Mortality is directly related to the duration of mechanical ventilation before the initiation of extracorporeal life support for severe respiratory failure

Publisher Website
Google Scholar
Abstract
Objective To investigate the relationship between the period of mechanical ventilation before extracorporeal life support and survival in patients with respiratory failure. Design Retrospective review. Setting Surgical intensive care unit at a university medical center. Patients Thirty-six consecutive adult patients with severe respiratory failure managed with extracorporeal life support. Interventions Extracorporeal life support was utilized in 36 acute respiratory failure adult patients with a variety of diagnoses and an estimated mortality rate of >90%. Management protocols were followed before and during extracorporeal life support. The 36 patients were physiologically similar before extracorporeal life support was initiated: shunt of 48 +/- 17%; FIO2 of 1.0 +/- 0.1; peak inspiratory pressure of 56 +/- 16 cm H2 O; positive end-expiratory pressure of 14 +/- 6 cm H2 O; and respiratory rate of 23 +/- 10 breaths/min. Ventilation was utilized for 1 to 17 days before extracorporeal life support. Typical lung rest settings during extracorporeal life support were FIO2 of 0.40, peak inspiratory pressure of 30 cm H2 O, positive end-expiratory pressure of 10 cm H2 O, and respiratory rate of 6 breaths/min. Death was almost always secondary to end-stage pulmonary failure. Measurements and Main Results Survival (hospital discharge) in these 36 patients was inversely associated with the number of days of preextracorporeal life support ventilation, with a 50% mortality rate predicted by logistic regression after 5 days of mechanical ventilation. The overall survival rate was 18 (50.0%) of 36 patients. Conclusion In severe acute respiratory failure treated with lung rest and extracorporeal life support, a predicted 50% mortality rate was associated with 5 days of preextracorporeal life support mechanical ventilation. (Crit Care Med 1997; 25:28-32)