Increased Driving Pressure During Assisted Ventilation for Hypoxemic Respiratory Failure Is Associated with Lower ICU Survival: The ICEBERG Study

Abstract

<b>Rationale:</b> Driving pressure (ΔP) is a marker of severity and a possible target for lung protection during controlled ventilation, but its value during assisted ventilation is unknown. Inspiratory holds provide an estimate of ΔP that is quasi-static. Expiratory holds provide an estimate of the inspiratory effort, which is useful to estimate the transpulmonary dynamic ΔP. <b>Objectives:</b> To assess the correlation between ΔPs measured during assisted ventilation and ICU outcomes. <b>Methods:</b> Multicenter prospective observational study. Patients with acute hypoxemic respiratory failure were enrolled within 48 hours of triggering the ventilator. Respiratory mechanics were measured daily, and the variables of interest were averaged over the first 3 days of partial assistance. ICU outcomes were collected until Day 90. <b>Measurements and Main Results:</b> A total of 298 patients from 16 centers were enrolled. Vt, peak airway pressure, positive end-expiratory pressure, and inspiratory effort during the first 3 days of assisted ventilation did not differ between survivors and nonsurvivors. Quasi-static ΔP and transpulmonary dynamic ΔP were higher in nonsurvivors than in survivors (13 [IQR, 11-14] vs. 11 [9-13] cm H₂O [<i>P</i> < 0.001]; and 19 [16-23] vs. 16 [13-18] cm H₂O [<i>P</i> < 0.001], respectively), whereas compliance normalized to predicted body weight was lower (0.65 [0.54-0.84] vs. 0.79 [0.64-0.97] mL/cm H₂O/kg; <i>P</i> < 0.001). Multivariable analysis confirmed the association with outcomes. During the study days, static ΔP significantly diverged between survivors and nonsurvivors. <b>Conclusions:</b> During assisted ventilation, ΔP and normalized compliance are associated with ICU outcome, despite some overlap. Although our study does not allow an estimation of whether ΔP is a marker of severity or a cause of lung injury, it highlights the potential value of monitoring and targeting it during spontaneous assisted breathing.