204: The Role of Respiratory Phenotype in Outcomes of ARDS Patients Supported with ECMO

Background: Due to differences in gas transportation in the blood, ECMO can completely eliminate CO2 across a range of blood flow rates, but is limited in oxygen delivery by the fraction of total venous return entrained. We hypothesized that ECMO would restore homeostasis in hypercapnic respiratory failure and lead to improved outcomes when compared to hypoxic failure. Methods: We conducted a retrospective view of ARDS patients on VV-ECMO in the ELSO database (2010-19). Using pre-cannulation blood gases to define patient groups, we compared hypercapnic (pH < 7.25 and PaCO2>60 mmHg) to hypoxic failure (PaO2:FiO2<100) while excluding those who met neither or both criteria. Baseline characteristics, ABG data, and outcomes were compared using chi-squared, student t-tests, or Wilcoxon rank-sum tests. Logistic regression was performed for the primary outcome of in-hospital mortality. Results: 7,632 patients were included, with 6,798 (89%) defined as hypoxic and 834 (11%) as hypercapnic. Hypercapnic patients were more likely to achieve systemic homeostasis (pH>7.35 and PaO2>65 mmHg) within 24-hours (60.4% vs 51.4%, p<0.001). ECMO duration was similar between groups (211 vs 213 hours, p=0.39). Hypercapnic patients had a lower incidence of limb ischemia (0.4% vs 1.3%, p=0.023), and renal failure (26.1% vs 32.8%, p<0.001), but had higher mortality (43.2% vs 35.5%, p<0.001). After adjusting for age, ECMO duration, comorbidities, baseline lung disease, and stroke, hypercapnic patients had 1.35 times greater odds of in-hospital mortality (CI 1.02-1.80). Conclusion: ECMO was more efficacious in addressing the physiologic derangements of hypercapnic respiratory failure, but this did not translate into a mortality benefit.