Viral Etiologies of Encephalitis in Thai Children

A prospective study of childhood encephalitis was performed in Bangkok from 1996 through 1998. The viral agents identifiable in 26 (65%) of 40 children were dengue virus (8), Japanese encephalitis (6), herpes simplex virus (4), human herpes virus type 6 (3), mumps (2), enterovirus (1), varicella-zoster virus (VZV) (1) and rabies (1). Various viral agents cause encephalitis; the major ones are arboviruses and herpes simplex virus (HSV). Immunization against Japanese encephalitis virus (JEV), the most prevalent arbovirus in Asia, has been generally recommended in Thailand. This may have reduced the relative rate of JEV encephalitis in Thai children. In endemic areas dengue virus (DV) infection may involve the central nervous system 1 and cause illness similar to other viral encephalitis. Many other agents such as rabies, enteroviruses (EV), mumps virus (MV), human herpes virus type 6 (HHV-6), Epstein-Barr virus, adenovirus, human immunodeficiency virus and Mycoplasma are significant in some areas. The objective of this study was to determine the viral causes of encephalitis among Thai children at the time of widespread vaccination against JEV. We also examined the clinical findings and outcome associated with different viral causes. Materials and methods. A prospective cohort study was performed at the Department of Pediatrics, Siriraj Hospital, a tertiary care medical center with a 300-bed facility in Bangkok. From July, 1996, to December 1998, the children hospitalized for illnesses suggestive of encephalitis or meningoencephalitis, i.e. fever with alteration of consciousness, or with neurologic deficit secondary to central nervous system involvement, lasting >24 h and without other identifiable causes were enrolled into the study. In addition to the laboratory workups determined by the attending physician, blood and cerebrospinal fluid (CSF) samples were collected for further investigations of viral etiologic agents including HSV, HHV-6, EV, JEV, DV and MV. Clinical data, treatment and outcome were recorded. Decisions regarding treatment were not altered by the study. All the patients were followed up for at least 6 months after leaving the hospital. Paired blood specimens were tested simultaneously for antibody against HSV, HHV-6, JEV, DV and MV. The CSF specimens were processed for PCR to identify the presence of HSV, HHV-6 and EV genomes. In the patients who had symptoms or serology suggestive of dengue infection, dengue PCR and dengue viral isolation were performed in acute blood and CSF specimens. The nested PCR for HSV used oligonucleotide primers derived from the thymidine kinase (TK) gene. The length of the amplified products designed by the outer and inner primer pairs were 540 and 250 bp, respectively, for HSV-1 and HSV-2. The oligonucleotide primers derived from the immediate early (IE) gene 2 were used to amplify HHV-6 DNA in CSF samples. The primary PCR products were 321- and 554-bp fragments and the nested PCR products were 193- and 421-bp fragments for variant A and B, respectively. Detection of EV RNA was performed with the Amplicor test kit (Roche Diagnostic System). Detection of DV RNA in acute blood samples and CSF was performed by reverse transcriptase-PCR with serotype-specific primers in a seminested PCR according to protocols described previously. 3 Dengue viral isolation was attempted with laboratory-reared mosquitoes (Toxorhynchites splendens) by an intrathoracic inoculation technique. 4 Identification of dengue types was further performed by indirect fluorescence antibody staining using serotype specific monoclonal antibody. Specific antibodies against HSV (IgG and IgM) and EV (IgG, IgM and IgA) in paired sera were determined by enzyme-linked immunosorbent assay (ELISA) with test kits from Genzyme Virotech GmbH (Russelsheim, Germany). Anti-HHV-6 (IgG and IgM) was detected by an ELISA test kit from Panbio (East Brisbane, Australia). Anti-MV antibody testing was performed by complement fixation with specific antigen from Dade-Behring (Marburg, Germany). Anti-JEV and anti-DV IgG and IgM antibodies in CSF and serum samples were assayed by an ELISA capture test as described earlier 5 and by a hemagglutination-inhibition test modified from that of Clarke and Casals. 6 Diagnosis of JEV or DV infection was confirmed by the results of both serologic methods. The patients who had final diagnoses other than encephalitis were excluded. The diagnosis of encephalitis or encephalopathy caused by DV was referred to as dengue infection with central nervous system involvement (DCI). The clinical findings and outcome of encephalitis from the major pathogens were compared by chi square, Fisher’s exact and Student’s t tests. Results. During the period of 30 months, 61 children were enrolled into the study. At the conclusion of the illnesses, 21 patients had final diagnoses other than encephalitis, i.e. epilepsy (4), febrile convulsion (5), bacterial meningitis (3), tuberculosis (3), brain tumor (1), congenital cytomegalovirus infection (1), acute disseminated encephalomyelitis (1), progressive multifocal leukoencephalopathy (1), subarachnoid hemorrhage (1) and lead poisoning (1). These patients were excluded from the analysis. Of the 40 patients with encephalitis, the mean age was 6.6 ± 3.84 years (range, 1 month to 13 years) with boys predominating (M:F = 25:15). The viral etiologies were identified in 26 (65%) patients (Table 1). The three most common agents, DV, JEV and HSV, accounted for 69% of the identifiable agents and the etiologies of 45% of the total encephalitis cases. There was no seasonal pattern of any etiology except for DV that occurred only during the rainy season from August to November.Table 1: Viral etiologies of and diagnostic methods used in 40 Thai children with encephalitisThe neurologic presentations were mostly alteration of consciousness (82%), headache (47%), meningism (45%) and seizure (37%). There were some clinical clues associated with viral etiologies identified. Seven of 8 patients with DCI developed typical manifestations of dengue hemorrhagic fever (DHF), i.e. fluid leakage from the intravascular compartment and thrombocytopenia within 48 h of hospitalization. Two of these developed convalescent rash. Both patients with mumps encephalitis had parotitis. The patient with VZV encephalitis had typical varicella rash. The patient with rabies had hydrophobia with laryngospasm and a history of dog bite. The patients with HHV-6 infections did not have rash. In most cases, however, the clinical manifestations caused by each agent were indistinguishable. None of the six patients with JEV encephalitis had received the vaccination against JEV. One of the two patients with mumps encephalitis had been vaccinated with mumps-containing vaccine. The patient with VZV encephalitis never received VZV vaccination. CSF findings were abnormal in 88% (28 of 32) of encephalitis cases caused by etiologies other than DV, in which almost all CSF were normal. Polymorphonuclear cells were predominant in 15% (4 of 27) of children with pleocytosis. Of the eight patients with serologically confirmed dengue infection, six had positive viral isolations or PCR RNA in the blood. One of these who presented with left facial palsy (lower motor neurone type) also had viral RNA detected in the CSF. DV serotypes identified were types 1, 2 and 3, each in two patients. The neurologic symptoms gradually declined during a 2-week period in all. Except for one patient who had a protein concentration of 69 mg/dl without pleocytosis, all of the patients had normal CSF findings. According to serologic findings DCI occurred as a primary infection in one, a secondary infection in five and of uncertain status in two patients. The severity of DHF by WHO classification was Grade I in one, Grade II in 6 and Grade III in 1 patient. Acyclovir was given to 25 (62.5%) patients. Seven (17%) patients died from the illnesses. The overall rate of complete recovery was 70% (28 of 40), lowest in the group of unknown cause (7 of 14, 50%). Five (12.5%) patients recovered with significant neurologic sequelae that persisted for >6 months. All of the patients with encephalitis caused by EV, MV, VZV and DV recovered completely. The mean duration of hospitalization from encephalitis was 26.25 ± 35.7 days (range, 4 to 155 days). Discussion. Viral etiologic agents in this cohort were identified in 65%, similar to those of other studies. 7 Arboviral encephalitis occurs at different rates in different parts of the world. Seasonal variation in arbovirus infection is dependent on the survival and proliferation of the vectors. Aedes aegypti, the vector of DV, proliferates well during rainy months when the eggs are laid on stagnant pools in the environment. In contrast the presence of reservoirs (rodents, pig) and vectors (Culex mosquitoes) for JEV are stable year-round in Thailand; therefore there is little seasonal variation for JEV. In this study encephalitis caused by JEV was significant, accounting for 15% of the cases despite the availability of effective vaccine. 8 Vaccination against JEV has been in the routine immunization schedule in endemic provinces (34 of 76) in Thailand since 1985. Bangkok has not been considered an endemic province for JEV; therefore vaccination against JEV has been optional. Taking into account data from this cohort and the safety, efficacy and affordability of the vaccine, 8 routine vaccination against JEV in Bangkok should be considered. The clinical presentations of encephalitis from various causes were mostly indistinguishable. The exception was in the cases caused by MV, VZV, rabies and DV. Dengue infection with central nervous system involvement in our cohort may be better defined as encephalopathy based on normal CSF findings and inability to isolate the virus from CSF, except in one patient. Moreover all the patients recovered without neurologic sequelae. However, the patient with facial palsy who also had virus present in CSF implies the possibility of true encephalitis. The virus has been rarely found in brain specimens of fatal cases of DCI. 9 We believe that most of the DCI cases were encephalopathies and that only in a small number of cases were they true encephalitis. 1 This is the first study to our knowledge that examined the contribution of DV to encephalitic illness. The high relative rate of DV infection in this study may be a result of the high incidence of DHF in Thailand in 1997. Dengue infection should always be considered in children presenting with encephalitis-like illness, particularly during the rainy months, in endemic areas. HSV was responsible for ∼10% of viral encephalitis reported in other studies 10 and also in this study. Recovery of HHV-6 DNA in CSF could be from acute infection or reactivation of latent infection. The unknown etiologies in 14 children (35%) could be Mycoplasma, HSV with false negative PCR, EBV, adenovirus or respiratory viruses that we did not investigate. Besides the supportive therapy that is crucial in treatment of encephalitis, acyclovir was given to most patients with encephalitis of uncertain etiology or suspected HSE. Given the 10% proportion of HSE, however, with grave prognosis and the inability to distinguish HSE from encephalitides of other causes, a safe antiviral agent such as acyclovir may be worthwhile for empiric treatment. The findings from our study underscore the importance of vaccination. A better coverage for JE and MMR vaccination might prevent 20% (8 of 40) of encephalitis. Moreover better mosquito control would also reduce DCI, which comprises a significant portion of encephalitic illness. Acknowledgment. This study was partially funded by Siriraj China Medical Board Foundation Grant 75-348-258.