Importance Information about the severity of Omicron is scarce. Objective To report the respective risk of ICU admission in patients hospitalized with Delta and Omicron variants and to compare the characteristics and disease severity of critically ill patients infected with both variants according to vaccination status. Design Analysis from the APHP database, called Reality, prospectively recording the following information in consecutive patients admitted in the ICU for COVID-19: age, sex, type of variant, immunosuppression, vaccination, pneumonia, need for invasive mechanical ventilation, time between symptom onset and ICU admission, and in-ICU mortality. Retrospective analysis on an administrative database, Systeme Information pour le Suivi des Victimes (SI-VIC), which lists hospitalized COVID-19 patients. Setting 39 hospitals in the Paris area from APHP group. Participants Patients hospitalized from December 1, 2021 to January 18, 2022 for COVID-19. Main outcomes and measures Risk of ICU admission was evaluated in 3761 patients and Omicron cases were compared to Delta cases in the ICU in 888 consecutive patients. Results On January 18, 45% of patients in the ICU and 63.8% of patients in conventional hospital units were infected with the Omicron variant (p < 0.001). The risk of ICU admission with Omicron was reduced by 64% than with Delta (9.3% versus 25.8% of cases, respectively, p < 0.001). In critically ill patients, 400 had the Delta variant, 229 the Omicron variant, 98 had an uninformative variant screening test and 161 did not have information on variant screening test. 747 patients (84.1%) were admitted for pneumonia. Compared to patients infected with Delta, Omicron patients were more vaccinated (p<0.001), even with 3 doses, more immunocompromised (p<0.001), less admitted for pneumonia (p<0.001), especially when vaccinated (62.1% in vaccinated versus 80.7% in unvaccinated, p<0.001), and less invasively ventilated (p=0.02). Similar results were found in the subgroup of pneumonia but Omicron cases were older. Unadjusted in-ICU mortality did not differ between Omicron and Delta cases, neither in the overall population (20.0% versus 27.9%, p = 0.08), nor in patients with pneumonia (31.6% versus 29.7%, respectively) where adjusted in-ICU mortality did not differ according to the variant (HR 1.43 95%CI [0.89;2.29], p=0.14). Conclusion and relevance Compared to the Delta variant, the Omicron variant is less likely to result in ICU admission and less likely to be associated with pneumonia. However, when patients with the Omicron variant are admitted for pneumonia, the severity seems similar to that of patients with the Delta variant, with more immunocompromised and vaccinated patients and no difference in adjusted in-ICU mortality. Further studies are needed to confirm our results.

Background Prone position is frequently used in patients with acute respiratory distress syndrome (ARDS), especially during the Coronavirus disease 2019 pandemic. Our study investigated the ability of pulse pressure variation (PPV) and its changes during a tidal volume challenge (TVC) to assess preload responsiveness in ARDS patients under prone position. Methods This was a prospective study conducted in a 25-bed intensive care unit at a university hospital. We included patients with ARDS under prone position, ventilated with 6 mL/kg tidal volume and monitored by a transpulmonary thermodilution device. We measured PPV and its changes during a TVC (ΔPPV TVC6–8) after increasing the tidal volume from 6 to 8 mL/kg for one minute. Changes in cardiac index (CI) during a Trendelenburg maneuver (ΔCITREND) and during end-expiratory occlusion (EEO) at 8 mL/kg tidal volume (ΔCI EEO8) were recorded. Preload responsiveness was defined by both ΔCITREND ≥ 8% and ΔCI EEO8 ≥ 5%. Preload unresponsiveness was defined by both ΔCITREND < 8% and ΔCI EEO8 < 5%. Results Eighty-four sets of measurements were analyzed in 58 patients. Before prone positioning, the ratio of partial pressure of arterial oxygen to fraction of inspired oxygen was 104 ± 27 mmHg. At the inclusion time, patients were under prone position for 11 (2–14) hours. Norepinephrine was administered in 83% of cases with a dose of 0.25 (0.15–0.42) µg/kg/min. The positive end-expiratory pressure was 14 (11–16) cmH2O. The driving pressure was 12 (10–17) cmH2O, and the respiratory system compliance was 32 (22–40) mL/cmH2O. Preload responsiveness was detected in 42 cases. An absolute change in PPV ≥ 3.5% during a TVC assessed preload responsiveness with an area under the receiver operating characteristics (AUROC) curve of 0.94 ± 0.03 (sensitivity: 98%, specificity: 86%) better than that of baseline PPV (0.85 ± 0.05; p = 0.047). In the 56 cases where baseline PPV was inconclusive (≥ 4% and < 11%), ΔPPV TVC6–8 ≥ 3.5% still enabled to reliably assess preload responsiveness (AUROC: 0.91 ± 0.05, sensitivity: 97%, specificity: 81%; p < 0.01 vs. baseline PPV). Conclusion In patients with ARDS under low tidal volume ventilation during prone position, the changes in PPV during a TVC can reliably assess preload responsiveness without the need for cardiac output measurements. Trial registration: ClinicalTrials.gov (NCT04457739). Registered 30 June 2020 —Retrospectively registered, https://clinicaltrials.gov/ct2/show/record/NCT04457739