Abstract
BACKGROUND: During control mechanical ventilation (CMV), the driving pressure of the respiratory system (ΔP(rs)) serves as a surrogate of transpulmonary driving pressure (ΔP(lung)). Expiratory muscle activity that decreases end-expiratory lung volume may impair the validity of ΔP(rs) to reflect ΔP(lung). This prospective observational study in patients with acute respiratory distress syndrome (ARDS) ventilated with proportional assist ventilation (PAV+), aimed to investigate: (1) the prevalence of elevated ΔP(lung), (2) the ΔP(rs)-ΔP(lung) relationship, and (3) whether dynamic transpulmonary pressure (Plung(sw)) and effort indices (transdiaphragmatic and respiratory muscle pressure swings) remain within safe limits. METHODS: Thirty-one patients instrumented with esophageal and gastric catheters (n = 22) were switched from CMV to PAV+ and respiratory variables were recorded, over a maximum of 24 h. To decrease the contribution of random breaths with irregular characteristics, a 7-breath moving average technique was applied. In each patient, measurements were also analyzed per deciles of increasing lung elastance (E(lung)). Patients were divided into Group A, if end-inspiratory transpulmonary pressure (P(LEI)) increased as E(lung) increased, and Group B, which showed a decrease or no change in P(LEI) with E(lung) increase. RESULTS: In 44,836 occluded breaths, ΔP(lung) ≥ 12 cmH(2)O was infrequently observed [0.0% (0.0-16.9%) of measurements]. End-expiratory lung volume decrease, due to active expiration, was associated with underestimation of ΔP(lung) by ΔP(rs), as suggested by a negative linear relationship between transpulmonary pressure at end-expiration (P(LEE)) and ΔP(lung)/ΔP(rs). Group A included 17 and Group B 14 patients. As E(lung) increased, ΔP(lung) increased mainly due to P(LEI) increase in Group A, and P(LEE) decrease in Group B. Although ΔP(rs) had an area receiver operating characteristic curve (AUC) of 0.87 (95% confidence intervals 0.82-0.92, P < 0.001) for ΔP(lung) ≥ 12 cmH(2)O, this was due exclusively to Group A [0.91 (0.86-0.95), P < 0.001]. In Group B, ΔP(rs) showed no predictive capacity for detecting ΔP(lung) ≥ 12 cmH(2)O [0.65 (0.52-0.78), P > 0.05]. Most of the time Plung(sw) and effort indices remained within safe range. CONCLUSION: In patients with ARDS ventilated with PAV+, injurious tidal lung stress and effort were infrequent. In the presence of expiratory muscle activity, ΔP(rs) underestimated ΔP(lung). This phenomenon limits the usefulness of ΔP(rs) as a surrogate of tidal lung stress, regardless of the mode of support.