JOURNAL ARTICLE
A Clear and Present Danger: Encephalitis Due to Infection With Normocellular Cerebrospinal Fluid
Karen Bloch, Carol Glaser
Clinical Infectious Diseases, Volume 81, Issue 1, 15 July 2025, Pages 188–189, https://doi.org/10.1093/cid/ciae387
Can encephalitis due to an infectious cause be present despite normocellular cerebral spinal fluid (CSF)? In the article “Absence of cerebrospinal fluid pleocytosis in encephalitis” by Habis et al, more than a quarter of patients with acute encephalitis lacked CSF pleocytosis, making the answer an emphatic “Yes!” This study, which included almost 600 adult patients hospitalized at 2 geographically diverse urban settings between 2005 and 2023 reported an absence of pleocytosis in 19% of encephalitis cases due to an infectious cause, 39% of autoimmune encephalitis cases, and 26% of encephalitis cases without an identified cause. Most strikingly, almost a quarter of patients with herpes simplex virus 1 (HSV-1) encephalitis (HSE) had normocellular CSF.
Perhaps this finding is not surprising when considering the pathophysiology of central nervous system infections. Encephalitis denotes inflammation of brain parenchyma, whereas meningitis is inflammation of the meninges. In many cases, an overlap syndrome (meningoencephalitis) is present. However, when infection is restricted to the brain, the CSF characteristics may not reflect compartmentalized infection. Diagnostic criteria developed by the International Encephalitis Consortium and widely adapted for clinical use includes CSF pleocytosis as only 1 of 6 minor criteria [1], underscoring that this is an imperfect surrogate marker for brain infection. The absence of pleocytosis has been well-documented for encephalitis due to West Nile virus [2], rabies virus [3], tickborne encephalitis virus [4], and, increasingly, HSV-1.
Recognition that encephalitis due to an infectious cause may be present with a normal CSF white blood cell (WBC) count has important implications for the management of this notoriously challenging syndrome. Cost and concern for inappropriate test utilization has led some health systems to restrict the use of multiplex molecular assays for meningoencephalitis to patients with CSF pleocytosis [5–7]. Not only is this an artificial dichotomy (is 5 WBC/mm3 truly different than 4 WBC/mm3?), the study by Habis et al challenges the wisdom of this approach to diagnostic stewardship. Limiting testing to only those with ≥5 WBC/mm3, almost a quarter of HSE cases would have gone undiagnosed and, more importantly, untreated. For a disease such as HSE, where the mortality of untreated infection is >70% and treatment delays are associated with poor neurologic outcomes, this could be a catastrophic oversight.
As the authors discuss, normocellular CSF has been reported with HSE early in infection and in specific populations: at the extremes of age, in immunocompromised patients, and in those receiving whole-brain radiation therapy. Based on the current study, this phenomenon is likely more common than has been previously appreciated. A contemporary study of HSE patients conducted between 2007–2017 at 47 intensive care units in France reported 16.5% had a CSF WBC of <5/mm3, and among the subset with an initially negative HSV-1 polymerase chain reaction (PCR) in CSF, this rose to an astounding 60% of cases [8]. This begs the question of whether the converse is true as well: Are patients with HSE and pauci-cellular CSF more likely to have false negative CSF PCR testing on the initial sample?
Although the current study does not address this question, false negative HSV PCR tests are well documented in HSE, particularly early in infection [9]. This has led to societal guidelines recommending that acyclovir be continued and repeat testing be performed on a subsequent CSF sample when suspicion for this diagnosis is high [1, 10]. Often, this is interpreted as visualization of a temporal lobe abnormality on neuroimaging, bolstered by the presence of pleocytosis and fever. What then to make of the finding in the study by Habis et al that almost 20% of HSE patients had a normal MRI scan, rising to 40% among the subset without pleocytosis? Or the finding that only 44% of HSE patients without pleocytosis had fever? In the absence of these classic signs, risk stratification for HSE is not possible. Thus, based on the paper by Habis et al, not only should HSV-1 (standalone or multiplex) PCR be performed on CSF regardless of WBC count, repeating testing on a second CSF sample collected 3–7 days later should be considered despite an initial negative test, even when classic findings like fever or temporal lobe focality are absent.
Although the primary focus of the study was the frequency of normocellular CSF with encephalitis, several other features deserve mention. Among cases due to an infectious cause, 10% were diagnosed with a pathogen more commonly associated with community-acquired meningitis (S. pneumoniae, H. influenzae, S. agalactiae, N meningitidis) or healthcare-associated meningitis (S. aureus, Enterococcus) rather than encephalitis. This highlights the challenges of differentiation of meningitis from encephalitis at the time of presentation. The inability to discriminate between these syndromes clinically underscores the importance of multiplex molecular testing to allow rapid diagnosis and optimize antimicrobial use. More importantly, this also points out the necessity of continuing empiric antibiotic treatment for bacterial meningitis until an alternative pathogen is identified or cultures are negative at >72 hours, as, similar to HSE, failure to appropriately treat bacterial meningitis is associated with adverse outcomes including death.
A further striking finding of the study was the frequency and diversity of causes of autoimmune encephalitis. More than 16% of encephalitis cases were due to autoimmune causes, and slightly less than half (45%) of these were diagnosed with anti-N-methyl-D-aspartate receptor (NMDAR)encephalitis. The relatively frequent occurrence of anti-NMDAR encephalitis coupled with the need for immunosuppressive therapy for this disease underscores contemporary recommendations for upfront testing for anti-NMDAR antibody in serum and CSF in all patients with encephalitis, or at least in those with an initially negative infectious evaluation [11]. However, limiting testing to just anti-NMDAR encephalitis would miss many other autoimmune diagnoses. Therefore, a full autoimmune panel, which is available through several commercial laboratories, is preferrable. And when the diagnosis remains elusive, neurologic consultation is indicated to determine additional testing and the role of empiric immunosuppression.
A final point bears emphasis, namely that 42% of cases presenting with encephalitis remain undiagnosed. The high proportion of undiagnosed cases in this and other studies suggests a failure of our current approach to encephalitis. If the preceding sentence sounds familiar, this is because it was copied without modification from an editorial we wrote in 2009 [12] in response to a study where 50% of encephalitis cases remained undiagnosed [13]. It is difficult to say that significant progress has been made over the last 15 years when the proportion of diagnosed cases has risen by a scant 8%, particularly when this earlier study excluded cases of autoimmune encephalitis. The microbiology of encephalitis in the study by Habis et al reflects pathogens well known to be associated with meningoencephalitis rather than newly discovered or emerging organisms, despite diagnostic advances such as next generation sequencing. This raises the troubling idea that the inability to find a cause in some cases may be due to a lower sensitivity of molecular testing with acellular CSF or a failure to adequately search for an infectious cause of encephalitis in the absence of pleocytosis.
We stand by the comment made in the 2009 editorial [12]: “We need to keep pushing the envelope” to improve the diagnosis and prognosis of patients with encephalitis. The study by Habis et al suggests that one way to do this is to aggressively pursue diagnostic evaluation on all patients, including those who lack pleocytosis but fulfill other criteria for encephalitis. Further research to guide optimal management of encephalitis in patients with normocellular CSF is a priority, with the potential to save lives and improve patient outcomes.
