Transpulmonary Pressure as a Predictor of Successful Lung Recruitment: Reanalysis of a Multicenter International Randomized Clinical Trial

Abstract

Background: Recruitment maneuvers are used in patients with ARDS to enhance oxygenation and lung mechanics. Heterogeneous lung and chest-wall mechanics lead to unpredictable transpulmonary pressures and could impact recruitment maneuver success. Tailoring care based on individualized transpulmonary pressure might optimize recruitment, preventing overdistention. This study aimed to identify the optimal transpulmonary pressure for effective recruitment and to explore its association with baseline characteristics. Methods: We performed post hoc analysis on the Esophageal Pressure Guided Ventilation (EpVent2) trial. We estimated the dose-response relationship between end-recruitment end-inspiratory transpulmonary pressure and the change in lung elastance after a recruitment maneuver by using logistic regression weighted by a generalized propensity score. A positive change in lung elastance was indicative of overdistention. We examined how patient characteristics, disease severity markers, and respiratory parameters predict transpulmonary pressure by using multivariate linear regression models and dominance analyses. Results: Of 121 subjects, 43.8% had a positive change in lung elastance. Subjects with a positive change in lung elastance had a mean ± SD transpulmonary pressure of 15.1 ± 4.9 cm H₂O, compared with 13.9 ± 3.9 cm H₂O in those with a negative change in lung elastance. Higher transpulmonary pressure was associated with increased probability of a positive change in lung elastance (adjusted odds ratio 1.35 per 1 cm H₂O of transpulmonary pressure, 95% CI 1.13-1.61; P = .001), which indicated an S-shaped dose-response curve, with overdistention probability > 50% at transpulmonary pressure values > 18.3 cm H₂O. The volume of recruitment was transpulmonary pressure-dependent (P < .001; R² = 0.49) and inversely related to a change in lung elastance after adjusting for baseline lung elastance (P < .001; R² = 0.43). Negative correlations were observed between transpulmonary pressure and body mass index, PEEP, Sequential Organ Failure Assessment score, and P_aO2/F_IO2, whereas baseline lung elastance showed a positive correlation. The body mass index emerged as the dominant negative predictor of transpulmonary pressure (ranking 1; contribution to R² = 0.08), whereas pre-recruitment elastance was the sole positive predictor (contribution to R² = 0.06). Conclusions: Higher end-recruitment transpulmonary pressure increases the volume of recruitment but raises the risk of overdistention, providing the rationale for transpulmonary pressure to be used as a clinical target. Predictors, for example, body mass index, could guide recruitment maneuver individualization to balance adequate volume gain with overdistention.

Trial registration: ClinicalTrials.gov NCT01681225.

Keywords: Acute Respiratory Distress Syndrome; esophageal pressure; lung overdistension; recruitment maneuver; transpulmonary pressure; volume of recruitment.

Fig. 1.

The variability of subjects in chest-wall mechanics. The distribution of 121 subjects with an esophageal balloon, illustrating airway and end-recruitment transpulmonary pressures during a standard recruitment maneuver that demonstrates the relative contribution of the transpulmonary pressure to the total measured airway recruitment pressure. Transpulmonary pressure was calculated as the airway pressure minus the esophageal pressure.

Fig. 2.

Dose-response association of end-recruitment end-inspiratory transpulmonary pressure and lung elastance change after recruitment maneuver (ΔEL). A positive change in lung elastance represents lung overdistention. Analysis was done through a generalized propensity score–weighted model with logistic link function model. The wide confidence intervals at low recruitment transpulmonary pressure are secondary to few subjects in this range.

Fig. 3.

Dominance analysis, depicting the individual contribution.

Fig. 4.

Locally weighted scatterplot smoothing (LOWESS) model showing significant association between transpulmonary pressure (P_L) and lung elastance change (ΔEL). Transpulmonary pressures > 19.3 cm H₂O were suggested to result in a positive change in lung elastance and lung overdistention.