Gram stains ought to be part of any workup for bacterial or aseptic meningitis, which apparently has not been consistently applied in our institution in the past. False-negative CSF cultures are not uncommon [37] and a diagnosis of bacterial meningitis should not be ruled out in the absence of gram stain data [15], [17], [38] and [39]. Had gram stain data been available in all cases in this study, 39 additional cases could have met the BC criteria for ASM and the rates of agreement would have been: buy Veliparib OPA = 85%, PPA= 89%, and NPA = 77%. Second, as stated in
the BC case definition document for aseptic meningitis, “an upper reference AZD6738 value for pleocytosis is not used as a criterion in the case definition to distinguish bacterial from aseptic meningitis because pleocytosis of several thousand leukocytes/μl of CSF has been described in patients with aseptic meningitis of confirmed viral etiology [7] and [40].” Based
on purulent CSF samples, several cases in the reported study were labeled as “bacterial meningitis” in the discharge summary, even though gram stain and culture results remained negative. The differential diagnosis of aseptic meningitis should always be considered, even if CSF cell counts are highly elevated [37] and [41]. Third, encephalitis was underrecognized in the discharge diagnoses whenever a concomitant diagnosis of aseptic meningitis seemed to “fit”. Encephalitis, however, is often associated with concomitant meningitis but the prognosis worsens considerably with the presence of parenchymal infection [42]. Therefore, the Brighton Collaboration Aseptic Meningitis and Encephalitis
Working Groups recommended that “aseptic meningitis should be reported only for cases in which meningeal inflammation is present in the absence of clinical or diagnostic features of encephalitis [7] and [8].” Overlapping cases should be listed as “(meningo-)encephalitis”. The limited case numbers in this study for encephalitis, myelitis, and ADEM, however, allow only limited conclusions. Additional evaluation studies are needed for these much BC case definitions. The design of the reported study also shows several strengths: the study used a closed system with a standardized tool for the diagnosis of complex medical entities. Several approaches (ICD-10 search and electronic search of discharge summaries by pre-defined terms) were used to identify cases consistently representing the clinical assessment as accurately as possible. The investigator was independent from the clinical care of the patients and blinded to the discharge diagnoses during the data entry and case evaluation process.