Management of Acute Respiratory Distress Syndrome and Refractory Hypoxemia. A Multicenter Observational Study.

RATIONALE: Clinicians' current practice patterns in the management of acute respiratory distress syndrome (ARDS) and refractory hypoxemia are not well described.

OBJECTIVES: To describe mechanical ventilation strategies and treatment adjuncts for adults with ARDS, including refractory hypoxemia.

METHODS: This was a prospective cohort study (March 2014-February 2015) of mechanically ventilated adults with moderate-to-severe ARDS requiring an FiO2 of 0.50 or greater in 24 intensive care units.

RESULTS: We enrolled 664 patients: 222 (33%) with moderate and 442 (67%) with severe ARDS. On Study Day 1, mean Vt was 7.5 (SD = 2.1) ml/kg predicted body weight (n = 625); 80% (n = 501) received Vt greater than 6 ml/kg. Mean positive end-expiratory pressure (PEEP) was 10.5 (3.7) cm H2 O (n = 653); 568 patients (87%) received PEEP less than 15 cm H2 O. Treatment adjuncts were common (n = 440, 66%): neuromuscular blockers (n = 276, 42%), pulmonary vasodilators (n = 118, 18%), prone positioning (n = 67, 10%), extracorporeal life support (n = 29, 4%), and high-frequency oscillatory ventilation (n = 29, 4%). Refractory hypoxemia, defined as PaO2 less than 60 mm Hg on FiO2 of 1.0, occurred in 138 (21%) patients. At onset of refractory hypoxemia, mean Vt was 7.1 (SD = 2.0) ml/kg (n = 124); 95 patients (77%) received Vt greater than 6 ml/kg. Mean PEEP was 12.1 (SD = 4.4) cm H2 O (n = 133); 99 patients (74%) received PEEP less than 15 cm H2 O. Among patients with refractory hypoxemia, 91% received treatment adjuncts (126/138), with increased use of neuromuscular blockers (n = 87, 63%), pulmonary vasodilators (n = 57, 41%), and prone positioning (n = 32, 23%).

CONCLUSIONS: Patients with moderate-to-severe ARDS receive treatment adjuncts frequently, especially with refractory hypoxemia. Paradoxically, therapies with less evidence supporting their use (e.g., pulmonary vasodilators) were over-used, whereas those with more evidence (e.g., prone positioning, neuromuscular blockade) were under-used. Patients received higher Vts and lower PEEP than would be suggested by the evidence.