Non-invasive assessment of inspiratory effort in patients undergoing non-invasive ventilation: The INSPIRE study

Objective: To evaluate whether occlusion maneuvers allow to estimate inspiratory effort, respiratory drive, lung stress and mechanics during noninvasive ventilation (NIV). Methods: In this multicenter study, sixty hypoxemic patients underwent two randomized 1-h NIV sessions (oro-nasal and full-face) after extubation. End-expiratory and end-inspiratory occlusions were performed to measure inspiratory airway-pressure deflection (Pocc), 100-ms airway-pressure drop (P0.1) and plateau pressure. Esophageal manometry, which was calibrated during invasive ventilation before extubation, was used as reference. To derive effort (ΔPes:K×Pocc) and dynamic transpulmonary driving pressure (ΔPL,dyn=[Pressure support-(K×Pocc)]), interface-specific conversion factors K=ΔPes/Pocc were computed in a derivation cohort (n=20), and validated in a separated cohort (n=40). Results: Pocc could be measured in all patients. In the derivation cohort, K was −0.71 (oro-nasal), and −0.80 (full-face). In the validation cohort, predicted ΔPes agreed with actual ΔPes (oro-nasal: bias −0.4 cmH2O [limits of agreement:-3.4 to +2.6]; full-face: −0.1 cmH2O [−2.9 to +3.1]); predicted ΔPL,dyn agreed with actual ΔPL,dyn (oro-nasal: 0.4 cmH2O [−2.9 to +3.7]; full-face: -0.06 cmH2O [−4.1 to +4.2]). Predicted ΔPes identified actual ΔPes ≥10 cmH2O with area under the curve of 0.96 (oro-nasal) and 0.94 (full-face). P0.1 did not reflect respiratory drive. Plateau pressure was unstable in 78% (oro-nasal) and 90% (full-face) of patients. Higher predicted ΔPes and ΔPL,dyn, and lower predicted compliance (tidal volume/ΔPL,dyn) were associated with subsequent reintubation. Conclusions: During NIV, Pocc allows to estimate effort and lung stress, whereas P0.1 and plateau pressure are unreliable.