Hemolysis-Associated Nitric Oxide Dysregulation during Extracorporeal Membrane Oxygenation

Acute intravascular hemolysis during extracorporeal membrane oxygenation (ECMO) leads to increased levels of cell-free hemoglobin (FHb). Our aim was to investigate whether FHb levels are associated with nitric oxide (NO) consumption and clinical outcomes. A prospective observational study was performed involving pediatric patients on ECMO. Blood samples were collected before, during, and after the ECMO run, and plasma was evaluated for FHb, oxyhemoglobin, and NO consumption. Clinical data were collected including baseline patient characteristics, indications for ECMO, circuit changes, and mortality. Correlations between laboratory measures and associations between laboratory measures and clinical observations were evaluated. Twenty-three patients (11 male, 17 neonates) were enrolled with a median weight of 3.1 kg (interquartile range, 2.8-14.0 kg) and median ECMO run of 12 days (interquartile range, 5-19 day). There was a significant increase in FHb over time on ECMO (p = .007), and significant correlations were present between NO consumption and both FHb (r = .41, p = .01) and oxyhemoglobin levels (r = .98, p < .0001). Patients on ECMO for sepsis (n = 6) had lower average levels of oxyhemoglobin (mean [standard deviation {SD}] 14.5 [4.4] versus 19.0 [5.0] μM, p = .07) and NO consumption (mean [SD] 15.8 [4.1] versus 19.8 [3.7] μM, p = .04) during ECMO than patients with other indications. In the 3 days leading up to a circuit change, there were increases in mean total cell-free hemoglobin levels (24%/day, p = .08), oxyhemoglobin (37%/day, p = .005), and NO consumption (40%/day, p = .006) (n = 5). There were no significant associations identified between peak or average plasma measures of hemolysis and type of ECMO (venovenous versus venoarterial) or mortality. For children on ECMO, we observed a strong correlation between increased levels of plasma FHb and elevations in oxyhemoglobin and NO consumption; however, these changes were not associated with increased mortality. Increased hemolysis before circuit changes may be both a marker and a contributor to circuit failure.