Impact of direct ICU admission of pneumococcal meningitis in France: a retrospective analysis of a French medico-administrative (PMSI) database

Abstract

Abstract Background Current guidelines for adult patients with pneumococcal meningitis (PM) recommend initial management in intermediate or intensive care units (ICU), but evidence to support these recommendations is limited. We aimed to describe ICU admission practices of patients with PM. Methods We conducted a retrospective analysis of the French medico administrative database of consecutive adult patients with PM and sepsis criteria hospitalized between 2011 and 2020. We defined two groups, “Direct ICU” corresponding to a direct ICU admission and “Delayed ICU” corresponding to a secondary ICU admission. Results We identified 4052 patients hospitalized for a first episode of PM, including 2006 “Direct ICU” patients (50%) and 2046 “delayed ICU” patients (50%). The patients were mainly males [n = 2260 (56%)] with median age of 61 years [IQR 50–71] and a median Charlson index of 1 [0–3]. Among them, median SAPS II on admission was 46 [33–62], 2173 (54%) had a neurological failure on admission with 2133 (53%) in coma, 654 (16%) with brainstem failure, 488 (12%) with seizures and 779 (19%) with focal signs without coma. PM was frequently associated with pneumonia [n = 1411 (35%)], and less frequently with endocarditis [n = 317 (8%)]. The median ICU length of stay and hospital length of stay were 6 days [2–14] and 21 days [13–38], respectively. In-hospital mortality was 27% (n = 1100) and 640 (16%) patients were secondarily transferred to rehabilitation care unit. Direct ICU group was significantly more severe but after adjustment for age, sex, comorbidities, organ failures on admission and admission from home, direct ICU admission was significantly associated with a lower mortality (Odds ratio 0.67 [0.56–0.80], p < 0.01). This corresponded to one death avoided for 11 PM directly admitted in ICU. Conclusions Among patients with PM and sepsis, direct ICU admission was associated with lower mortality rates when compared to delayed admission.