“…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). Therefore, considering that the database included anamnestic/heteroanamnestic data on disease course before hospital admission, appropriate antibiotic timing relative to the onset of meningitis symptoms was assessed as "within 24 hours" or "later", based on agreement between two investigators unaware of the patients outcome and dexamethasone treatment.…”