Impact of Conservative Vs. Liberal Extracorporeal Oxygen Target During Venoarterial Extracorporeal Membrane Oxygenation Support for Cardiogenic Shock: A Pilot Randomized Control Trial.

Abstract

To evaluate if titration of sweep gas oxygen fraction (F D O 2 ) targeting post oxygenator conservative oxygen strategy might decrease organ dysfunction compared with a liberal extracorporeal oxygenation strategy. Pilot multicenter randomized controlled trial. Four ICUs in three teaching hospitals of the eastern France. Cardiogenic shock supported by venoarterial extracorporeal membrane oxygenation (VA ECMO). F D O 2 titration targeting postoxygenator oxygen partial pressure (P POST O 2 ) 100-150 mm Hg in the conservative arm, and F D O 2 maintained at 100% in the liberal arm. Primary outcome was mean plasmatic intestinal fatty acid binding protein (IFABP) level at day 2, a marker of gut damage. Secondary outcome were feasibility criteria, biomarkers of hepatic and renal dysfunction, biomarkers of inflammation and antioxidant system. Among the 55 patients analyzed, 29 were assigned to the conservative arm and 26 to the liberal arm. The two arms were well separated over the first 7 days: F D O 2 61 (± 7) % vs. 98 (± 5) %, and P POST O 2 139 (± 40) mm Hg vs. 420 (± 50) mm Hg, in the conservative and liberal arms, respectively. However, in the conservative arm, the mean proportion of time within the targeted window was only 33 (± 20) %. IFABP on day 2 did not differ between conservative and liberal arms (407 [206-549] pg/mL vs. 569 [247-708] pg/mL, p = 0.25). Lactate, vasoactive inotropic score, aspartate aminotransferase, prothrombin time, serum creatinine, need for renal replacement therapy, tumor necrosis factor-α, IL-6 and IL-8 did not differ between the two arms at day 0, 2, and 6 after randomization. In patients supported by VA ECMO for cardiogenic shock, a conservative oxygen target strategy was not easily feasible, as P POST O 2 could be maintained within the targeted window only 33% of time. However, the conservative and liberal arms were well separated regarding P POST O 2 , and we did not find any difference in biomarkers of organ dysfunction.