Severity of hypoxemia and effect of high-frequency oscillatory ventilation in acute respiratory distress syndrome

Rationale: High-frequency oscillatory ventilation (HFOV) is theoretically beneficial for lung protection, but the results of clinical trials are inconsistent, with study-level meta-analyses suggesting no significant effect on mortality. Objectives: The aim of this individual patient data meta-analysis was to identify acute respiratory distress syndrome (ARDS) patient subgroups with differential outcomes from HFOV. Methods: After a comprehensive search for trials, two reviewers independently identified randomized trials comparing HFOV with conventional ventilation for adults with ARDS. Prespecified effect modifiers were tested using multivariable hierarchical logistic regression models, adjusting for important prognostic factors and clustering effects. Measurements and Main Results: Data from 1,552 patients in four trials were analyzed, applying uniform definitions for study variables and outcomes. Patients had a mean baseline Pa_O2/FI_O2 of 114 - 39 mm Hg; 40% had severe ARDS (Pa_O2/FI_O2 ,100 mm Hg). Mortality at 30 days was 321 of 785 (40.9%) for HFOV patients versus 288 of 767 (37.6%) for control subjects (adjusted odds ratio, 1.17; 95% confidence interval, 0.94–1.46; P = 0.16). This treatment effect varied, however, depending on baseline severity of hypoxemia (P = 0.0003), with harm increasing with Pa_O2/FI_O2 among patients with mild-moderate ARDS, and the possibility of decreased mortality in patients with very severe ARDS. Compliance and body mass index did not modify the treatment effect. HFOV increased barotrauma risk compared with conventional ventilation (adjusted odds ratio, 1.75; 95% confidence interval, 1.04–2.96; P = 0.04). Conclusions: HFOV increases mortality for most patients with ARDS but may improve survival among patients with severe hypoxemia on conventional mechanical ventilation.