Noninvasive Respiratory Support Before Extracorporeal Membrane Oxygenation for Acute Respiratory Distress Syndrome: A Meta-Analysis.

Abstract

The optimal timing for transitioning from noninvasive respiratory support to invasive mechanical ventilation and extracorporeal membrane oxygenation (ECMO) in severe acute respiratory distress syndrome (ARDS) remains controversial. PubMed and Embase were searched for studies reporting the duration of noninvasive respiratory support and the total duration of pre-ECMO respiratory support among survivors and non-survivors. A random-effects model with the Hartung-Knapp adjustment was used. Certainty of evidence was rated using the GRADE approach. Ten studies involving 718 patients were included. Survivors spent significantly fewer days on noninvasive respiratory support before ECMO (mean difference [MD]: -1.88 days [95% confidence interval (CI): -2.81 to -0.96]; p < 0.01), with a 95% prediction interval (PI) ranging from -3.71 to -0.05; GRADE certainty: Low. In the total pre-ECMO respiratory support duration (noninvasive plus invasive) analysis, the MD widened to -3.49 days (95% CI: -6.20 to -0.78, p = 0.03), with a PI ranging from -8.04 to 1.06; GRADE certainty: Very Low. These data suggest that the duration of pre-ECMO noninvasive respiratory support could jeopardize the success of the treatment; however, they are based on observational studies and should be considered hypothesis-generating, advocating further research on this topic.