Progressive hemolysis with combined VA-ECMO and Impella-CP support in a porcine model of cardiogenic shock

We have previously shown that veno-arterial extracorporeal membrane oxygenation (VA-ECMO) combined with Impella-CP improves cardiac energetics compared with VA-ECMO alone in a porcine model of cardiogenic shock (1). Further, in patients with cardiogenic shock, the DanGer shock trial demonstrated that Impella-CP reduced mortality compared to standard treatment, but was associated with higher rates of severe acute kidney injury (2). Device-related hemolysis may contribute to this finding, as free hemoglobin is nephrotoxic through tubular obstruction, oxidative stress, and renal vasoconstriction (3). However, the temporal dynamics of hemolysis induced by microaxial flow pumps have not been systematically characterized.

To compare urinary hemoglobin excretion, as a marker of hemolysis, between VA-ECMO alone and VA-ECMO + Impella-CP support in a porcine model of cardiogenic shock.

Fourteen female pigs (Danish Landrace x Yorkshire, ~75 kg) underwent left main coronary artery embolization to induce cardiogenic shock. Following induction of shock, animals were allocated to receive either VA-ECMO alone (n=7) or VA-ECMO + Impella-CP (n=7) support for 4 hours. Urine samples were collected at baseline, at the time of shock inducement, and hourly during mechanical circulatory support. Urinary hemoglobin was quantified by Western blot, normalized to urine osmolality, and expressed as fold-change from baseline. Linear mixed-effects modeling with random intercept was used to assess the rate of change over time between intervention groups.

Baseline characteristics and hemodynamic parameters were similar between groups. Total urine output was significantly higher with VA-ECMO + Impella-CP compared to VA-ECMO alone (median 600 mL vs. 265 mL, p=0.017; Fig. 1A). Temporal urine samples were available from 11 animals (VA-ECMO n=4, VA-ECMO + Impella-CP n=7), as three VA-ECMO animals became anuric during support. Urinary hemoglobin fraction was comparable at baseline and after shock induction. During mechanical circulatory support, the VA-ECMO group showed no change in urinary hemoglobin over time, whereas the VA-ECMO + Impella-CP group demonstrated a progressive increase throughout the study period (Group×Time interaction p=0.028, Fig. 1B). Urinary creatinine and protein fractions both increased over time in both groups, but with no differences between the two interventions (p=0.905 and p=0.633, respectively).

In a porcine model of cardiogenic shock, combined VA-ECMO and Impella-CP support was associated with significantly greater progressive urinary hemoglobin compared to VA-ECMO alone. Despite higher urine output with VA-ECMO + Impella-CP, the progressive nature of urinary hemoglobin may suggest device-related hemolysis and explain the higher incidence of acute kidney injury observed clinically with Impella-CP, particularly with prolonged support duration.Figure 1