Resident (PGY1) University of Toronto University of Toronto Toronto, Ontario, Canada
Disclosure(s):
Mansur Naeem, MD, PhD: No financial relationships to disclose
Background: Patients on veno-arterial extracorporeal membrane oxygenation (VA ECMO) are at risk of hemolysis and death. Magnetically levitated pumps (CMAG) are designed to reduce this risk by superior design. In this single center retrospective study, we compared hemolysis and survival in two different pumps (CardioHelp vs CMAG) in VA-ECMO patients.
Methods: We reviewed 102 patients who received VA-ECMO with CardioHelp (n=42) or CMAG (n = 60) treatment from 2018-2021 at University Health Network. Baseline characteristics including patient demographics and clinical indication between the two groups was compared using Mann-Whitney U-test for continues variables and chi-square test for nominal variables. Interval censored Kaplan-Meier (KM) curves were generated for visualization of survival data. Semi-parametric cox-proportional hazard (Cox-PH) model was (R-package IcenReg 2.0.16) was used to assess the effect of covariates on survival. A subset (n =61) of the study population, where laboratory data available during ECMO period, was used for hemolysis analysis. Hemolysis Index (HI) was used to assess for hemolysis. HI in CardioHelp vs CMAG was compared for each day on ECMO starting at day -1 (pre-ECMO). We then completed a Time-To-Event analysis where hemolysis was defined as HI > 0.5. Subsequently a competing risk analysis (to account for competing event of death) between the two pump groups. Lastly, the definition of hemolysis was expanded to include cases where pump thrombus was observed in addition to cases where HI>0.5. Relative risk (RR) was generated for hemolysis in each group as well as RR for death in patients with hemolysis vs no-hemolysis.
Outcome: Overall 102 patients were enrolled with mean age of 47, 70% male, median days on ECMO is 5 and these variables were not significantly different between CardioHelp and CMAG (Table 1). KM curves demonstrate no difference in survival curves between CMAG and CardioHelp (Figure 1). Furthermore, the Cox-PH model demonstrated none of the co-variates (age, sex, indication and days on ECMO) had statistically significant hazard ratio (HR). The hemolysis data showed that prior to ECMO patients on CardioHelp and CMAG had comparable HI. However, by day 2 of ECMO patients on CardioHelp had significantly higher HI and remained persistently high until Day 3 of ECMO (Figure 2). However, Time-To-Event analysis of hemolysis (HI > 0.5) in CardioHelp and CMAG were not statistically significant, this was confirmed with competing risk analysis (Figure 3). Lastly, RR analysis for hemolysis (HI > 0.5 and/or documented visible thrombus in the ECMO circuit) showed no difference in CardioHelp vs CMAG. Although patients with hemolysis had higher number of deaths this was not statistically significant (Figure 4).
Conclusion: Currently CMAG is marketed as a pump option with reduced hemolysis as such it costs more. Here we report the first retrospective head-to-head comparison of CMAG and CardioHelp hemolysis and survival. Our data suggest that CardioHelp is non-inferior to CMAG in patients on VA-ECMO with respect to survival and hemolysis.
.jpg "Mansur Naeem, MD, PhD photo")