“…The contrast between the two clusters was also estimated as a measure of the effect of "CSF inflammation" on the outcomes. The following further independent variables were considered in multivariate analysis: age; sex; time (days) elapsed since first symptoms (any) to admission; use of antibiotics (any) before the diagnosis of meningitis; presence of serious comorbidity [includes malignancy, immunodeficiency (immunosuppressants, human immunodeficiency virus infection or splenectomy), diabetes mellitus (DM), other endocrinological diseases, alcohol abuse and liver cirrhosis, other chronic organ diseases (lungs, heart, kidney, liver)]; presence of focal neurological symptoms on admission (includes aphasia, cranial nerve palsy, monoparesis or hemiparesis); leukocyte count on admission; pathophysiological mechanism of the disease (e.g., meningitis following septicemia, or following middle ear infection or trauma; dichotomized as "following septicemia" and "other"); microbiologically verified BM (considered as yes/no, and also as pneumococcal/other bacterial/probable); worst Glasgow Coma Score (GCS) within 24 hours since admission as a continuous variable and also categorized into levels of consciousness disturbance as: none (GCS ≥15), mild (GCS [13][14], moderate (GCS 10-12 ) or severe (GCS ≤9); and timing of the appropriate antibiotic treatment (empirical as per in-house guidelines, or bacteriologically targeted, see above) commencement specifically in relation to the onset of consciousness disturbance and/or overt meningitis symptoms (e.g., fever, headache, vomiting, malaise) (16). Namely, although the "door-to-antibiotic" delay negatively affects the outcomes in community-acquired adult BM (particularly if > 2hours), timing of the appropriate antibiotic treatment relative to the onset of consciousness disturbance and/or other specific meningitis symptoms appears to be a particularly relevant predictor of the disease outcome (17,18).…”