Encephalitis remains associated with a substantial burden of morbidity and mortality in children worldwide. The true incidence is unknown and various rates of up to 27.7 cases per 100,000 child per year have been reported.1,2 Although more than 100 pathogens have been associated with encephalitis, the etiology often remains undefined.3,4 Definition of the precise etiology should be vigorously sought, because apart from allowing for optimization of treatment, it can lead to the development of preventive strategies.
Not much is known about the epidemiology of encephalitis worldwide. Following an initial report from the island of Crete, Greece,5 we aimed to prospectively investigate the characteristics of childhood encephalitis in 3 regions of Greece with emphasis on causative pathogens, clinical, laboratory, and neuroimaging findings and outcome.
PATIENTS AND METHODS
A prospective study was conducted from January 2005 to December 2007 at 3 tertiary-care university-affiliated general hospitals. The hospitals were referral centers in each area, covering different regions (northern, north-western and southern) in Greece. Definitions and investigation protocol were identical in all 3 participating centers.
Children aged 1 month to 15 years meeting our case definition of encephalitis were enrolled. Encephalitis was defined as encephalopathy (depressed or altered level of consciousness lasting ≥24 hours, lethargy, or change in personality requiring hospitalization) with ≥1 of the following symptoms: fever, seizures, focal neurologic findings, cerebrospinal fluid (CSF) pleocytosis (>5 WBC/μL), or electroencephalogram (EEG) or neuroimaging findings suggestive of encephalitis.3 Diagnosis was further classified as confirmed, probable or possible based on the certainty of pathogen identification and on the documented ability of the identified pathogen to affect brain tissue.3 In brief, diagnosis was defined as confirmed in the cases where agent was both a well-established cause of encephalitis and detected in CSF specimens; diagnosis was defined as probable in the cases where agent was either a well-established cause of encephalitis but detected in other than CSF specimens or not a well-established cause of encephalitis but detected in CSF specimens; and diagnosis was defined as possible in cases that agent was a well-established cause of encephalitis but serological evidence of infection only was suggestive or agent was not well-established cause of encephalitis but strong serological or culture-based evidence existed outside the central nervous system (CNS). Patients with known immunodeficiency or neurologic, metabolic, vascular and neoplastic disorders were excluded.
Patient data included demographics, clinical manifestations, microbiology and virology results, EEG and neuroimaging findings, and outcome. Acute-phase and convalescence-phase serum samples were tested for herpes simplex virus 1 and 2 (HSV 1-2), varicella-zoster virus, cytomegalovirus (CMV), Epstein-Barr virus (EBV), human herpes virus 6 (HHV-6), adenoviruses, group A and B coxsackieviruses and enteroviruses, Bartonella spp., Mycoplasma pneumoniae, and chlamydia. Serologic testing for viruses was done by the use of ELISA, while serologic testing for EBV included the test for heterophile antibody and Epstein-Barr nuclear and anticapsular antigen. Serologic evidence of infection was determined by the presence of specific IgM or by seroconversion, which was defined by a 4-fold increase in titer between acute and convalescence sera. All children underwent tuberculin skin tests and all CSF samples were cultured. Additional testing, including polymerase chain reaction (PCR) for Streptococcus pneumoniae, Neisseria meningitidis, M . pneumoniae, Mycobacterium tuberculosis, Bartonella, HSV 1-2, EBV, HHV-6, and enteroviruses, was performed in CSF samples for individual patients, depending on clinical presentation, potential pathogens, and history. The differences in continuous variables were tested by using Mann-Whitney U test. For categorical variables χ2 test was performed. P < 0.05 was considered significant.
During the study period, 42 children (24 boys, 18 girls; mean age: 5.9, median: 5.05, range: 0.2–14.8 years) fulfilled the study criteria for encephalitis. Of these patients, 13 were residents of southern, 17 of North-Western and 12 of Northern Greece. No outbreaks were observed and no particular geographic or seasonal predilection was documented (χ2 test for goodness of fit, P = 0.22).
Clinical manifestations included fever (64%), gastrointestinal (43%) and respiratory (19%) symptoms, altered consciousness (45%), hemiparesis or cranial nerve involvement (14%), ataxia, dysmetria or unsteadiness (33%), and seizures (14%) including 3 children with status epilepticus. CSF pleocytosis at presentation was documented in all but 1 patient (median: 40, range: 2–9500 WBC/μL). Neutrophils ranged from 4 to 9405 (median: 80 cells/μL), lymphocytes from 3 to 510 (median: 89 cells/μL), monocytes from 0 to 306 (median: 0 cells/μL), CSF protein from 7.5 to 128 (median: 37 mg/dL), and CSF glucose from 37 to 107 (median: 63 mg/dL). Values of CSF WBC/μL, protein and glucose concentrations did not differ significantly among cases with confirmed, probable, possible, or unknown etiology, as was the case for bacterial or viral encephalitis cases.
The cause of encephalitis was identified in 24 of 42 (57.1%) children (Table, Supplemental Digital Content 1, http://links.lww.com/A1262 and Table, Supplemental Digital Content 2, http://links.lww.com/A1263). In 10 of 24 cases (41.7%), diagnosis was classified as confirmed, while in 8 of 24 (33.3%) diagnosis was probable and in 6 of 24 (25%) possible according to the aforementioned criteria. CSF PCR contributed to diagnosis and classification in 8 children. Viruses were the leading causes of infection (17/24, 70.8%). Herpes viruses predominated (3 HSV cases, 3 varicella-zoster virus, 1 EBV, 1 CMV, and 2 HHV-6), followed by enteroviruses (6 children). The viral origin of encephalitis was confirmed in 1 case each with CSF PCR for HSV, EBV, HHV-6, and enterovirus. Evidence for bacterial cause was detected in 7 of 24 children (29.2%) and in 6 cases diagnosis was classified as confirmed. S. pneumoniae was detected in 2 patients by CSF PCR and in one of them grew in blood culture as well. CSF culture yielded L. monocyotogenes in 1 patient and M. pneumoniae and B. quintana encephalitis was confirmed by CSF PCR in 1 patient each. In a child with positive interferon-γ release assay (Quantiferon; Cellestis Ltd, Carnegie, Victoria, Australia) but negative tuberculin skin test, M. tuberculosis was confirmed by CSF PCR.
Among 28 patients who underwent EEG, abnormalities were demonstrated in 12 (42.9%). Neuroimaging (CT, MRI, or both) was performed in all but 3 patients, of whom one underwent brain ultrasonogram. Neuroimaging abnormalities were observed in 4 of 22 (18.2%) of the CTs performed and in 14 of 34 (41.1%) of the MRIs. Four children, 3 with HSV encephalitis, had initially normal CTs but demonstrated MRI abnormalities in 5 to 8 days. MRI findings, consistent with acute disseminated encephalomyelitis (ADEM) were identified in 4 patients, 2 of whom had HSV infection and the other 2 did not have an etiology identified.
Median duration of hospitalization was 12 days and intensive care was required for 7 (17%) patients. There were no deaths in our study. At discharge, 10 of 42 (23.8%) children had remarkable neurologic deficits, including truncal ataxia (7), hyperflexia (3), hemiparesis (2), dysarthria (2), behavioral disorder (1), visual (1), and auditory (1) impairment. Three patients developed hydrocephalus (following Bartonella, Listeria, and M. tuberculosis infection), requiring ventriculoperitoneal shunt or external CSF drainage. All patients were followed after discharge and 5 had persisting abnormalities (3 with neurologic signs, 1 with hyperactivity disorder, and 1 with phobias); among these abnormalities, hyperactivity disorder and phobias could not be definitely attributed to encephalitis. The girl with M. tuberculosis encephalitis was still hospitalized 7 months after onset, with deteriorating hydrocephalus, Aspergillus ventriculitis, and guarded prognosis. One child continued to receive anticonvulsants and the child with hyperactivity was receiving psychologic support.
Using a standardized investigation protocol, evidence of acute encephalitis was found in 42 children in 3 regions of Greece during the 3-year period 2005 to 2007, with an estimated incidence of 4.5 admissions for encephalitis per 100,000 children per year. The annual incidence in the island of Crete was estimated to be 3.3 admissions per 100,000 children, very close to the rate of 2.6 reported for the years 2000 to 2004.5 Incidence rates for encephalitis cannot be precisely calculated, as diagnosis may be confused with that of meningitis or meningoencephalitis, or even not considered at all in mild cases. Previous studies have reported 2.6 to 10.5 cases per 100,000 children per year, although rates may be as high as 27.7 cases per 100,000 child-years.1,2,5
Our attempts to identify the confirmed or probable causative agent was only fruitful for 42.8% of the cases, a finding consistent with previous reports where etiology was established in 12% to 70% of cases.1,4–9 We managed to identify 13 different pathogens. Herpes viruses and enteroviruses were the leading agents and bacterial agents were proportionally less. HSV-related encephalitis seemed to present rather atypically, as temporal MRI findings were only revealed in 1 child, with the remaining 2 presenting with ADEM and medullar MRI findings. On the other hand, in only 1 case, RSV diagnosis was classified as “confirmed” based on PCR results. Recent studies from western and central European countries have reported considerable frequency of M. pneumoniae and arthropod-born encephalitis,7,10,11 but this may partially reflect extensive use of serology, as serology alone can lead to overestimation or cross-reaction of these particular pathogens. Of particular interest were the cases of B. quintana, S. pneumoniae, and M. tuberculosis encephalitis. B. quintana infection was complicated by Guillain-Barre syndrome and hydrocephalus; although B. quintana is not a known pathogen for encephalitis, both serology and CSF PCR were positive in this case.12 S. pneumoniae with thrombosis and widespread brain lesions and M. tuberculosis with salt wasting syndrome, profound hyponatremia, and serious neurologic sequelae. It is often difficult to distinguish bacterial encephalitis or meningoencephalitis from meningitis; in our study, we defined these cases as encephalitis rather than meningitis based on the prominent clinical, electrophysiological, and imaging features of brain damage.
A confirmed or probable diagnosis, using the criteria of Glaser et al,3 was only reached in 18 of 42 (42.9%) children. Reported rates of confirmed or probable diagnosis have ranged widely, depending on study geography and time.3,4 In older studies, etiologic diagnosis was based on isolation of the pathogen from brain at biopsy or autopsy.4 More recently, PCR has substantially contributed to identification of causative pathogens,13 because serology alone may lead to an overestimation of the confirmed or probable etiologies identified, particularly for CMV and enterovirus. These limitations of serology were taken into consideration when using the aforementioned criteria for diagnosis classification as confirmed, probable or possible. In our experience, CT and EEG were less helpful than MRI, which provided definite help in diagnosis, especially in cases of ADEM.14 Survival and long-term sequelae in our patients showed a favorable outcome. Persisting neurodevelopmental impairment signs were observed in 5 of 42 children, a rate lower than what previously reported.5,7,15 Wide-spread use of acyclovir seems to have contributed to this better outcome.13
Considerable limitations of this study included the identification of responsible pathogens by serology rather than by cultures or nucleic acid detection tests, and the definition difficulties in differentiating encephalitis or meningoencephalitis from meningitis, particularly in cases of common bacterial or mycobacterial pathogens. In 7 cases, bacterial CNS infection was defined as encephalitis or meningoencephalitis rather than as meningitis, as clinical and imaging findings clearly pointed to the presence of an inflammatory process of the brain in association with evidence of neurologic dysfunction.13 Additionally, our findings reflect only severe cases of encephalitis treated in referral hospitals and cannot be used as a reliable marker of incidence.
Despite these limitations, our study provides information on the epidemiology of childhood encephalitis in Southern Europe and Eastern Mediterranean basin, where information is scarce and where, to our knowledge, no prospective multicenter report has been conducted. The use of acyclovir and intensive care seems to have dramatically improved outcome; furthermore, the wide-spread use of vaccines is expected to contribute to alteration in etiology and epidemiology in the future.16 Despite the advent of new molecular and neuroimaging techniques, etiological diagnosis of encephalitis remains a challenge.
1. Koskiniemi M, Korppi M, Mustonen K, et al. Epidemiology of encephalitis
in children. A prospective multicentre study. Eur J Pediatr
2. Beghi E, Nicolosi A, Kurland LT, et al. Encephalitis
and aseptic meningitis, Olmsted County, Minnesota, 1950–1981: I, Epidemiology. Ann Neurol
3. Glaser CA, Gilliam S, Schnurr D, et al. In search of encephalitis
etiologies: diagnostic challenges in the California Encephalitis
Project, 1998–2000. Clin Infect Dis.
4. Cizman M, Jazbec J. Etiology of acute encephalitis
in Slovenia. Pediatr Infect Dis J
5. Ilias A, Galanakis E, Raissaki M, et al. Childhood encephalitis
in Crete, Greece
. J Child Neurol
6. Rantala H, Uhari M. Occurrence of childhood encephalitis
: a population-based study. Pediatr Infect Dis J
7. Kolski H, Ford-Jones EL, Richardson S, et al. Etiology of acute childhood encephalitis
at The Hospital for Sick Children, Toronto, 1994–1995. Clin Infect Dis
8. Fowler A, Stodberg T, Eriksson M, et al. Childhood encephalitis
in Sweden: etiology, clinical presentation and outcome. Eur J Paediatr Neurol
9. Ishikawa T, Asano Y, Morishima T, et al. Epidemiology of acute childhood encephalitis
. Aichi Prefecture, Japan, 1984–1990. Brain Dev
10. Christie LJ, Honarmand S, Talkington DF, et al. Pediatric encephalitis
: what is the role of Mycoplasma pneumoniae
11. Lesnicar G, Poljak M, Seme K, et al. Pediatric tick-borne encephalitis
in 371 cases from an endemic region in Slovenia, 1959 to 2000. Pediatr Infect Dis J
12. Mantadakis E, Spanaki AM, Psaroulaki A, et al. Encephalopathy complicated by Guillain-Barre syndrome and hydrocephalus and associated with acute Bartonella quintana infection. Pediatr Infect Dis J
13. Tunkel AR, Glaser CA, Bloch KC et al; Infectious Diseases Society of America. The management of encephalitis
: clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect Dis
14. Murthy SN, Faden HS, Cohen ME, et al. Acute disseminated encephalomyelitis in children. Pediatrics
15. Klein SK, Hom DL, Anderson MR, et al. Predictive factors of short-term neurologic outcome in children with encephalitis
. Pediatr Neurol
16. Koskiniemi M, Vaheri A. Effect of measles, mumps, rubella vaccination on pattern of encephalitis
in children. Lancet