Causes of encephalitis and differences in their clinical presentations in England: a multicentre, population-based prospective study

Background Encephalitis has many causes, but for most patients the cause is unknown. We aimed to establish the cause and identify the clinical differences between causes in patients with encephalitis in England. Methods Patients of all ages and with symptoms suggestive of encephalitis were actively recruited for 2 years (staged start between October, 2005, and November, 2006) from 24 hospitals by clinical staff. Systematic laboratory testing included PCR and antibody assays for all commonly recognised causes of infectious encephalitis, investigation for less commonly recognised causes in immunocompromised patients, and testing for travel-related causes if indicated. We also tested for non-infectious causes for acute encephalitis including autoimmunity. A multidisciplinary expert team reviewed clinical presentation and hospital tests and directed further investigations. Patients were followed up for 6 months after discharge from hospital. Findings We identified 203 patients with encephalitis. Median age was 30 years (range 0-87). 86 patients (42%, 95% CI 35-49) had infectious causes, induding 38 (19%, 14-25) herpes simplex virus, ten (5%, 2-9) varicella zoster virus, and ten (5%, 2-9) Mycobacterium tuberculosis; 75 (37%, 30-44) had unknown causes. 42 patients (21%, 15-27) had acute immune-mediated encephalitis. 24 patients (12%, 8-17) died, with higher case fatality for infections from M tuberculosis (three patients; 30%, 7-65) and varicella zoster virus (two patients; 20%, 2-56). The 16 patients with antibody-associated encephalitis had the worst outcome of all groups nine (56%, 30-80) either died or had severe disabilities. Patients who died were more likely to be immunocompromised than were those who survived (OR=3.44). Interpretation Early diagnosis of encephalitis is crucial to ensure that the right treatment is given on time. Extensive testing substantially reduced the proportion with unknown cause, but the proportion of cases with unknown cause was higher than that for any specific identified cause.