Résumé
The COVID-19 pandemic abruptly increased the inflow of patients requiring intensive care units (ICU). French health institutions responded by a twofold capacity increase with temporary upgraded beds, supplemental beds in pre-existing ICUs, or newly created units (New-ICU). We aimed to compare outcomes according to admission in expert pre-existing ICUs or in New-ICU. Methods: this multicenter retrospective observational study was conducted in two 20-bed expert ICUs of an University Hospital (Expert-ICU) and in one 16-bed New-ICU in a private clinic managed respectively by 3 and 2 physicians during daytime and by one physician during the night shift. All consecutive adult patients with Covid-19 related acute hypoxemic respiratory failure admitted after centralized regional management by a dedicated crisis cell were included. Primary outcome was 180-day mortality. Propensity score matching and restricted cubic spline for predicted mortality over time were performed. Results: during the study period, 165 and 176 patients were enrolled in Expert-ICU and New-ICU respectively, 162 (98%) and 157 (89%) patients were analyzed. The unadjusted 180-day mortality was 30.8% in Expert-ICU and 28.7% in New-ICU, (log-rank test, p= 0.7). After propensity score matching, 123 pairs (76 and 78%) of patients were matched, with no significant difference in mortality (32% vs 32%, OR 1.00 [0.89; 1.12], p= 1). Adjusted predicted mortality decreased over time (p<0.01) in both Expert-ICU and New-ICU. Conclusions: in COVID-19 patients with acute hypoxemic respiratory failure, hospitalization in a new ICU was not associated with mortality at day 180. Our results suggest the feasibility of providing safe and effective care for COVID-19 patients when workload was fairly shared and staff training was well ensured in a New-ICU.