The experience of extracorporeal membrane oxygenation for severe acute respiratory failure in adults

Abstract

To summarize the experience of extracorporeal membrane oxygenation (ECMO) for patients with severe acute respiratory failure in adults and to investigate the factors associated with death. The clinical data of patients with severe acute respiratory failure supported with ECMO in respiratory intensive care unit of Beijing Chaoyang Hospital from November 2009 to December 2011 were prospectively collected and analyzed. The data included general condition before EMCO, blood gas analysis, hemodynamics, ventilator settings of mechanical ventilation and complications during ECMO. The primary outcome was death or severe disability within 3 months. Statistical software of SPSS (version 16.0) was used for data analysis. Twenty-five patients with severe respiratory failure received ECMO treatment, of which 16 patients were analyzed. The mean age was (45 ± 14) years old (range, 22 - 64 years old). Thirteen patients were male. Before ECMO, all of the patients were treated with invasive positive pressure ventilation for (72 ± 64) hours. Eight patients had been treated with noninvasive ventilation for a median of 55(10-114) hours. Patients had severe respiratory failure despite advanced mechanical ventilator support. The mean PaO2/fraction of inspired oxygenation (FiO2) ratio was (54 ± 18) mm Hg (1 mm Hg = 0.133 kPa), positive end-expiratory pressure (PEEP) was (11 ± 6) cm H2O(1 cm H2O = 0.098 kPa), Murray lung injury score was 3.6 ± 0.5, serum lactate was (2.5 ± 2.0) mmol/L, serum white blood cell count was (16 ± 6)×10(9)/L, and APACHEII score was 17 ± 8. All of the patients were treated with venous-venous ECMO (VV-ECMO). The change of mechanical ventilation settings were (pre-ECMO vs 2 hours post-ECMO): FiO2 1.0 vs 0.55 ± 0.21, PEEP (11 ± 6) vs (9 ± 6) cm H2O, V(T) (6.8 ± 2.2) vs (4.4 ± 2.0) ml/kg PBW, peak airway pressure (27 ± 8) vs (24 ± 7) cm H2O, respiratory rate (37 ± 10) vs (23 ± 10) breaths/min. Arterial blood gas, including pH, PaO2 and PaCO2 were significantly improved after ECMO running 24 and 48 hours (P < 0.05). The mean VV-ECMO support interval was (9.7 ± 9.6) days (range, 2 - 41 days). Ten patients were successfully weaned from ECMO, of whom 2 died in ICU. Three patients died during ECMO, while another 3 patients died after withdrawal of further treatment. Eight patients survived to 3 months without severe disability. In a multi-variate Cox regression model, pre-ECMO factors including lower PaO2/FiO2 and increasing white blood cell count were associated with increased risks of death (RR was 0.733, 1.701 respectively, both P values < 0.05). VV-ECMO is a potentially effective approach for severe acute respiratory failure. PaO2/FiO2 and white blood cell count pre-ECMO may be the risk factors for poor outcome.