Late Breaking Abstract - Effect of a conservative approach to the start of mechanical ventilation on ventilator-free days in coronavirus disease 2019 (COVID-19) pneumonia after adjustment by inverse probability of treatment weighting

ABSTRACT

Introduction: High flow nasal cannula (HFNC) decreases the need for intubation in patients with acute respiratory failure (ARF). However, delaying intubation has been linked to worse patients' outcome in the acute respiratory distress syndrome (ARDS). Methods: Prospectively once-daily collected database including patients admitted to 30 Spanish ICUs. On May 13th, we identified 390 patients where outcome data was complete. Patients were divided into 2 groups 1) 272 subjects who received mechanical ventilation (MV) at admission or early after ICU admission (before day 2 considering admission as day 0), and 2) 118 patients who did not receive MV in this time period. The probability of not being ventilated was estimated by means of multivariate logistic regression including age, sex, number of comorbidities, respiratory and non-respiratory SOFA score, minimal systolic blood pressure on days 1 and 2, maximal heart rate on days 1 and 2, respiratory rate and hospital (divided in 4 groups based on the number of cases). Treatment weighting was defined as the inverse of the probability for group 2 and as the inverse of 1 minus the probability for group 1, as described elsewhere. Ventilator-free days (VFD) at 28 days were compared between groups. Follow-up was continued until ICU discharge, death or 28 days after ICU admission. Death was computed as 0 VFDs. Results: VFD was 4.8 days longer in those not receiving early MV (95% CI 2.2 to 7.4). Conclusion: After adjusting for confounding, avoiding early MV increased VFD by 4.8 days in this group of COVID-19 patients.