Incidence, timing, and outcomes of tracheostomy in COVID-19 acute respiratory distress syndrome patients across three nations-an individual patient data analysis.

Background: Tracheostomy is often performed to facilitate weaning in invasively ventilated patients. There are studies regarding tracheostomy in acute respiratory distress syndrome (ARDS) patients, but its practice in ARDS patients due to coronavirus disease 2019 (COVID-19) remains uncertain. The aim of the study was to compare incidences of tracheostomy among three nations and to analyze outcomes associated with tracheostomy in COVID-19 ARDS patients. Methods: Post hoc analysis of patient-level data on tracheostomy in patients with COVID-19 ARDS from nationwide ventilation studies in Argentina, Spain, and the Netherlands. The primary endpoint was incidence and timing of tracheostomy. A propensity matched analysis was used to correct for factors with a known association with mortality, and a sensitivity analysis was performed to correct for the risk of death and the chance of receiving a tracheostomy. All three studies included patients that were admitted to a participating intensive care unit (ICU); aged >18 years; receiving ventilatory support; confirmed to have COVID-19 pneumonia. Patients were excluded from participation when there was an alternate cause for pneumonia. For the current analysis, we additionally excluded patients not having ARDS. Results: The analysis included 5,781 invasively ventilated patients: 1,469 (25%) patients from Argentina, 3,349 (58%) patients from Spain, and 963 (17%) patients from the Netherlands. Tracheostomies were performed 24% in Argentina {median 20 [16-24] days}, 40% in Spain {median 16 [12-21] days}, and 18% in the Netherlands {median 21 [17-27] days}. In unmatched and matched analyses 60-day mortality was lower in patients that received a tracheostomy. Conclusions: Both the incidence and timing of tracheostomy in COVID-19 ARDS patients differed among the three nations. Tracheostomy was associated with lower mortality rates.