (Khetsuriani et al., 2002). In the United Kingdom, based on the data from hospital admission surveillance, about 700 cases of viral encephalitis are recorded every year (Davison et al., 2003), while in the United States, an estimated 19,000 hospitalizations and 1,400 deaths take place each year for a diagnosis of viral encephalitis (Khetsuriani et al., 2002).
Geographic Distribution of Viral Encephalitis
Viruses that cause encephalitis can be divided in two groups based on their epidemiological behavior: those that cause sporadic encephalitis and those that are responsible for epidemics of encephalitis. The majority of epidemics of viral encephalitis are due to arboviruses, that is, viruses that are transmitted to humans by arthropods, including mosquitoes and ticks. The various arboviruses have characteristic geographical distributions; for instance, in the United States, the California serogroup of viruses (including the La Crosse virus) are the most common cause of viral encephalitis (Beghi et al., 1984). Other viral agents responsible for encephalitis in the United States include St. Louis virus and eastern equine encephalitis and western equine encephalitis viruses. Epidemics of eastern equine encephalitis and western equine encephalitis were recorded in 1970s; apart from these, there have been only handful of documented cases every year (http://www.cdc.gov/ncidod/dvbid/arbor/arbocase.htm). Small focal outbreaks have been documented for the St. Louis encephalitis virus. Encephalitis caused by a tick-borne virus known as Powassan virus was described in 1999–2000 in northeastern United States (CDC, 2001). In South America, Venzualian equine encephalitis and eastern and western equine encephalitis viruses are responsible for encephalitis (Weaver et al., 1996; Figueiredo, 2007). In the United Kingdom, arboviral encephalitis has not been reported (Davison et al., 2003). However, clusters of infection of unknown etiology have been described from here and it would be interesting to speculate an arboviral aetiology for these clusters. In most of the central and eastern Europe, arboviral encephalitis is caused by a tick-borne virus. Japanese encephalitis (JE) is the leading cause of encephalitis in Southeast Asia, where 30,000–50,000 cases are recorded annually (Tsai, 1997). The World Health Organisation estimated nearly 14,000 deaths due to JE in the year 2002. Of these, 8,500 occurred in Southeast Asia, 3,000 in the western Pacific region and about 2,000 in the eastern Mediterranean region. Arboviruses of medical importance in Australia include the Murray valley encephalitis and Kunjin viruses, both of which have been documented to cause encephalitis mostly in western Australia and the Northern Territory (Broom et al., 2003).
Herpes simplex virus type 1 (HSV-1) is the most common cause of sporadic encephalitis (Davison et al., 2003; Mailles et al., 2007). In some recent reports from Scandinavia and central Europe, however, varicella zoster has been identified as the most common viral agent responsible for encephalitis (Cizman & Jazbec, 1993; Studahl et al., 1998; Koskiniemi et al., 2001). In many of these reports, PCR was used extensively in virological diagnosis; hence, the observation may reflect improved diagnostic methods. Besides HSV and varicella, other viral pathogens that can cause encephalitis include influenza virus type A, mumps, measles, and enteroviruses.
Outbreaks of encephalitis due to novel viral pathogens continue to be described. For instance, an outbreak of West Nile virus infection was documented in New York in 1999, Nipah virus encephalitis due to a paramyxovirus was reported from Malaysia in 1999 and eastern India in 2001, and Chandipura virus encephalitis due to a novel virus belonging to the rhabdoviridae family was reported from South India in 2002 (Mostashari et al., 2001; Chong & Tan, 2003; Tandale et al., 2008).
Pathogenesis of Viral Encephalitis
Viruses producing encephalitis have certain distinctive features that render them pathogenic to the brain parenchyma. These are essentially neurotropic, that is, they have the ability to invade, infect, and subsequently replicate within the human nervous system. The basis of neurotropism is not well understood. Certain viruses, for example, HSV-1, varicella zoster virus, and human herpes virus (HHV)-6 and -7 can remain dormant in the neural tissue for long periods of time following initial infection. This ability is considered responsible for producing latent and persistent infections, relapses as well as reactivation several years following initial infection (Baringer & Pisani, 1994).
In the case of HSV-1, initial infection is followed by axoplasmic transport of the virus to the trigeminal sensory ganglion where it establishes latency. Latent HSV-1 virus is detectable in the trigeminal ganglia in nearly all seropositive (anti-HSV-1 antibody) individuals (Baringer & Pisani, 1994). Reactivation results in the retrograde transport of virus usually resulting in herpes labialis. However, the pathway by which HSV-1 reaches the brain parenchyma in humans in order to produce encephalitis remains unknown. In primary infection, the virus possibly gains access to the olfactory bulbs through the nose and subsequently spreads via the olfactory pathway to orbitofrontal and mesial temporal lobes (Johnson, 1998). The HSV-I has affinity for basi-frontal and mesial temporal (limbic) cortex and spares most of the remaining cerebral cortex, the deep gray matter nuclei, and the white matter (Fig. 1). Perhaps, the proximity of dural nerves to basi-frontal and temporal lobes renders them susceptible. Moreover, the virus lies dormant in anterior and middle cranial fossae and it is easy to invade the frontotemporal cortices from here. The virus may travel by cell to cell contact and across the meninges into adjacent cortices (Damasio & van Hoesen, 1985).
Seizures in Viral Encephalitis
Viral encephalitis frequently manifests with seizures in its acute phase. The exact frequency of early seizures in viral encephalitis is underestimated perhaps because the manifestations of seizures may be subtle. Nonconvulsive seizures remain underdiagnosed as these are detectable only by continuous electroencephalographic monitoring. Seizures may occur in 40–60% of cases during the acute stage of HSE. The high frequency of seizures reflects the propensity of HSV-1 to involve the highly epileptogenic mesial temporal lobe structures including the hippocampus. In the experimental setting, application of HSV-1 to hippocampal cultures ex vivo produces both acute electrographic and clinical seizures as well as induces long-term changes in hippocampal excitability as a result of neuronal loss in the CA3 sector of the hippocampus (Wu et al., 2003; Chen et al., 2004).
Seizures have also been described by several investigators in JE (Gourie Devi et al., 1995; Misra & Kalita, 1998, 2001; Solomon et al., 2002; Kalita et al., 2003). The reported frequency of seizures in JE is highly variable (7–67%). A Vietnamese study revealed that the occurrence of seizures correlated with elevated CSF opening pressures, signs of brain herniation and mortality (Solomon et al., 2002). Predictors of the occurrence of seizures included young age, level of consciousness, and presence of cortical involvement on imaging (Fig. 2A). Sixty-one percent of children had seizures compared to 37% of adults and 54% of those with cortical involvement demonstrated on imaging had seizures in comparison to 15% without cortical signal abnormalities. Seizures were not related to mortality but were associated with poor outcome (Misra & Kalita, unpublished observations). In Nipah encephalitis, acute symptomatic seizures have been reported in 24% of acute encephalitis, but 50% of relapsed or late-onset encephalitis (Tan et al., 2002).
Late unprovoked seizures
Although early seizures are frequent in all types of acute viral encephalitides irrespective of aetiology, late unprovoked seizures may or may not be common in various encephalitides. The risk of unprovoked seizures or epilepsy after an episode of viral encephalitis was reported in a retrospective community-based study in Rochester, MN, USA (Annegers et al., 1988). An episode of viral encephalitis complicated by early seizures increased the risk of developing a subsequent unprovoked seizure by 22 times and an episode that was not accompanied by early seizures increased this risk by 10 times. Overall, there was 16 times increased risk of developing an unprovoked seizure following viral encephalitis. Most seizures occurred within the first 5 years following the encephalitic episode though the risk of unprovoked seizures remained elevated till 20 years.
The incidence of late unprovoked seizures in different viral encephalitides has not been systematically evaluated. Hospital-based studies have revealed that late unprovoked seizures occur in 40–65% of patients after an episode of HSE (Chen et al., 2006; Elbers et al., 2007). The high incidence of late unprovoked seizures following HSE may be owing to the necrotizing nature of HSV-1 infection and involvement of the highly epileptogenic mesial temporal and basi-frontal cortices. The incidence of epilepsy following La Crosse encephalitis is 10–12% (Chun, 1983). Follow-up studies of encephalitides due to other viruses such as West Nile encephalitis, eastern equine encephalitis and JE have emphasized the occurrence of neurocognitive sequelae and movement disorders rather than epilepsy. The predominant involvement of subcortical structures (thalamus, substantia nigra, and the basal ganglia) (Fig. 2A) and sparing of the cerebral cortices may be the reason for lower incidence of late unprovoked seizures following these encephalitides (Deresiewicz et al., 1997; Kalita & Misra, 2000). Similarly, acute Nipah virus encephalitis produces subcortical and deep white matter lesions in magnetic resonance imaging (MRI) (Fig. 3) (Ahmad Sarji et al., 2000). Hence, the incidence of unprovoked seizures is low. The incidence of late unprovoked seizures was found to be 2.2% in a follow-up study of 137 patients over 8 years and was higher for relapsed (4.0%) than acute (1.8%) encephalitis (Tan, unpublished observations).
The anatomic distribution of lesions in acute viral encephalitis is of relevance to determining the basis of postencephalitic epilepsy. Because, these lesions represent putative epileptogenic foci, they are critical determinants of the semiological characteristics, course, and outcome of postencephalitic epilepsy. In general, patients with postencephalitic epilepsy are likely to present with multifocal epilepsy. A comparative study of patients with intractable epilepsy following meningitis and encephalitis, who were subjected to presurgical assessments, found that those with meningitis were more likely to have a mesial temporal onset of seizure, while those with encephalitis more often had neocortical seizures (Marks et al., 1992).
Association between Herpes Virus Infections and Epilepsy
Although HSE is typically an acute monophasic illness, both chronic persistent as well as relapsing forms have been described (Jay et al., 1998; Ito et al., 2000; De Tiége et al., 2003; Yamada et al., 2003). Some cases of HSE, in particular those in older individuals are thought to represent reactivation of the latent virus in the CNS (Whitley et al., 1982). The incidence of relapse following an episode of HSE is 5–26% (Ito et al., 2000; De Tiége et al., 2003). Most relapses occur soon after completion of acyclovir therapy. However, there are reports of progressive deterioration in seizure control, focal neurological deficits, and cognitive impairment for several years after an acute episode of HSE (Yamada et al., 2003). Pathological examination of brain tissue obtained during epilepsy surgery or post mortem has revealed chronic inflammation within the mesial temporal lobe structures in addition to the identification of HSV genome and antigen in some patients (Jay et al., 1998; Yamada et al., 2003). Conversely, in some patients, HSV-1 DNA has been detected without evidence of inflammation (Jay et al., 1995; Sanders et al., 1997) suggesting a chance finding. The reports of chronic and relapsing encephalitis need to be interpreted in the context of the tendency of the HSV to remain dormant in neural tissue for long periods of time. The HSV-1 genome can be detected at autopsy by PCR in brains of individuals who have not had any neurological disorder in their life (Baringer & Pisani, 1994). Hence, it remains debatable whether the detection of viral genome in cases ostensibly described as chronic or recurrent HSE represent pathogenicity of the detected virus or merely the coincidental detection of dormant viral matter.
Both acute symptomatic and late unprovoked seizures (and epilepsy) may occur in relation to viral encephalitis. It appears that epilepsy following HSE is common and is frequently difficult to control. There is limited data regarding the incidence, course, and outcome of epilepsy following most other viral encephalitides. It is possible that the propensity to develop late seizures following viral encephalitis depends on the distribution of pathologic lesions in different regions of the brain. The seizure burden and characteristics and outcome of epilepsy associated with viral encephalitis should be determined independently for each of the several known viral agents.
We acknowledge Rakesh Kumar Nigam for secretarial help.
Conflict of interest: We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. The authors have declared no conflicts of interest.