Extracorporeal membrane oxygenation (ECMO) for the treatment of cardiogenic shock : identification of pre-implant outcome predictors

Sammanfattning: Background: Refractory cardiogenic shock (RCS), cardiac arrest (CA) and postarrest cardiogenic shock (CS) are associated with high mortality. Venoarterial extracorporeal membrane oxygenation (VA-ECMO) is increasingly used and can offer acute cardiopulmonary life support in this critically ill population but selection of VA-ECMO candidates remains challenging. There are limited data on which pre-VA-ECMO variables that predict outcome. Aims: To identify pre-VA-ECMO predictors for 90-day mortality in Study I unselected, Study II postcardiotomy and Study III non-surgical patients with RCS, and Study IV in patients with CA or postarrest cardiogenic shock prior to VA-ECMO were studied. Methods: Study I-IV were observational and retrospective. Study I included 181 mixed patients with RCS, Study II 105 patients with refractory postcardiotomy CS, Study III 76 non-surgical patients with RCS and Study IV 72 patients with CA or postarrest CS prior to VA-ECMO cannulation. The association between pre-implant variables and all-cause mortality at 90 days was analyzed with uni- and multivariable logistic regression. Results: Study I. Main etiologies of RCS were post-cardiotomy failure 58%, acute myocardial infarction 22% and other medical etiologies 20%. Median duration of VA-ECMO support was 7 days (interquartile range [IQR]: 3-13). The 90-day overall mortality was 54%. Arterial lactate (odds ratio [OR] per unit: 1.14; 95% confidence interval [CI]: 1.06-1.23; p <0.001), number of inotropes and vasopressors (OR per agent: 1.58; 95% CI: 1.13-2.21; p = 0.008), and ischemic heart disease (IHD), (OR: 2.90; 95% CI: 1.31-6.39; p = 0.008) were independent predictors of 90-day mortality. Study II. Main surgical subgroups were single non-CABG 29%, isolated CABG 20%, 2 and 3 concomitant surgical procedures, 31% and 20%, respectively. Median duration of VA-ECMO was 7 days (IQR: 3-14). The 90-day overall mortality and in-hospital mortality was 57% and 56%, respectively. Arterial lactate (OR per unit: 1.22; 95% CI:1.07-1.40; p = 0.004), and IHD (OR: 7.87; 95% CI: 2.55-24.3; p <0.001) were independent predictors of 90-day mortality. Study III. Main etiologies of RCS were acute myocardial infarction 51% and acute heart failure of other etiologies 49%. Median duration of VA-ECMO was 5 days (IQR: 2-11). The 90-day overall mortality was 49% and in-hospital mortality was 50%. Arterial lactate (OR per mmol/L: 1.15; 95% CI: 1.06-1.24; p = 0.001) and number of inotropes and vasopressors (OR per agent: 2.14; 95% CI: 1.26-3.63; p = 0.005) were independent predictors of 90-day mortality. Study IV. Out-of-hospital CA occurred in 12% of the patients. Initial cardiac rhythm was non-shockable in 57%, and shockable in 43% of the patients. Median cardiopulmonary resuscitation duration was 21 minutes (IQR: 10-73, range:1-197). No return of spontaneous circulation (ROSC) was present in 64% and postarrest CS in 36% of the patients at VA-ECMO cannulation. Median duration of VA-ECMO was 5 days (IQR: 2-12). The 90-day overall mortality and in-hospital mortality were 57%, respectively, and 53% died during VA-ECMO. All survivors (43%) had cerebral performance category score 1-2 at discharge to home. Initial non-shockable CA rhythm (OR: 12.2; 95% CI 2.83-52.7; p = 0.001), arterial lactate (OR per unit: 1.15; 95% CI: 1.01-1.31; p = 0.041), and IHD (OR: 7.39; 95% CI: 1.57-34.7; p = 0.011) as independent predictors of 90-day mortality. Conclusions: Identified independent pre-implant predictors for 90-day mortality after VA-ECMO initiation were in Study I arterial lactate, number of inotropes and vasopressors, and IHD, Study II arterial lactate and IHD, Study III arterial lactate, number of inotropes and vasopressors, Study IV initial non-shockable CA rhythm, arterial lactate and IHD. These predictors are easily available for pre-VA-ECMO risk prediction.