Extracorporeal membrane oxygenation (ECMO) is an increasingly used supportive measure for patients with refractory cardiogenic shock (CS). Despite its increasing use, there remain minimal data regarding which patients with refractory CS are most likely to benefit from ECMO. We retrospectively studied all patients (n = 123) who underwent initiation of ECMO for CS from February 2009 to September 2014 at a single center. Baseline patient characteristics, including demographics, co-morbid illness, cause of CS, available laboratory values, and patient outcomes were analyzed. Overall, 69 patients (56%) were weaned from ECMO, with 48 patients (39%) surviving to discharge. Survivors were younger (50 vs 60 years; p ≤0.0001), had a lower rate of previous smoking (27 vs 56%; p = 0.01) and chronic kidney disease (2% vs 13%; p = 0.03), and had lower lactate measured soon after ECMO initiation (3.1 vs 10.2 mmol/l; p = 0.01). Patients with pulmonary embolism (odds ratio 8.0, 95% confidence interval 2.00 to 31.99; p = 0.01) and acute cardiomyopathy (odds ratio 7.5, 95% confidence interval 1.69 to 33.27; p = 0.01) had a higher rate of survival than acute myocardial infarction, chronic cardiomyopathy, and miscellaneous etiologies compared to postcardiotomy CS as a referent. In conclusion, survival after ECMO initiation differs based on underlying cause of CS. Survival may be lower in older patients and those with early evidence of persistent hypoperfusion after initiation of ECMO for CS.
Cardiogenic shock (CS) is characterized by impaired cardiac output secondary to myocardial dysfunction with resultant hypoperfusion and end-organ dysfunction. Optimal selection and prompt initiation of circulatory assistance for patients with CS is an emerging area of interest in mechanical circulatory support. Of the varied options for mechanical support in CS, extracorporeal membrane oxygenation (ECMO) has emerged as a portable, rapidly deployable support method for single or biventricular heart failure, regardless of cause, as a bridge to recovery, definitive intervention, or a more permanent mechanical support solution. The use of ECMO to support patients with CS has grown in the past 10 years, with reported survival approaching 40% in large registries. The increasing use of ECMO support in the management of CS refractory to conventional management has prompted a critical focus on appropriate use and patient selection for optimal clinical outcome. Previous studies have demonstrated a potential survival benefit in these critically ill patients but are limited in size and do not provide information on which patients may be most likely to survive to definitive therapy or ECMO discontinuation. As a consequence, consensus guidelines on the use of ECMO remain limited. In this study, we sought to evaluate clinical and biochemical features suggestive of outcome in patients with CS supported by ECMO.
Methods
We retrospectively reviewed all adult patients who underwent ECMO support for refractory CS at Massachusetts General Hospital from February 2009 to September 2014. The study was approved by the institutional review board. Refractory CS consisted of evidence of persistent hypoperfusion or cardiac arrest associated with myocardial failure on optimal vasopressor, inotropic, or conventional percutaneous mechanical circulatory support, such as an intra-aortic balloon pump (IABP) or percutaneous ventricular assist device (Impella; Abiomed, Danvers, Massachusetts). The presence of CS was confirmed by a multidisciplinary “shock team” consisting of advanced heart failure specialists and cardiac surgeons to determine eligibility and appropriateness for ECMO support.
Baseline patient characteristics (demographics, co-morbid illness, cause of CS, and available laboratory values) were extracted from the medical record, where available. Echocardiography data were obtained from the medical record, as assessed by the interpreting cardiologist. Causes of CS were categorized as acute myocardial infarction (AMI), pulmonary embolism (PE), acute cardiomyopathy (CM), chronic cardiomyopathy, postcardiotomy shock, or other (miscellaneous) causes. Definitions for each cause are provided in Table 1 . We reviewed concurrent therapies, complications, and outcomes, including ability to wean from ECMO and survival to discharge. Complications included ischemic limb injury, bleeding, and neurologic injury. Ischemic limb injury was defined as significant injury requiring surgical or procedural intervention, not including those who underwent vascular repair on removal of the ECMO cannulas, which is commonly required. Significant bleeding was defined as a bleed requiring transfusion of >2 U of packed red blood cells or bleeding requiring procedural intervention. Neurologic injury was defined as evidence of ischemic or hemorrhagic stroke on computed tomography while on ECMO. Weaning from ECMO was considered successful if ECMO support was discontinued with intent for survival, with or without bridge to further mechanical support.
| Etiology | Definition |
|---|---|
| Acute myocardial infarction | Clinical syndrome suggestive of myocardial ischemia associated with characteristic ischemic electrocardiographic abnormalities and/or serum troponin elevation, with a clinical diagnosis by the treating physician of a type 1 myocardial infarction; (2) cardiac arrest felt to be secondary to myocardial ischemia; (3) post-coronary artery bypass surgery with CS felt by the treating physician to be secondary to myocardial ischemia (Type 5 myocardial infarction). |
| Pulmonary embolism | Evidence of embolism on diagnostic imaging or at time of pulmonary embolectomy. |
| Acute cardiomyopathy | No known prior history of heart failure with CS secondary to a primary myocardial process unrelated to ischemia; this included presumed or biopsy-proven myocarditis, tachycardia-induced cardiomyopathy, cardiomyopathy related to toxic exposure (e.g., alcohol), or idiopathic cardiomyopathy. |
| Chronic cardiomyopathy | Decompensation of a known cardiomyopathy (regardless of underlying etiology of myopathy, including those secondary to valvular disease) |
| Post-cardiotomy | Inability to wean from cardiopulmonary bypass or development of CS after cardiac surgery (including post-heart transplantation), not related to new myocardial ischemia. |
| Miscellaneous | All remaining patients, including post-cardiac arrest CS without another predisposing condition noted above and CS developing after any cardiac or non-cardiac procedure, toxic ingestion, or in those with an undetermined cause of CS. |
A multidisciplinary ECMO team consisting of cardiac surgeons, intensivists, cardiologists, perfusionist support, and respiratory therapists was responsible for the management of each patient. Cannulation was performed by the cardiac surgery team and consisted of both peripheral and central cannulation, as deemed appropriate by the performing surgeon, with use of a distal perfusion catheter as per the surgeon or cardiologist. In general, ECMO care (pump management, anticoagulation, and so forth) followed Extracorporeal Life Support Organization (ELSO) guidelines for all patients, unless clinically mandated otherwise (ELSO Guidelines for Cardiopulmonary Extracorporeal Life Support, unpublished, 2013). ECMO support was continued until adequate cardiac function returned for weaning, death, or health care proxy determination to withdraw support in the setting of multisystem organ failure or irreversible neurologic injury. Ability to wean was assessed by echocardiography and/or central hemodynamics (through a central venous or pulmonary arterial catheter) during decreased ECMO support.
Continuous variables are presented as medians and interquartile ranges, and categorical variables are presented as numbers and percentages. Nonparametric tests (Wilcoxon rank-sum and Kruskal–Wallis) were used to compare continuous covariates, and a chi-square test was used to compare categorical variables. A 2-tailed p value <0.05 was considered statistically significant. SAS 9.3 (SAS Institute, Cary, North Carolina) was used for statistical analyses.
Results
From February 2009 to September 2014, 123 patients were supported with ECMO for refractory CS. Baseline patient demographics and clinical characteristics are provided in Table 2 . Fifty-seven patients (46%) had a cardiac arrest before ECMO ( Figure 1 ). Patients with CS referred for ECMO had significant utilization of optimal medical therapy for CS, including 96% on vasopressor agents (76% on ≥2 agents) and 49% on inotropic support. Fifty-three patients (43%) had mechanical circulatory support before ECMO (40 with IABP, 10 with Impella, 3 with more advanced ventricular assist devices). The distribution of indications for initiation of ECMO support is shown in Figure 2 . Of 123 patients studied, 106 patients (86%) were classified into predetermined categories of CS, and 17 patients (14%) had a “miscellaneous” cause. Initiation of ECMO occurred at our institution in 102 patients (83%), with the remaining 21 patients (17%) transferred from another hospital with ongoing ECMO support. Of the patients cannulated at our institution, 61 (60%) were cannulated in the operating room, 24 (23%) in the cardiac catheterization laboratory, 16 (16%) at the bedside, and 1 (1%) in the emergency department. Cannulation was predominantly through peripheral vasculature (92 patients; 75%), with central cannulation in 31 patients (25%), of which 5 cases were right atrium to pulmonary artery for right ventricular support. Concomitant mechanical support devices were used in a small fraction of patients (13 with IABP, 6 with Impella percutaneous ventricular assist device).
| Variable | Total (n=123) | Wean (n=69) | Unable to Wean (n=54) | P Value | Survival to Discharge (n=48) | Death Prior to Discharge (n=75) | P Value |
|---|---|---|---|---|---|---|---|
| Age (years) | 56(41-65) | 55(40-62) | 57(44-66) | 0.30 | 50(34-57) | 60(47-67) | <0.0001 |
| Men | 85(69%) | 52(75%) | 33(61%) | 0.09 | 37(77%) | 48(64%) | 0.13 |
| Body Mass Index (kg/m 2 ) (121) | 27.4(24.2-32.9) | 27(23.9-32.9) | 28.3(24.9-32.7) | 0.53 | 27.0(23.3-31.7) | 28.2(24.9-33.1) | 0.31 |
| Coronary Artery Disease | 42(34%) | 24(35%) | 18(33%) | 0.87 | 14(29%) | 28(37%) | 0.35 |
| Prior Heart Failure | 37(30%) | 19(28%) | 18(33%) | 0.49 | 11(23%) | 26(35%) | 0.17 |
| Hypertension | 52(42%) | 31(45%) | 21(39%) | 0.50 | 18(38%) | 34(45%) | 0.39 |
| Chronic Kidney Disease | 11(9%) | 6(9%) | 5(9%) | 0.91 | 1(2%) | 10(13%) | 0.03 |
| Dialysis | 3(2%) | 1(1%) | 2(4%) | 0.42 | 0(0%) | 3(4%) | 0.16 |
| Diabetes | 25(20%) | 13(19%) | 12(22%) | 0.64 | 8(17%) | 17(23%) | 0.42 |
| Chronic Obstructive Pulmonary Disease | 5(4%) | 2(3%) | 3(6%) | 0.46 | 1(2%) | 4(5%) | 0.37 |
| Active Malignancy | 11(9%) | 6(9%) | 5(9%) | 0.91 | 4(8%) | 7(9%) | 0.85 |
| Cirrhosis | 5(4%) | 1(1%) | 4(7%) | 0.10 | 0(0%) | 5(7%) | 0.07 |
| Current or Former Smoker | 55(45%) | 27(39%) | 28(52%) | 0.28 | 13(27%) | 42(56%) | 0.01 |
| Cardiac Arrest | 57(46%) | 33(48%) | 24(44%) | 0.71 | 26(54%) | 31(41%) | 0.16 |
| Extracorporeal-Assisted Cardiopulmonary Resuscitation | 22(18%) | 10(15%) | 12(22%) | 0.27 | 6(13%) | 16(21%) | 0.21 |
| Troponin T (μg/L) (69) | 0.93(0.16-4.66) | 0.71(0.14-3.21) | 1.22(0.27-8.84) | 0.30 | 0.34(0.11-1.74) | 2.14(0.31-8.13) | 0.05 |
| Arterial pH (101) | 7.23(7.09-7.33) | 7.26(7.1-7.38) | 7.19(7.08-7.29) | 0.15 | 7.28(7.10-7.39) | 7.19(7.09-7.30) | 0.17 |
| Lactate (mmol/L) (54) | 7.5(3.8-12.2) | 8.3(3.9-12.2) | 6.9(3-12.6) | 0.72 | 6.2(2.7-12.1) | 8.6(4.5-14.4) | 0.23 |
| Anion Gap (mmol/L) (96) | 15(12-22) | 15(12-18) | 16(11-23) | 0.29 | 14(10-17) | 16(12-23) | 0.06 |
| Vasopressor (114) | 109(96%) | 58(94%) | 51(98%) | 0.24 | 38(93%) | 71(97%) | 0.25 |
| Number of Vasopressors (101) | 2(1-3) | 2(1-3) | 2(1-3) | 0.11 | 2(1-3) | 2(1-3) | 0.30 |
| Inotropes (108) | 53(49%) | 27(47%) | 26(51%) | 0.60 | 20(51%) | 33(47%) | 0.71 |
| Ejection Fraction Prior to Extracorporeal Membrane Oxygenation (80) | 30(17-57) | 25(14-59) | 38(21-55) | 0.25 | 21(14-59) | 40(21-55) | 0.15 |
| Right Ventricular Dysfunction (88) | 70(80%) | 42(82%) | 28(76%) | 0.44 | 31(91%) | 39(72%) | 0.03 |
| Percutaneous Coronary Intervention (122) | 22(18%) | 10(15%) | 12(22%) | 0.28 | 5(11%) | 17(23%) | 0.09 |
| Surgery | 45(37%) | 23(33%) | 22(41%) | 0.40 | 14(29%) | 31(41%) | 0.17 |
| Mechanical Circulatory Support | 53(43%) | 30(43%) | 23(43%) | 0.92 | 21(44%) | 32(43%) | 0.91 |
Sixty-nine patients (57%) were ultimately weaned off of ECMO, and 48 patients (39%) survived to discharge. In 70 patients with lactate measured 6 to 24 hours after ECMO initiation, median lactate was 5.6 mmol/L (interquartile range 2.8 to 13.0 mmol/l). Concomitant procedures from ECMO support were rare, with 7 patients undergoing percutaneous coronary intervention with ongoing ECMO support and 5 taken for coronary artery bypass grafting directly from ECMO support. Five patients underwent pulmonary embolectomy and 3 underwent catheter-directed fibrinolysis and/or thromboembolectomy while on ECMO; 2 patients had received systemic fibrinolysis and 2 had undergone embolectomy before ECMO initiation, with the remaining 5 receiving anticoagulation alone. Cannulation strategy was changed in 11 patients (peripheral to central in 8; 6 of those 8 transitioned from venoarterial [VA] ECMO to right atrial to pulmonary artery cannulation for right ventricular support alone).
Thirty-two patients (26% overall and 46% of those weaned) were bridged from ECMO to another form of mechanical circulatory support or transplant: durable left ventricular assist device (LVAD) in 16 patients, right ventricular assist device in 7, short-term biventricular assist devices in 4, right ventricular assist device plus durable LVAD in 2, IABP in 1, and heart transplant in 2 patients. An additional 8 patients were transitioned from VA ECMO to venovenous ECMO. Seventeen of the 29 patients (59%) who were bridged to any type of VAD survived to discharge, whereas 12 of the 18 patients (67%) who were bridged to a durable LVAD survived to discharge. Of the 48 patients discharged, 35 had follow-up at our institution after discharge; 32 of the 35 (91%) were alive at 6 months. Of the 75 patients who did not survive to discharge, cause of death was determined to be multisystem organ failure in 52 patients (69%), neurologic injury in 9 patients (12%), bleeding in 7 patients (6%), hypoxemic respiratory failure in 2 patients (2%), and other causes in the remaining 5 patients.
Baseline patient demographics and characteristics were analyzed to evaluate potential differences between those weaned and unable to wean from ECMO and survivors and nonsurvivors ( Tables 2 and 3 ). Survivors were more likely to be younger with a low rate of survival in those aged >65 years (7%, 2 of 28), with no patient over age 71 years surviving to discharge (0 of 5). Those with a PE (odds ratio 8.0; p = 0.01) and acute CM (odds ratio 7.5; p = 0.01) were more likely to survive to discharge compared to postcardiotomy patients as referent group ( Table 4 ). Of those with a cardiac arrest before ECMO, 22 had an AMI, 14 had a PE, 8 had decompensation of a chronic CM, 4 had acute CM, 1 postcardiotomy, and 8 from a miscellaneous cause. Further characteristics in patients with cardiac arrest are detailed in Figure 1 . In analysis of variables occurring after ECMO insertion, lactate and anion gap 6 to 24 hours after ECMO initiation were significantly lower in those who were successfully weaned from ECMO and in those who survived.
| Variable | Total (n=123) | Wean (n=69) | Unable to Wean (n=54) | P Value | Survival to Discharge (n=48) | Death Prior to Discharge (n=75) | P Value |
|---|---|---|---|---|---|---|---|
| Transfer on Extracorporeal Membrane Oxygenation | 21(17%) | 14(20%) | 7(13%) | 0.28 | 12(25%) | 9(12%) | 0.06 |
| Cannulation in Operating Room at Massachusetts General Hospital (102) | 61(60%) | 31(56%) | 30(64%) | 0.44 | 20(56%) | 41(62%) | 0.52 |
| Site of Cannulation (peripheral) | 92(75%) | 54(78%) | 38(70%) | 0.32 | 39(81%) | 53(71%) | 0.19 |
| Lactate (mmol/L) (70) | 5.6(2.8-13.0) | 4.1(2.5-8.7) | 10.2(4.4-18.7) | 0.01 | 3.1(2.3-6.5) | 10.2(4.4-15.5) | 0.01 |
| Anion Gap (mmol/L) (114) | 14(10-20) | 13(9-16) | 18(13-25) | 0.01 | 12(9-15) | 16(11-23) | 0.01 |
| Continuous Renal Replacement Therapy | 56(46%) | 27(39%) | 29(54%) | 0.11 | 17(35%) | 39(52%) | 0.07 |
| Concomitant Mechanical Circulatory Support | 19(15%) | 12(17%) | 7(13%) | 0.50 | 9(19%) | 10(13%) | 0.42 |
| Extracorporeal Membrane Oxygenation plus Vasopressor | 114(97%) | 63(95%) | 51(98%) | 0.43 | 43(93%) | 71(99%) | 0.13 |
| Extracorporeal Membrane Oxygenation plus Inotrope (113) | 73(65%) | 43(66%) | 30(63%) | 0.69 | 31(70%) | 42(61%) | 0.30 |
| Revascularization on Extracorporeal Membrane Oxygenation | 12(10%) | 9(13%) | 3(6%) | 0.16 | 2(4%) | 10(13%) | 0.09 |
| Bleeding | 82(67%) | 43(62%) | 39(72%) | 0.5 | 26(54%) | 56(75%) | 0.02 |
| Limb Ischemia | 18(15%) | 10(15%) | 8(7%) | 0.96 | 6(13%) | 12(16%) | 0.60 |
| Neurologic Injury | 15(12%) | 7(10%) | 8(15%) | 0.43 | 4(8%) | 11(15%) | 0.30 |
| Duration of Extracorporeal Membrane Oxygenation (hours) | 94(43-172) | 119(62-176) | 65(21-168) | 0.02 | 120(60-178) | 84(35-169) | 0.11 |
| Length of Stay (days) | 18(8-43) | 34(18-66) | 8.5(5-17) | 0.01 | 42(23-79) | 10(5-21) | 0.01 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
