as co-investigators (Centre Hospitalier d'Aix-en-Provence, France). Michael Aubourg, MSc, and Pierre Castel, MiM, provided English editing service. None of these individuals received compensation for their contributions.

We aimed to compare the effects of vitamin C, glucocorticoids, vitamin B1, combinations of these drugs, and placebo or usual care on longer-term mortality in adults with sepsis or septic shock. MEDLINE, Embase, CENTRAL, ClinicalTrials.gov and WHO-ICTRP were searched. The final search was carried out on September 3rd, 2021. Multiple reviewers independently selected randomized controlled trials (RCTs) comparing very-high-dose vitamin C (≥ 12 g/day), high-dose vitamin C (< 12, ≥ 6 g/day), vitamin C (< 6 g/day), glucocorticoid (< 400 mg/day of hydrocortisone), vitamin B1, combinations of these drugs, and placebo/usual care. We performed random-effects network meta-analysis and, where applicable, a random-effects component network meta-analysis. We used the Confidence in Network Meta-Analysis framework to assess the degree of treatment effect certainty. The primary outcome was longer-term mortality (90-days to 1-year). Secondary outcomes were severity of organ dysfunction over 72 h, time to cessation of vasopressor therapy, and length of stay in intensive care unit (ICU). Forty-three RCTs (10,257 patients) were eligible. There were no significant differences in longer-term mortality between treatments and placebo/usual care or between treatments (10 RCTs, 7,096 patients, moderate to very-low-certainty). We did not find any evidence that vitamin C or B1 affect organ dysfunction or ICU length of stay. Adding glucocorticoid to other treatments shortened duration of vasopressor therapy (incremental mean difference, − 29.8 h [95% CI − 44.1 to − 15.5]) and ICU stay (incremental mean difference, − 1.3 days [95% CI − 2.2 to − 0.3]). Metabolic resuscitation with vitamin C, glucocorticoids, vitamin B1, or combinations of these drugs was not significantly associated with a decrease in longer-term mortality. Supplementary Information The online version contains supplementary material available at 10.1007/s00134-021-06558-0.

Objective: Continuous positive airway pressure (CPAP) is an important therapeutic tool in COVID-19 acute respiratory distress syndrome (ARDS) since it improves oxygenation, reduces respiratory rate and can prevent intubation and intensive care unit (ICU) admission. CPAP during pronation has seldom been described and never during sedation. Design: Case series. Setting: High dependency unit of San Carlo University Hospital (Potenza, Italy). Patients: Eleven consecutive patients with COVID-19 ARDS. Intervention. Helmet CPAP in prone position after failing a CPAP trial in the supine position. Main variable of interest: Data collection at baseline and then after 24, 48 and 72 hours of pronation. We measured PaO2/FIO2, pH, lactate, PaCO2, SpO2, respiratory rate and the status of the patients at 28-day follow up. Results: Patients were treated with helmet CPAP for a mean +SD of 7+2.7 days. Prone positioning was feasible in all patients, but in 7 of them dexmedetomidine improved comfort. PaO2/FIO2 improved from 107.5±20.8 before starting pronation to 244.4±106.2 after 72 hours (p<.001). We also observed a significantly increase in Sp02 from 90.6±2.3 to 96±3.1 (p<.001) and a decrease in respiratory rate from 27.6±4.3 to 20.1±4.7 (p=.004). No difference was observed in PaCO2 or pH. At 28 days two patients died after ICU admission, one was discharged in the main ward after ICU admission and eight were discharged home after being successfully managed outside the ICU. Conclusions: Helmet CPAP during pronation was feasible and safe in COVID-19 ARDS managed outside the ICU and sedation with dexmedetomidine safely improved comfort. We recorded an increase in PaO2/FIO2, SpO2 and a reduction in respiratory rate.