The California Encephalitis Project — 1998-2005

Abstract & Commentary

By Carol A. Kemper, MD, FACP, Clinical Associate Professor of Medicine, Stanford University, Division of Infectious Diseases; Santa Clara Valley Medical Center, Section Editor, Updates; Section Editor, HIV, is Associate Editor for Infectious Disease Alert.

Dr. Kemper reports no financial relationships relevant to this field of study.

Synopsis: Despite the use of a large number of diagnostic tests, the etiology of the majority of cases remains obscure.

Source: CA. Glaser et al. Beyond viruses: Clinical profiles and etiologies associated with encephalitis. Clin Infect Dis 2006; 43: 1565-1577.

Since 1998, researchers at the viral and Rickettsial Disease Laboratory in Richmond, CA (near Berkeley) have conducted a large-scale epidemiologic project to identify causes of encephalitis in California. Requirements for participation included hospitalization with symptoms of encephalitis for ≥ 24 hours, along with at least one other criteria (fever, seizures, focal neurologic deficits, CSF pleocytosis, or abnormal neuroimaging studies or eeg). Patients were "referred" by their treating physician, who collected a detailed history, and forwarded acute and convalescent sera, and CSF and respiratory specimens. A core battery of tests was performed for each patient, including CSF PCR for HSV-1, HSV-2, VZV, enterovirus, and Mycoplasma pneumoniae; CSF viral culture; and serologic tests for SLE, WEE, WNV, M. pneumoniae, measles, HSV, VZV, Adenovirus, Chlamydia, Influenza A and B, and Bartonella spp. In addition, PCR testing and viral culture was performed on respiratory specimens. Other selective tests were performed depending on the clinical history (eg, rabies, borrelia, Colorado Tick Fever, etc.).

A total of 2494 patients were referred from 195 institutions throughout California from 1998-2005, of which 1570 met the criteria for analysis. Of these 98% provided adequate CSF and acute sera, and 44% provided adequate convalescent samples. The median age was 23 years (0-92 years); 56% were male; 58% required admission to ICU; 42% had seizures, and 18% were comatose. Two-thirds of the patients presented with fever, one-third had respiratory symptoms, one-third had gastrointestinal symptoms, and 13% had rash. The median CSF cell count was 23 (0-13,000), the median protein level was 57 (7-11,723), and the median glucose was 64 (6-533). Slightly more than half had abnormal neuroimaging studies. Eleven percent died.

A viral etiology was identified for 170 cases (10.8%), and was most frequently due to infection with either enteroviruses (25%) or HSV-1 (24%). Not surprisingly, the mean age of patients with HSV-1 (54 years), VZV (44 years), and WNV (66 years) was considerably older than those with enterovirus (12 years), measles (12 years), or EBV (11 years). Three-fourths of patients with a viral etiology presented with fever, which was more common than those with a non-viral etiology, and seizures occurred in 38% (most commonly, in those with measles, HSV-1 and human herpesvirus-6 infection). Arboviruses were only infrequently detected in this study, and there were no cases of WEE or SLE.

Non-viral infectious causes of encephalitis were identified in 78 persons (5%), and included a broad range of bacterial, mycobacterial, parasitic and, rarely, fungal etiologies. For example, bacterial etiologies included pyogenic bacteria in 14 cases; M. tuberculous, presenting as an acute encephalitis, occurred in 19; Bartonella spp. (n = 13), M. pneumonia (n = 2), and Tropheryma whippelii (n = 1). Parasitic infections were identified in 7 cases, including the raccoon tapeworm, Baylisascaris procyonis (n = 4). Fungal infections were identified in 3 cases, including coccidioidomycosis (n =2) and cryptococosis (n = 1).

An additional 204 patients (13%) were identified with possible etiologies. This generally involved a positive culture or PCR from a non-CSF site, but the organism could not be detected in CSF. The most common diagnosis in this category was M. pneumoniae in 94 cases (6%), but also included cases of enterovirus, influenza, HSV, chlamydia, human metapneumovirus, VZV, HHV-6, RSV, brucella, rotavirus, parainfluenza virus, bartonella, EBV, and CJD.

A non-infectious etiology was identified in 8% of patients, typically an autoimmune disorder or vasculitis, neoplasm, or metabolic disorder.

Commentary

Many of us in the local community participated in the project for years, and have been eager to see the results, which were surprising in their frequent inability to identify a causative agent in the majority of cases. I guess most of us believed that the newer molecular tools and the broad panel of assays available to the researchers (many of which were considered "experimental" and otherwise not available to clinicians) would provide a diagnostic advantage. However, despite the comprehensive nature of the laboratory investigation, a confirmed or probable infectious cause was identified in 248 (16%) of patients, and a non-infectious etiology was identified in 122 (8%). No etiology was found in 68% of cases. Roughly one-third of patients presenting with encephalitis had a definite, probable or possible infectious etiology, about two-thirds of which was viral and one-third was non-viral. Although 13% of the total cases had non-CSF evidence of an infectious etiology (most commonly, a respiratory illness) without confirmation of a CNS infection, a causal relationship was considered likely. Thus, thorough testing of respiratory specimens for potential causes of encephalitis, including testing for enterovirus, influenza, adenovirus, M. pneumoniae, RSV, and human metapneumovirus, may be helpful.

Some broad generalizations can be drawn from this data. Temporal lobe seizures or MRI findings were common (9%), and were most frequently associated with HSV-1 infection, although were also seen in patients with VZV, EBV, and HHV-6. Seizures were common in bartonellosis (85%), measles (83%), EBV (63%), as well as HSV-1 (59%) but were less common in enterovirus (28%), VZV (13%) and WNV (12%). Movement and extra-pyramidal disorders, sometimes with evidence of basal ganglia involvement on neuroimaging studies, tended to occur in young patients, with no predominant etiology identified, but the survival rate was not much better than that of the group as a whole (91%). Both seizures and extrapyramidal and movement disorders tended to result in long hospital stays (>1 month). Cerebellar disorders also tended to occur in young people, and were associated with a better survival outcome (98%).

Diffuse cerebral edema, occurring within the first 7 days of presentation, was associated with worse outcome (72% died). Although this presentation was most consistent with a Reyes-like condition, VZV infection could not be confirmed. A non-viral etiology was found in one-half of cases presenting with hydrocephalous, including mycobacterial, fungal, and parasitic causes, and was also associated with a worse prognosis.

The authors suggest that the inability to identify a causative agent in the majority of cases in the state-wide survey may be the result of a selection bias towards diagnostically challenging cases. As a physician-participant in this project, I beg to differ. One of the frustrations of caring for these patients is the wait time between presentation and the results of any "real-time" PCR or serologic testing, which for most community-based hospitals requires referral to a reference laboratory for many of the tests required. It is not uncommon to wait up to 2 weeks before the initial test results are available. And, by definition, the results of convalescent serologies, take even longer. Thus, in my experience, most patients with encephalitis present a "diagnostic challenge". This extensive survey demonstrates that a thoughtful and thorough approach to patients with encephalitis can provide an infectious etiology in 16% to 29% of patients, and occasionally yields some unusual diagnoses. It may be helpful to the families of these patients to understand that even with the most sophisticated and current techniques, the diagnosis in two-thirds of cases remains elusive.