Abstract
Background. Encephalitis is a complex syndrome, and its etiology is often not identified. The California Encephalitis Project was initiated in 1998 to identify the causes and further describe the clinical and epidemiologic characteristics of encephalitis.
Methods.A standardized report form was used to collect demographic and clinical data. Serum, cerebrospinal fluid, and respiratory specimens were obtained prospectively and were tested for the presence of herpesviruses, arboviruses, enteroviruses, measles, respiratory viruses, Chlamydia species, and Mycoplasma pneumoniae. The association between an identified infection and encephalitis was defined using predetermined, organism-specific criteria for confirmed, probable, or possible causes.
Results. From 1998 through 2005, a total of 1570 patients were enrolled. Given the large number of patients, subgroups of patients with similar clinical characteristics and laboratory findings were identified. Ten clinical profiles were described. A confirmed or probable etiologic agent was identified for 16% of cases of encephalitis: 69% of these agents were viral; 20%, bacterial; 7%, prion; 3%, parasitic; and 1%, fungal. An additional 13% of cases had a possible etiology identified. Many of the agents classified as possible causes are suspected but have not yet been definitively demonstrated to cause encephalitis; these agents include M. pneumoniae (n = 96), influenza virus (n = 22), adenovirus (n = 14), Chlamydia species (n = 10), and human metapneumovirus (n = 4). A noninfectious etiology was identified for 8% of cases, and no etiology was found for 63% of cases.
Conclusions. Although the etiology of encephalitis remains unknown in most cases, the recognition of discrete clinical profiles among patients with encephalitis should help focus our efforts toward understanding the etiology, pathogenesis, course, and management of this complex syndrome.
Encephalitis is a complex, severe, neurological syndrome that is associated with significant morbidity and mortality, and the etiology of the syndrome often is not identified [1]. The California Encephalitis Project (CEP) was initiated to identify the etiologic agents and define the clinical and epidemiologic characteristics associated with encephalitis. A previously published report summarized data for the first 334 patients with encephalitis enrolled in the CEP [2]. In addition, individual agent-specific case series have also been reported [3–8]. The present article provides an updated summary and overview. Based on >1500 cases of encephalitis, the study has identified (1) patient profiles with similar clinical characteristics and laboratory findings, (2) etiologies that might otherwise have been missed, and (3) potential novel etiologic agents of encephalitis. These data provide new insights into the features of this devastating syndrome, and they provide new ways to consider possible causes and patient management.
Patients, Materials, and Methods
Case finding and enrollment. Patients were referred to the study by treating physicians and were enrolled if they were immunocompetent, were ⩾6 months of age, and met the CEP case definition of encephalitis. A “case patient” was defined as a patient hospitalized with encephalopathy (defined by a depressed or altered level of consciousness lasting ⩾24 h, lethargy, or a personality change) with ⩾1 of the following characteristics: fever, seizure, focal neurological findings, pleocytosis, or electroencephalography or neuroimaging findings consistent with encephalitis.
The referring physician completed a case history form that included information on exposures (e.g., animal or arthropod contact, recent immunization, medications, etc.), travel, laboratory findings, and clinical and demographic characteristics. CSF, respiratory, and acute- and convalescent-phase serum specimens were requested for diagnostic testing. When a brain biopsy or autopsy was performed, brain tissue specimens were also requested by the CEP.
CEP testing. A core battery of tests was performed on the specimens obtained from each patient (table 1). Testing for 16 potential infectious agents of encephalitis, including herpesviruses (herpes simplex virus [HSV] 1 and HSV-2), varicella-zoster virus (VZV), enteroviruses (EVs), measles, St. Louis encephalitis virus, Western equine encephalitis virus, West Nile virus (WNV), Epstein-Barr virus (EBV), respiratory viruses, and bacteria (Mycoplasma pneumoniae and Chlamydia species). Details about CEP testing are described in table 1. Specialized testing for serum antibodies was done using EIA or indirect fluorescent antibody testing performed using standard methods [12]. Agent-specific PCR analyses were performed at the Viral and Rickettsial Disease Laboratory (California Department of Health Services; Richmond, CA) [9–11] and other reference laboratories (e.g., the Centers for Disease Control and Prevention; Atlanta, GA). Testing for additional agents was performed on the basis of exposure or travel history, time of year, clinical symptoms, the request of physicians, and the availability of appropriate specimens (table 2). Fungal and prion testing were performed by the referring institution. Methods are further described elsewhere [2].
California Encephalitis Project core testing, 1998–2005.
Details of California Encephalitis Project (CEP) selective testing, 1998–2005.
Input from the referring institution. Staff from the CEP were in telephone contact with the referring institution for ⩾2 weeks, depending on the severity of the case. The follow-up calls provided information on additional laboratory and neuroimaging findings, as well as updates on each patient's condition.
Classification of the association between a pathogen and encephalitis. We defined the association between an identified agent and the encephalitis case as confirmed, probable, or possible, on the basis of the type of specimen in which the potential etiologic agent was detected, the strength of the previously established associations between the agent and encephalitis, and the clinical and epidemiologic characteristics of the disease (table 3).
Diagnosis category definitions.
Statistical analysis. The data were analyzed using Fisher's exact test or the Kruskal-Wallis test, as appropriate, with statistical significance denoted by P ⩽.05.
Results
Characteristics of Study Patients
From 1998 through 2005, a total of 2494 patients were referred to the CEP from 195 institutions throughout California. Of these patients, 924 were excluded for ⩾1 of the following reasons: insufficient samples or data were provided or patients had severe immunosuppression, had not been hospitalized, had a history of prior severe neurological impairment, or were <6 months old. A total of 1570 patients were eligible for evaluation. Of the 1570 study patients, 1533 (98%) provided CSF specimens, 1540 (98%) provided acute-phase serum specimens, and 696 (44%) provided convalescent-phase serum specimens. Demographic and clinical characteristics and laboratory data are summarized in table 4. The demographic characteristics of CEP patients differed from those of the California population in that pediatric patients (45% vs. 27%) and male patients (56% vs. 50%) were slightly overrepresented in the CEP. In addition, blacks (10% vs. 6%) were overrepresented in the CEP population, whereas whites (41% vs. 45%) and Hispanics (30% vs. 35%) were slightly underrepresented.
Summary of characteristics of and findings for 1570 California Encephalitis Project patients, 1998–2005.
Infectious and Noninfectious Etiologies
A confirmed or probable infectious cause of encephalitis was determined for 248 CEP patients, and a noninfectious cause was determined for 122 patients. For cases of encephalitis with an infectious cause, viral pathogens were most commonly detected, followed by bacterial, parasitic, prion, and fungal etiologies (table 5). Some patients were found to have infectious meningitis rather than encephalitis. A possible infectious cause was identified for an additional 208 patients (13%).
Aggregate data for California Encephalitis Project patients with encephalitis with a confirmed or probable etiology, by etiologic agent.
Confirmed and Probable Viral Agents
A total of 170 patients had encephalitis with a confirmed or probable viral etiology, and the most commonly identified viral agents were EV (for 25% of cases) and HSV-1 (for 24% of cases). The demographic, clinical, and laboratory data for patients with encephalitis due to viral and other etiologic agents, which were identified in association with ⩾5 cases, are summarized in table 6. The median age of individuals with HSV-1 (54.0 years), VZV (44.0 years), and WNV (66.0 years) encephalitis was greater than that of individuals with EV (12.0 years) and EBV (11.0 years) encephalitis and measles causing subacute sclerosing panencephalitis (12.0 years). WNV and EV encephalitis occurred more commonly in the summer, but no other seasonal association was observed.
Characteristics of and findings for patients with encephalitis, according to viral or bacterial etiology.
As expected, many (75%) of the patients with viral encephalitides presented with fever. Important exceptions included individuals with measles causing subacute sclerosing panencephalitis, VZV infection, and hepatitis C. No consistent prodromes were seen in association with any specific virus except influenza virus (respiratory prodromes) and rotavirus (gastrointestinal prodrome). Seizures were noted in 38% of patients with viral encephalitis, most commonly among patients with measles causing subacute sclerosing panencephalitis (83%), human herpesvirus 6 infection (75%), and HSV-1 infection (59%). The initial MRI findings were abnormal for 87 patients (60%) and were most frequently abnormal for patients with HSV-1 (93%).
Nonviral Infectious Agents
A confirmed or probable nonviral agent was identified for 78 cases (5%), including cases of meningitis. The bacterial agents identified were diverse, including Mycobacterium tuberculosis (n = 19), Bartonella species (n = 13), M. pneumoniae (n = 2), and Tropheryma whippelii (n = 1); in addition, 14 cases of pyogenic bacteria were identified (table 6).
Seven parasitic infections were identified (4 due to Balamuthia mandrillaris and 3 due to Baylisascaris procyonis). Of the patients with infection due to B. mandrillaris, 75% were male, all were Hispanic [13], and their age range was 3–64 years. Pleocytosis (median CSF WBC count, 126 cells/mm3), an elevated protein level (median, 945 mg/dL), and abnormal neuroimaging findings were noted for all patients. The infections due to B. procyonis occurred in children whose age ranged from 11 months to 17 years, and all of these children had CSF and peripheral eosinophilia.
Fungal testing for Cryptococcus neoformans and Coccidioides immitis was performed by the referring hospital at the discretion of the clinician. Three individuals in the CEP were identified as having fungal infections (2 due to C. immitis and 1 due to C. neoformans). All 3 patients had an elevated CSF WBC count (median, 117 cells/mm3), an elevated CSF protein level (median, 176 mg/dL), and a depressed CSF glucose level (median, 25 mg/dL).
Possible Agents and/or Causes
A total of 204 patients who had a possible etiologic agent identified were classified as belonging to 1 of 2 categories, as outlined in table 3. Acute infection with M. pneumoniae was identified in 96 patients, a higher frequency than was noted for any other agent. However, most of these cases were classified as “possible” because, although there was serological evidence (in 88 cases) and/or PCR evidence (in 18 cases) of an acute infection, the organism was not detected from a CNS site. Other possible etiologies included influenza A and B viruses (22 cases), adenovirus (14), HSV-1 (13), Chlamydia species (10), human metapneumovirus (4), VZV (4), human herpesvirus 6 (2), respiratory syncytial virus (2), Brucella species (2), rotavirus (2), parainfluenza virus (1), Bartonella species (1), EBV (1), Creutzfeldt-Jakob prion disease (1), and mixed respiratory infections (5). Twenty-eight cases of possible EV infection were identified; 15 were identified by detection of EV in respiratory samples, whereas the remainder were identified by serum EV IgM assay.
Noninfectious Etiologies
For ∼8% of cases, a noninfectious etiology was identified by the referring institution. These noninfectious causes were associated with 52 cases of autoimmune disease and/or vasculitis (43%), 33 neoplastic cases (27%), 7 metabolic cases (6%), and 30 cases due to other disorders (25%).
Comparative CSF Values
As shown in figure 1 and figure 2, patients with cases due to viral and bacterial agents had a wide range of CSF WBC counts and protein levels, as did patients with cases due to noninfectious etiologies. Not surprisingly, comparison of the CSF laboratory values indicates that patients who had an infectious agent diagnosed had a higher CSF WBC count than did patients who had a noninfectious agent diagnosed (median CSF WBC count, 53.5 vs. 9.5 cells/mm3; P < .001). However, the difference in CSF protein levels was not significant (median level, 71.0 vs. 67.0 mg/dL).
Number of patients with confirmed or probable etiologic agents of encephalitis identified, by CSF WBC count and type of etiologic agent. The etiologic agent(s) and the number of patients with the etiologic agent(s) identified are listed according to diagnosis category definition. 1a comprises 6 patients with Epstein-Barr virus (EBV) infection, 5 with enterovirus (EV) infection, 5 with herpes simplex virus 1 (HSV-1) infection, 3 with varicella-zoster virus (VZV) infection, 2 with West Nile virus (WNV) infection, 2 with measles causing subacute sclerosing panencephalitis (SSPE), 2 with rabies, and 1 with rotavirus infection. 1b comprises 15 patients with HSV-1 infection, 7 with EV infection, 6 with WNV infection, 4 with VZV infection, 4 with EBV infection, 3 with measles causing SSPE, 1 with acute HIV infection, 1 with hepatitis C, 1 with human herpesvirus-6 (HHV-6) infection, 1 with rabies, and 1 with rotavirus infection. 1c comprises 13 patients with EV infection, 9 with WNV infection, 9 with HSV-1 infection, 7 with VZV infection, 3 with EBV infection, 2 with HHV-6, 2 with hepatitis C, 1 with measles causing SSPE, and 1 with acute HIV infection. 1d comprises 8 patients with HSV-1 infection, 7 with VZV infection, 6 with EV infection, 1 with WNV infection, 1 with HSV-2 infection, 1 with acute HIV infection, and 1 with EBV infection. 1e comprises 4 patients with EV infection, 3 with HSV-2 infection, 2 with HSV-1 infection, 1 with EBV infection, 1 with HHV-6 infection, and 1 with rotavirus infection. 2a comprises 11 patients with infection due to Bartonella species, 1 with infection due to Mycoplasma pneumoniae, 1 with Rocky Mountain spotted fever, 1 with poststreptococcal glomerulonephritis, 1 with infection due to Streptococcus agalactiae, 1 with urosepsis and infection due to Klebsiella species, and 1 with Sydenham chorea. 2b comprises 2 patients with infection due to Mycobacterium tuberculosis, 2 with infection due to Streptococcus pneumoniae, 1 with infection due to M. pneumoniae, 1 with infection due to Tropheryma whippelii, 1 with infection due to Staphylococcus aureus, and 1 with endocarditis. 2c comprises 9 patients with infection due to M. tuberculosis, 1 with infection due to Bartonella species, 1 with infection due to Neisseria meningitidis, and 1 with infection due to Streptococcus viridans. 2d comprises 4 patients with infection due to M. tuberculosis and 1 with group C β-hemolytic Streptococcus infection. 2e comprises 4 patients with infection due to M. tuberculosis, 2 with infection due to N. meningitides, 1 with infection due to methicillin-resistant S. aureus, and 1 with infection due to S. aureus. 3a comprises 1 patient with infection due to Baylisascaris procyonis. 3b comprises 2 patients with infection due to B. procyonis. 3c comprises 2 patients with infection due to Balamuthia mandrillaris. 3d comprises 1 patient with infection due to B. mandrillaris. 4a comprises 1 patient with infection due to Cryptococcus neoformans. 4b comprises 1 patient with infection due to Coccidioides immitis. 4c comprises 1 patient with infection due to C. immitis.
Number of patients with confirmed or probable etiologic agents of encephalitis identified, by CSF protein level and type of etiologic agent. The etiologic agent(s) and the number of patients with the etiologic agent(s) identified are listed according to diagnosis category definition. 1a comprises 16 patients with enterovirus (EV) infection, 11 with herpes simplex virus 1 (HSV-1) infection, 8 with Epstein-Barr virus (EBV) infection, 4 with measles causing subacute sclerosing panencephalitis (SSPE), 2 with hepatitis C, 2 with rabies, 2 with rotavirus infection, 1 with human herpesvirus 6 (HHV-6) infection, and 1 with varicella-zoster virus (VZV) infection. 1b comprises 16 patients with HSV-1 infection, 13 with EV infection, 13 with West Nile virus (WNV) infection, 9 with VZV infection, 5 with EBV infection, 2 with measles causing SSPE, 1 with hepatitis C, and 1 with HHV-6. 1c comprises 8 patients with HSV-1 infection, 8 with VZV infection, 5 with EV infection, 4 with WNV infection, 4 with HSV-2 infection, 2 with EBV infection, 2 with HHV-6 infection, 2 with acute HIV infection, and 1 with rabies. 1d comprises 1 patient with VZV infection, 1 with EV infection, and 1 with acute HIV infection. 2a comprises 8 patients with infection due to Bartonella species, 1 with infection due to Mycoplasma pneumoniae, 1 with poststreptococcal glomerulonephritis, 1 with urosepsis and infection due to Klebsiella species, and 1 with Sydenham chorea. 2b comprises 4 patients with infection due to M. tuberculosis, 3 with infection due to Bartonella species, 2 with infection due to Streptococcus pneumoniae, 1 with infection due to Tropheryma whippelli, 1 with infection due to M. pneumoniae, 1 with infection due to Streptococcus viridans, and 1 with Staphylococcus aureus endocarditis. 2c comprises 13 patients with infection due to M. tuberculosis, 1 with infection due to Neisseria meningitidis, and 1 with infection due to S. aureus. 2d comprises 2 patients with infection due to M. tuberculosis, 2 with infection due to N. meningitidis, 1 with infection due to Bartonella species, 1 with group C β-hemolytic Streptococcus infection, and 1 with infection due to Streptococcus agalactiae. 3a comprises 2 patients with infection due to Baylisascaris procyonis. 3b comprises 1 patient with infection due to B. procyonis. 3c comprises 2 patients with infection due to Balamuthia mandrillaris. 4a comprises 1 patient with infection due to Cryptococcus neoformans. 4b comprises 1 patient with infection due to Coccidioides immitis. 4c comprises 1 patient with infection due to C. immitis.
Clinical Profiles
Ten distinct constellations of clinical characteristics—4 focal and 6 generalized clinical profiles—were recognized in approximately one-half of the patients. For the remaining patients, we could not identify groups with common, distinct characteristics. Patients within a profile were analyzed to determine whether they had other characteristics in common, such as prodromal symptoms, course of illness, demographic characteristics, infectious or noninfectious causes of illness, and/or outcomes (table 7).
Clinical profiles of California Encephalitis Project cases, 1998–2005.
Focal Group of Encephalitides
Temporal lobe involvement. The most common profile (in 144 cases) was encephalitis with temporal lobe enhancement noted on MRI or CT. Not surprisingly, HSV-1 was the most common agent identified (in 32 [22%] of cases). Other herpesviruses identified in this profile group included VZV (5 cases), EBV (2 cases), and human herpesvirus 6 (1 case). An additional 42 patients had temporal lobe activity noted on electroencephalography but had no enhancement noted on CT or MRI. The etiologic agents identified for this group of patients were similar to those identified for patients with temporal lesions visualized on neuroimaging.
Movement and/or extrapyramidal disorders. Movement disorders were observed in 47 patients. These patients were younger (median age, 11 years) than other patients in the CEP (median age, 23 years) (P < .001). The length of hospitalization was prolonged for this group of patients (median duration, 39 days), compared with that noted for other patients in the CEP (median duration, 11 days) (P < .001), but the mortality rate for this group (9%) was comparable to the overall morality rate noted in the CEP study (11%).
Cerebellar disorders. Eighty-seven patients presented with a predominance of cerebellar signs (i.e., ataxia and dysmetria) and/or focal cerebellar lesions noted on MRI. The median age of this group was 15 years, compared with a median age of 23 years for other patients in the CEP (P = .001). The mortality rate for this group (2%) was lower than the overall mortality rate noted in the CEP (11%) (P = .01). No single infectious agent was found to be predominant in this group, and a relatively high percentage (16%) of noninfectious etiologies were identified.
Hydrocephalus. Thirty-two patients who presented with encephalitis were found to have new-onset hydrocephalus. A relatively high percentage of nonviral organisms (47%) were identified in this profile group, and these organisms included bacterial (11 cases), fungal (2 cases), and parasitic (2 cases) agents.
Generalized Group of Encephalitides
Diffuse cerebral edema. Forty-seven patients presented with or developed diffuse generalized cerebral edema within 7 days of admission, as evidenced by CT, MRI, or autopsy findings. Evidence of inflammation was minimal (median CSF WBC count, 8 cells/mm3). Although the CNS findings were similar to those associated with Reye syndrome, none of the patients had significant elevation of transaminase levels, hypoglycemia, or hyperammonemia. A confirmed or probable infectious cause was identified for only 2 cases. Of note, 34 patients (72%) died within 7 days after hospitalization.
Intractable seizures. Sixty-two patients either presented with or developed intractable seizures requiring general anesthesia or a barbiturate-induced coma to interrupt status epilepticus. Most patients (69%) were <18 years of age (median age, 10 years). These patients had prolonged hospital stays (median duration, 46 days), compared with other patients who were referred to the CEP (median duration, 11 days) (P < .001). Most cases (73%) in this patient group had no causative agents identified. Twenty percent of these patients died before discharge from the hospital, and of the patients who survived, most required extensive rehabilitation.
Seizures with rapid recovery. In contrast to the patient group with intractable seizures, 25 patients presented with seizures but had a rapid recovery and were discharged from the hospital within 7 days after hospital admission.
Psychosis presentation. Fifty-one patients presented with new-onset psychosis. Noninfectious causes were noted more frequently (in 20% of cases) than infectious causes (12%) in this group. Thirty (59%) of the cases in this group had no etiology identified.
Recurrent or chronic inflammatory CNS disease. Twenty-nine patients who were referred to the CEP had experienced at least 1 previous hospitalization for CNS symptoms. The median age of the patients in this profile group was 41 years. Other than a single case of Creutzfeldt-Jakob disease, no confirmed or probable infections were identified in this group, although a number of autoimmune and other noninfectious entities were identified.
Multifocal white matter disease. A total of 120 patients had multifocal white-matter lesions noted on neuroimaging. These patients often reported viral prodromal symptoms (34% reported upper respiratory tract infection, and 41% reported gastrointestinal symptoms). A number of possible agents were identified, including M. pneumoniae and respiratory viruses, consistent with a postinfectious disease process.
Discussion
Encephalitis is an inflammation of the brain parenchyma and presents as an alteration in consciousness, fever, headache, seizures, and/or focal neurologic signs [14]. However, without the identification of a neurotropic agent or confirmation by brain tissue analysis, the diagnosis of encephalitis is presumptive and is based on clinical characteristics. The case definition used in the present study reflects these nonspecific clinical manifestations of encephalitis and, thus, is inherently sensitive but not specific. Clinicians refer patients to the CEP on the basis of clinical characteristics that are compatible with encephalitis and may overlap with the clinical characteristics of other disease entities, including noninfectious conditions, bacterial and fungal infections, and traditional viral causes of encephalitis. Even after an exhaustive examination has been conducted, the condition of many patients remains undiagnosed, and management is problematic, demonstrating the need for further study.
The CEP study provided a unique opportunity to systematically investigate the cases of a large number of patients who presented with symptoms of encephalitis. The present report provides data on 1570 patients over a 7-year period. This large cohort made it possible to identify clinical profiles on the basis of the presence of predominant clinical signs and symptoms (e.g., intractable seizures and movement disorders) or neuroimaging findings (e.g., temporal lobe involvement and hydrocephalus). Grouping by clinical profile is a new approach and may clarify some aspects of this syndrome. First, the profiles have allowed us to estimate prognosis and consider management strategies. In the cerebral edema profile group, many patients experienced tentorial herniation and died within 7 days, suggesting that measures to control elevated intracranial pressure may improve outcomes [15]. Second, knowledge of the natural history of a given profile may also be helpful in counseling families. For example, patients who have seizures that terminate in a few days seem to have an excellent prognosis, whereas patients who have movement disorders or intractable seizures often have a poor prognosis. Additional studies, however, are needed to determine the precision of these estimates and the effectiveness of profile-specific management strategies. Third, the profiles may also provide a better way to consider etiology. Cases within a profile group may be more likely to have a single or related etiology, possibly making it easier to associate an agent with the disease. For example, HSV-1 is commonly found in patients with temporal lobe involvement, but our study identifies other etiologies that should be considered [16]. Finally, the clinical grouping may promote a systematic investigation into pathogenesis.
Interestingly, agents that are considered to be important causes of encephalitis worldwide, such as arboviruses [17], were infrequently detected in this study. No cases of Western equine encephalitis or St. Louis encephalitis were identified, and only 19 cases of WNV infection were identified through the CEP in 2004 and 2005 (the initial years of significant WNV activity in humans in California). There were, however, 111 cases of encephalitis caused by WNV infection in California in 2004 that were not included in the present study. The cases of WNV infection not referred to the CEP illustrate an important limitation of this study. The CEP is not a population-based study, and, thus, cases caused by agents that can be readily diagnosed with laboratory studies available in the community (e.g., HSV-1 and WNV) are underrepresented in this study. Indeed, clinicians often refer specimens to the West Nile Surveillance Project (which is also at the Viral and Rickettsial Disease Laboratory), and, if the test results are negative, the clinicians request that the patient be enrolled into the CEP. Furthermore, in the CEP, only 2.5% of cases were due to HSV-1, compared with other studies in which 8%–20% of all cases were due to HSV-1 [18, 19].
Bacterial, parasitic, and fungal agents were identified in a small but significant number of cases. The most frequently confirmed nonviral infection found in this cohort was caused by M. tuberculosis. M. tuberculosis infection of the CNS usually presents as chronic meningitis; however, patients with M. tuberculosis infection may appear to have encephalitis [20]. Furthermore, the CSF laboratory values for patients with M. tuberculosis infection of the CNS may be atypical and may resemble those associated with infection caused by viral pathogens. Other uncommon but nonetheless important pathogens that were etiologically confirmed included Bartonella species, B. mandrillaris, and B. procyonis.
More than 10% of patients had evidence of an acute infection without evidence sufficient to establish a causal relationship to the illness. This occurred often with respiratory pathogens when laboratory studies demonstrated an acute infection, either by serological tests or molecular testing of respiratory specimens. Although M. pneumoniae is commonly accepted as a cause of encephalitis, infections are typically diagnosed on the basis of an increase in serial serological titers and, rarely, by detection in CSF by PCR or by isolation from CSF or brain tissue specimens [21]. Chlamydia pneumoniae is a less-commonly described neuropathogen and is usually diagnosed by serological testing (of serum and CSF specimens) rather than by direct detection within the CNS [22]. Similarly, influenza viruses and other respiratory viral agents are associated with encephalitis, but the agents are rarely found in the CNS [23]. The large number of patients included in this study will provide opportunities to reconsider ways to determine causal relationships between these infections and encephalitis.
The failure to identify an etiologic agent for many of these cases is, in part, because of a referral bias toward diagnostically challenging cases; other factors may include insufficiently sensitive tests, lack of access to appropriate specimens (especially convalescent-phase serum specimens), suboptimal specimen handling, and the possible presence of a novel pathogen. In addition, some cases may have a noninfectious etiology that was not identified. Another factor in the low rate of identification of a causative agent is the use of stringent criteria for classification of confirmed or probable agents. With the use of these criteria, some agents that otherwise might be causally associated with the disease were considered to be only possible etiologic agents. The low rate of diagnosis of the etiology of encephalitis underscores the need for better tools and new approaches to identifying the etiology.
The present study demonstrates the diversity of clinical, epidemiologic, and etiologic characteristics of encephalitis. This diversity no doubt contributes to the difficulty in the diagnosis, management, and treatment of encephalitis in patients. The recognition of subgroups of patients with encephalitis as described in this report may be useful to clinicians in identifying the etiology of encephalitis, selecting the treatment strategy, and anticipating the course of illness and the prognosis for the patient.
Acknowledgments
We thank the laboratory staff in the Viral and Rickettsial Disease Laboratory and Microbial Diseases Laboratory for performing diagnostic testing. We also thank the clinicians who referred cases to the California Encephalitis Project. We gratefully acknowledge the assistance received from Karen Bloch, Abbie Collins, Audrey Foster-Barber, Shilpa Gavali, Sabrina Gilliam, Cynthia Jean, Nino Khetsuriani, Ashley LaMonte, Janice Louie, Jim Sejvar, and Allan Tunkel.
Financial support. Centers for Disease Control and Prevention Emerging Infections Program (U50/CCU915546-09).
Potential conflicts of interest. All authors: no conflicts.
Comments