(1166) Lung Recruitability for the Effect of Spontaneous Breathing in Patients at the Early Phase of ARDS

Introduction: Increasing evidence shows that spontaneous breathing (SB) creates a pendelluft effect (internal redistribution of gas) and can worsen lung injury by causing local overstretch and substantial tidal recruitment in the dependent lung, known as patient self-inflicted lung injury (P-SILI). However, intensivists worldwide have divided opinions on whether SB should be allowed during mechanical ventilation (MV) in patients with moderate to severe ARDS due to inconsistent findings in the limited clinical data, likely related to the heterogeneity of SB effects. The aim of this study is to test the hypothesis that, given the presence of collapsed but recruitable lung units as a prerequisite for the occurrence of P-SILI, the effects of SB depend on lung potential recruitment (LPR), with more substantial effects in patients with greater recruitability.

Methods: In this prospective observational study at two university hospitals, patients on invasive MV for early moderate or severe ARDS underwent whole-lung CT scans at 5 and 45 cm H2O after enrollment to assess the LPR, which was then dichotomized by the median value (10%) to define high or low recruiters. Variables including SB activity in dichotomized form and driving pressure (Paw) were recorded throughout the first 48 hours after enrollment. The primary outcome was 60-day mortality.

Results: Of 155 patients available for analysis, 60 (38.7%) died by Day 60. In a multivariate Cox regression model, after adjusting for age, Paw and Sequential Organ Failure Assessment score, SB during the first 2 days was not independently associated with 60-day mortality (hazard ratio [HR], 1.37; 95% confidence interval [CI], 0.82-2.29; P=0.22). However, the SB effect on 60-day mortality differed by dichotomized LPR (P = 0.02 for interaction). Preserving SB was associated with higher mortality among high recruiters (HR, 2.44; 95% CI, 1.19-5.0; P=0.02), but not among low recruiters (HR, 0.64; 95% CI, 0.29-1.44; P=0.27).

Conclusions: We found evidence of a differential response to the ventilation strategy allowing SB across two ARDS subphenotypes, suggesting possible harm in preserving SB during the first 2 days only in one subphenotype of moderate/severe ARDS patients with greater lung recruitability. These findings warrant prospective testing in a future clinical trial.