Encephalitis: Why We Need to Keep Pushing the Envelope

Encephalitis is one of the most challenging syndromes for physicians to manage. Disease onset is acute, symptoms progress rapidly, and previously healthy individuals become quickly, and possibly permanently, disabled. One of the difficulties in both managing and performing research on encephalitis is the vast number of causative agents [1]. Infectious causes are diverse, including viruses, bacteria, fungi, and parasites, and the microbiology of encephalitis represents a moving target. Historically important causes of encephalitis, such as measles, mumps, and rubella, have decreased, largely as a result of immunization campaigns [2]. At the same time, agents such as West Nile virus [3], enterovirus 71 [4], Australian bat Lyssavirus [5], and Nipah virus [6] have emerged as new or, in some cases, newly appreciated causes of encephalitis. Regardless of the cause, the nonspecific clinical presentation of encephalitis hinders diagnosis. Clinical findings, such as fever, headache, and altered level of consciousness, are common but fail to distinguish among various pathogens and can be seen with a variety of noninfections conditions that mimic encephalitis, such as autoimmune disease or vasculitis. Further compounding the diagnostic dilemma is determining the relevance of an infectious agent detected outside the central nervous system, especially for agents with unproven neurotropism. For instance, several large case series have identified Mycoplasma pneumoniae as the most commonly diagnosed infection among children with encephalitis [7, 8]. However, even when IgM antibody to Mycoplasma is present, the significance of this finding remains unclear given the high background incidence of acute infection in the pediatric population and the limitations of serologic testing [7–9]. Other organisms associated with encephalitis but with similar difficulties in establishing causality include influenza virus [10], rotavirus [11], human herpesvirus 6 [12], and parvovirus [13]. The article by Mailles et al [14] in the current issue of Clinical Infectious Diseases is an important step forward in understanding certain aspects of encephalitis. Mailles and colleagues conducted a 1-year prospective study to determine the microbiologic spectrum of encephalitis in France and evaluate the utility of the newly published Société de Pathologie Infectieuse de Langue Française (French Society of Infectious Diseases) encephalitis diagnostic guidelines [15]. Risk factors associated with death were also examined. The study identified an infectious cause in slightly >50% of cases; herpes simplex virus (HSV) and varicella-zoster virus were the most commonly identified infections, similar to other studies. A unique finding in this series was the unexpectedly high number of patients diagnosed as having Mycobacterium tuberculosis and Listeria monocytogenes . The authors secondarily excluded patients with a noninfectious cause of disease from their study cohort. These noninfectious entities are enumerated in Table 2 of the article by Mailles and colleagues and should remind physicians of the spectrum of noninfectious entities that mimic infectious encephalitis, many of which are treatable [16]. Ten percent of patients died during hospitalization. Not surprisingly, older age and immunosuppression were associated with higher mortality rates. Clinical severity at admission did not predict outcome, but patients who required mechanical ventilation and/or were comatose at day 5 after admission were at higher risk for death. The study alludes to the societal burden of encephalitis among survivors, a quarter of whom were discharged to skilled nursing facilities or to the care of relatives. Should the findings in this study influence physician behavior regarding testing and treatment of encephalitis? Although the relatively high proportion of cases with an identified infectious cause is laudable, several factors affect the generalizability of these findings. First, the study recruited patients from hospitals throughout the country, but it was not population based, and referral biases may have skewed the results. In support of this, pediatric patients comprised only 10% of the study population. Second, the authors excluded survivors hospitalized for <5 days because these patients were assumed to have aseptic meningitis rather than encephalitis. However, the duration of illness with encephalitis is highly variable, and the small decrease in misclassification bias may have been offset by a more significant exclusion bias. Specifically, patients with encephalitis due to enteroviruses and Bartonella species may have a fulminant presentation but often recover fully within several days after onset [17–19]. Not surprisingly, these diagnoses were underrepresented in this series. The exclusion of patients hospitalized for ⩾5 days also artificially inflated the median length of encephalitis hospitalization. An unexpected finding was the frequency with which HSV-1 was identified as a confirmed cause of encephalitis. Herpes simplex encephalitis accounted for 42% of cases in which a pathogen was identified and 22% of total cases, higher than what has been reported in other studies [20–22]. Whether this difference is due to variable inclusion or exclusion criteria, differences in age groups studied, methods of HSV detection, or differing diagnostic classification is unknown. Interestingly, a study from England found that herpes simplex encephalitis accounted for 59% of viral causes [23]. One possible explanation is that the European population may have a genetic polymorphism that increases their susceptibility to herpes simplex encephalitis, as has recently been described [24]. Alternatively, inhibitors of tumor necrosis factor-α activity have recently been associated with herpes simplex encephalitis [21], and the larger than expected number of cases in recent series may reflect increasing use of these medications [25]. A key finding of the present study is the need to consider bacterial pathogens as significant causes of encephalitis. M. tuberculosis and L. monocytogenes accounted for one-quarter of all encephalitis diagnoses during the study period. Despite the fact that both of these pathogens are treatable with prompt antibiotic therapy, infection with M. tuberculosis or L. monocytogenes accounted for almost one-half of all encephalitis-related deaths. Although the authors do not specify whether the excess mortality was due to a delay in initiating treatment, on the basis of the excess mortality with these agents it would be appropriate for French physicians to strongly consider initiation of empirical therapy for both M. tuberculosis and L. monocytogenes in patients presenting with meningoencephalitis. Mailles et al [14] prove that regional guidelines play an important role in optimizing diagnostic testing, identifying pathogens, and guiding empirical treatment in French patients with encephalitis. Whether these findings are generalizable to other populations can only be determined through similar large-scale, comprehensive studies to determine the regional microbiology of encephalitis. Because the microbiology of encephalitis is not static, the findings from this study period may not be reproducible for the same population in future years. As the authors point out, the finding of a large number of cases of M. tuberculosis and L. monocytogenes encephalitis during the study period likely reflects an unusually high background incidence of these infections. Only through ongoing prospective surveillance will it be apparent whether these agents will remain significant causes of encephalitis or whether they will be replaced by other emerging pathogens. Finally, it is striking that even using a rigorous case definition for encephalitis and with comprehensive testing the origin remained enigmatic in ∼50% of cases in this study. The high proportion of undiagnosed cases in this and other studies suggests a failure of our current approach to encephalitis. Diagnostic shortcomings include absence of tests for as yet undiscovered infectious agents, failure to perform testing for uncommon pathogens, and poor sensitivity of currently available diagnostic tools for well-associated pathogens. Undiscovered noninfectious entities also likely contribute to the pool of undiagnosed cases. For instance, several patients within the California encephalitis project have been identified with a newly discovered paraneoplastic entity, also known as anti-N -methyl-d-receptor encephalitis [26, 27]. Given the significant personal, financial, and societal burden of encephalitis, regional studies that seek to address these gaps in our understanding are urgently needed to improve the diagnosis and, ultimately, the prognosis of these devastating infections. Potential conflicts of interest. C.G.: no conflicts.