The influence of advanced age on venous–arterial extracorporeal membrane oxygenation outcomes

The influence of advanced age on venous–arterial extracorporeal membrane oxygenation outcomes Abstract OBJECTIVES Ethical and health care economic concerns surround the use of venous–arterial extracorporeal membrane oxygenation (VA-ECMO) in elderly patients. Patients requiring VA-ECMO are often in critical condition and the decision to cannulate is time-sensitive. We investigated the relationship between age and VA-ECMO outcomes to better inform this decision. METHODS This is a retrospective study of 355 patients placed on VA-ECMO between March 2007 and August 2016 at our institution. Using piecewise modelling, age became associated with in-hospital mortality after 63 years. Based on further analysis with the χ2 statistic maximization, patients were divided into 2 age groups: ≤72 years old [Group Y (Young), n = 310] and >72 years old [Group O (Old), n = 45]. Multivariable logistic regression was performed to identify preoperative predictors of in-hospital mortality. RESULTS Patients over the age of 72 had a significantly higher prevalence of comorbidities, including coronary disease, previous strokes and chronic kidney disease. Weaning from ECMO was achieved in 76% of Group Y and 47% of Group O (P < 0.001). In-hospital mortality was 52% among Group Y and 69% among Group O (P = 0.037). Multivariable logistic regression using preoperative risk factors identified coronary artery disease, acute decompensated heart failure and an age >72 years as independent predictors of mortality (age >72 years: odds ratio 2.71, 95% confidence interval 1.22–6.00; P = 0.01). CONCLUSIONS VA-ECMO in-hospital mortality is considerable across all age groups. However, age only becomes associated with mortality after 63 years and rises dramatically after 72 years. This study provides useful insight into these time-sensitive decisions for the development of possible practice guidelines. Extracorporeal membrane oxygenation, Extracorporeal life support, Acute myocardial infarction, Cardiogenic shock INTRODUCTION Extracorporeal life support (ECLS) using extracorporeal membrane oxygenation (ECMO) is an effective therapy for refractory cardiogenic shock. As new ECMO systems are developed and resuscitation protocols refined, patients previously not considered for ECLS are now being supported. While application of this therapy in patients with advanced age was reported to result in poor outcomes in previous small series [1, 2], there is insufficient literature to answer the question of ‘how old is too old?’ Many programmes entertain somewhat arbitrary upper age limits for ECMO use, but the immediate decision as to whether or not place an older patient on mechanical support can be complicated. Elderly patients in refractory cardiac failure often have significant and complex comorbidities that can affect outcomes. However, the decision to initiate ECMO oftentimes must be made quickly. While the left ventricular assist device literature has extensively investigated outcomes in elderly patients [3–5], there remains a paucity of literature involving the use of ECLS in the elderly. Moreover, often-cited guidelines that outline futility are nearly 2–3 decades old [6–8]. As the population ages, the prevalence of cardiovascular disease continues to rise, and the number of elderly patients undergoing cardiac surgery increases, an understanding of the impact of advanced age on outcomes will become increasingly more relevant. The goal of this study was to investigate the outcomes in patients of advanced age at our institution over the last 10 years and the utility of venous–arterial extracorporeal membrane oxygenation (VA-ECMO) in this population. MATERIALS AND METHODS Patients We retrospectively reviewed all patients >18 years who were placed on VA-ECMO at the New York Presbyterian Hospital/Columbia University Medical Center from January 2007 to August 2016. This study was approved by the institutional review board of Columbia University Medical Center, and individual consent was waived. Primary end-point and predictor variables The primary end-point was in-hospital mortality, and the primary predictor of interest was patient age at cannulation. A piecewise model was initially used to evaluate the association between age and in-hospital mortality. An age cut-off was identified by calculating the probabilities of 2 × 2 tables formed by the outcome and dichotomized age for all possible thresholds of age and calculating the associated χ2 statistic. The age with the maximum χ2 statistic was identified as the cut-off point. Simulation study has shown that if a cut-off point exists, maximizing the χ2 statistic can recover a true threshold for a continuous random variable [9]. Other predictor variables considered were pre-existing comorbidities such as coronary artery disease, hypertension, hyperlipidaemia, diabetes mellitus, prior stroke, chronic kidney disease, chronic obstructive pulmonary disease, baseline haemoglobin, platelets, creatinine, liver function tests and lactate as well as prior cardiac surgery through a midline sternotomy. Indications for venous–arterial extracorporeal membrane oxygenation Our mechanical circulatory support algorithm for refractory cardiogenic shock has been previously described [10]. We characterize cardiogenic shock by a systolic blood pressure of <90 mmHg, a cardiac index of <2.0 l/min/m2, a pulmonary capillary wedge pressure of >16 mmHg (or evidence of pulmonary oedema on chest radiography in the absence of a pulmonary artery catheter) and evidence of end-organ hypoperfusion. These patients were rapidly evaluated by a multidisciplinary ‘Shock Team’ comprising cardiac surgeons, interventional and heart failure cardiologists, intensivists and nurse practitioners to determine the most suitable device for each patient [11]. ECMO initiation occurs at the bedside, in the catheterization laboratory, or in the operating room—depending on the severity of haemodynamic compromise and patient status. Venous–arterial extracorporeal membrane oxygenation circuit and on-extracorporeal membrane oxygenation patient management The ECMO circuit is composed of a Quadrox D oxygenator (Maquet, Wayne, NJ, USA), Rotaflow pump (Maquet) and SMART-coated tubing (Sorin, Italy). ECMO flow was adjusted to provide systemic perfusion, which was monitored by mixed venous saturation and serum lactate, while maintaining native left ventricular (LV) ejection through the aortic valve. If clinically significant LV distension is noted after cannulation (increasing pulmonary artery diastolic pressure, worsening oxygenation, fulminant pulmonary oedema, refractory ventricular arrhythmias with enlarging LV end-diastolic diameter or significant stagnation of blood flow within the LV), the LV was vented with percutaneous femoral placement of an Impella 2.5 or CP LV assist device [12]. Patients were heparinized with a goal partial thromboplastin time of 60–80 s. Device weaning was considered when the patient demonstrated clinical improvement, as evidenced by improved end-organ function, reduction in vasoactive medication requirements and improved respiratory status. With adequate anticoagulation, device flow was temporarily decreased to 1 l/min at the bedside. Maintenance of satisfactory haemodynamics with acceptable central venous pressure (<13 mmHg) and mean arterial pressure (>70 mmHg) as well as echocardiographic parameters confirmed appropriate biventricular function, and VA-ECMO could be discontinued and subsequently explanted. When prolonged mechanical circulatory support was anticipated (generally longer than 7–10 days), VA-ECMO was aggressively converted to a short-term ventricular assist device [13, 14]. Bridging directly to an LV assist device was performed in a highly selected small number of cases. Statistical analysis Clinical and demographic variables were presented using standard summary statistics, including, mean ± standard deviation or median and interquartile range (25th, 75th percentile) depending on normality of distribution for continuous variables and frequencies and proportions for categorical variables. The distributions of continuous variables were tested with the Shapiro–Wilk test. To assess predictor variables and major morbidity, the χ2 or the Fisher’s exact test was used for categorical variables and the Student’s t-test or Mann–Whitney U-test for continuous variables. Variables with P-values ≤0.10 in univariate analyses were analysed in multivariable logistic regression and summarized as odds ratios and 95% confidence intervals. We made an a priori decision to include baseline lactate level in the model because of its widely reported association with mortality in shock patients. Backwards selection was adopted for variable selection in the final model. All statistical analyses were performed with SAS version 9.4, and a P-value of <0.05 was considered statistically significant. A propensity score-matched analysis was performed to create comparable risk groups between the advanced age and non-advanced age groups. Patients were matched according to significant differences between the 2 groups with respect to preoperative demographic factors and comorbidities. Matching was performed using the nearest neighbour algorithm, with Group O matched to Group Y based on a 4-digit match of the propensity score. RESULTS A total of 370 patients were placed on VA-ECMO at our institution during the study period. There were 15 patients over the age of 72 actively undergoing cardiopulmonary resuscitation at the time of ECMO cannulation (extracorporeal cardiopulmonary resuscitation) with mortality occurring in 12 (80%) patients. For the purpose of a more unbiased analysis, these patients were removed from subsequent analyses, leaving 355 patients for subsequent analyses. The overall median age was 59 years (interquartile range 48–68 years) and in-hospital mortality occurred in 191 (54%) patients. Cut-off age and baseline characteristics Based on the piecewise model, the estimated change point was at the age of 63 years (95% confidence interval 62.87–63.13, P <0.001). Prior to age 63 years, the estimated probability of death was the same across all ages—with no impact of age on the probability of death. After age 63 years, with every 1 year increase in age, the odds of death increased by 6% (odds ratio 1.06, P = 0.003). To establish an age cut-off, a distribution of the χ2 statistics was calculated for all possible cut-offs of age. The maximum χ2 was 7.374 when age was dichotomized at 72 years, generating 2 patient groups: >72 years (Group O) and ≤72 years (Group Y). These data are shown in Fig. 1. Patient characteristics are summarized in Table 1. Table 1: Characteristics and outcomes stratified by age   Overall (n = 355)  Age (years)   P-value  ≤72 (n =310)  >72 (n = 45)  Age (years)  59 (48–68)  56 (47–65)  78 (74–80)  <0.001  Female gender  113 (31)  96 (31)  17 (37)  0.39  Body mass index (kg/m2)  28.0 (23.9–32.2)  28.1 (24.4–31.7)  27.9 (23.7–31.9)  0.99  Hypertension  204 (57)  167 (54)  37 (82)  <0.001  Hyperlipidaemia  151 (43)  122 (39)  29 (64)  0.002  Coronary artery disease  161 (45)  144 (46)  28 (62)  0.034  Chronic obstructive pulmonary disease  29 (8)  22 (7)  7 (16)  0.074  Diabetes mellitus  100 (28)  85 (27)  15 (33)  0.48  Prior cerebrovascular accident  30 (8)  22 (7)  8 (18)  0.024  Chronic kidney disease  93 (26)  75 (24)  17 (38)  0.042  Previous cardiac surgery  67 (19)  59 (19)  8 (18)  1.00  Aetiology   Postcardiotomy shock  116 (33)  89 (29)  27 (60)  <0.001   Acute myocardial infarction  105 (30)  95 (31)  10 (22)  0.30   Primary graft dysfunction  41 (12)  41 (13)  0 (0.0)  0.010   Decompensated heart failure  55 (16)  52 (17)  3 (7)  0.088   Other (e.g. sepsis, myocarditis)  38 (11)  33 (11)  5 (11)  1.00  Central cannulation  50 (14)  39 (13)  11 (24)  0.040  Length of support (days)  4 (2–6)  4 (2–7)  3 (1–6)  0.506  Intensive care unit length of stay (days)  15 (5–30)  16 (5–31)  8 (4–20)  0.012  Hospital length of stay (days)  26 (9–48)  29 (10–51)  14 (5–33)  0.004  Major bleeding event  92 (26)  79 (26)  13 (29)  0.72  Survival to decannulation  257 (72)  236 (76)  21 (47)  <0.001  In-hospital mortality  191 (54)  160 (52)  31 (69)  0.037   Withdrawal of care  132 (69)  108 (67)  24 (77)  0.30    Overall (n = 355)  Age (years)   P-value  ≤72 (n =310)  >72 (n = 45)  Age (years)  59 (48–68)  56 (47–65)  78 (74–80)  <0.001  Female gender  113 (31)  96 (31)  17 (37)  0.39  Body mass index (kg/m2)  28.0 (23.9–32.2)  28.1 (24.4–31.7)  27.9 (23.7–31.9)  0.99  Hypertension  204 (57)  167 (54)  37 (82)  <0.001  Hyperlipidaemia  151 (43)  122 (39)  29 (64)  0.002  Coronary artery disease  161 (45)  144 (46)  28 (62)  0.034  Chronic obstructive pulmonary disease  29 (8)  22 (7)  7 (16)  0.074  Diabetes mellitus  100 (28)  85 (27)  15 (33)  0.48  Prior cerebrovascular accident  30 (8)  22 (7)  8 (18)  0.024  Chronic kidney disease  93 (26)  75 (24)  17 (38)  0.042  Previous cardiac surgery  67 (19)  59 (19)  8 (18)  1.00  Aetiology   Postcardiotomy shock  116 (33)  89 (29)  27 (60)  <0.001   Acute myocardial infarction  105 (30)  95 (31)  10 (22)  0.30   Primary graft dysfunction  41 (12)  41 (13)  0 (0.0)  0.010   Decompensated heart failure  55 (16)  52 (17)  3 (7)  0.088   Other (e.g. sepsis, myocarditis)  38 (11)  33 (11)  5 (11)  1.00  Central cannulation  50 (14)  39 (13)  11 (24)  0.040  Length of support (days)  4 (2–6)  4 (2–7)  3 (1–6)  0.506  Intensive care unit length of stay (days)  15 (5–30)  16 (5–31)  8 (4–20)  0.012  Hospital length of stay (days)  26 (9–48)  29 (10–51)  14 (5–33)  0.004  Major bleeding event  92 (26)  79 (26)  13 (29)  0.72  Survival to decannulation  257 (72)  236 (76)  21 (47)  <0.001  In-hospital mortality  191 (54)  160 (52)  31 (69)  0.037   Withdrawal of care  132 (69)  108 (67)  24 (77)  0.30  Data are presented as median (interquartile range) or frequency (%). Table 1: Characteristics and outcomes stratified by age   Overall (n = 355)  Age (years)   P-value  ≤72 (n =310)  >72 (n = 45)  Age (years)  59 (48–68)  56 (47–65)  78 (74–80)  <0.001  Female gender  113 (31)  96 (31)  17 (37)  0.39  Body mass index (kg/m2)  28.0 (23.9–32.2)  28.1 (24.4–31.7)  27.9 (23.7–31.9)  0.99  Hypertension  204 (57)  167 (54)  37 (82)  <0.001  Hyperlipidaemia  151 (43)  122 (39)  29 (64)  0.002  Coronary artery disease  161 (45)  144 (46)  28 (62)  0.034  Chronic obstructive pulmonary disease  29 (8)  22 (7)  7 (16)  0.074  Diabetes mellitus  100 (28)  85 (27)  15 (33)  0.48  Prior cerebrovascular accident  30 (8)  22 (7)  8 (18)  0.024  Chronic kidney disease  93 (26)  75 (24)  17 (38)  0.042  Previous cardiac surgery  67 (19)  59 (19)  8 (18)  1.00  Aetiology   Postcardiotomy shock  116 (33)  89 (29)  27 (60)  <0.001   Acute myocardial infarction  105 (30)  95 (31)  10 (22)  0.30   Primary graft dysfunction  41 (12)  41 (13)  0 (0.0)  0.010   Decompensated heart failure  55 (16)  52 (17)  3 (7)  0.088   Other (e.g. sepsis, myocarditis)  38 (11)  33 (11)  5 (11)  1.00  Central cannulation  50 (14)  39 (13)  11 (24)  0.040  Length of support (days)  4 (2–6)  4 (2–7)  3 (1–6)  0.506  Intensive care unit length of stay (days)  15 (5–30)  16 (5–31)  8 (4–20)  0.012  Hospital length of stay (days)  26 (9–48)  29 (10–51)  14 (5–33)  0.004  Major bleeding event  92 (26)  79 (26)  13 (29)  0.72  Survival to decannulation  257 (72)  236 (76)  21 (47)  <0.001  In-hospital mortality  191 (54)  160 (52)  31 (69)  0.037   Withdrawal of care  132 (69)  108 (67)  24 (77)  0.30    Overall (n = 355)  Age (years)   P-value  ≤72 (n =310)  >72 (n = 45)  Age (years)  59 (48–68)  56 (47–65)  78 (74–80)  <0.001  Female gender  113 (31)  96 (31)  17 (37)  0.39  Body mass index (kg/m2)  28.0 (23.9–32.2)  28.1 (24.4–31.7)  27.9 (23.7–31.9)  0.99  Hypertension  204 (57)  167 (54)  37 (82)  <0.001  Hyperlipidaemia  151 (43)  122 (39)  29 (64)  0.002  Coronary artery disease  161 (45)  144 (46)  28 (62)  0.034  Chronic obstructive pulmonary disease  29 (8)  22 (7)  7 (16)  0.074  Diabetes mellitus  100 (28)  85 (27)  15 (33)  0.48  Prior cerebrovascular accident  30 (8)  22 (7)  8 (18)  0.024  Chronic kidney disease  93 (26)  75 (24)  17 (38)  0.042  Previous cardiac surgery  67 (19)  59 (19)  8 (18)  1.00  Aetiology   Postcardiotomy shock  116 (33)  89 (29)  27 (60)  <0.001   Acute myocardial infarction  105 (30)  95 (31)  10 (22)  0.30   Primary graft dysfunction  41 (12)  41 (13)  0 (0.0)  0.010   Decompensated heart failure  55 (16)  52 (17)  3 (7)  0.088   Other (e.g. sepsis, myocarditis)  38 (11)  33 (11)  5 (11)  1.00  Central cannulation  50 (14)  39 (13)  11 (24)  0.040  Length of support (days)  4 (2–6)  4 (2–7)  3 (1–6)  0.506  Intensive care unit length of stay (days)  15 (5–30)  16 (5–31)  8 (4–20)  0.012  Hospital length of stay (days)  26 (9–48)  29 (10–51)  14 (5–33)  0.004  Major bleeding event  92 (26)  79 (26)  13 (29)  0.72  Survival to decannulation  257 (72)  236 (76)  21 (47)  <0.001  In-hospital mortality  191 (54)  160 (52)  31 (69)  0.037   Withdrawal of care  132 (69)  108 (67)  24 (77)  0.30  Data are presented as median (interquartile range) or frequency (%). Figure 1: View largeDownload slide Odds ratios and probability of in-hospital mortality in patients on VA-ECMO. Purple lines represent observed probability of death based on 10-year interval age cohorts (median age of each cohort is presented above the lines). The shaded grey region represents the 95% confidence interval of the predicted probability of death model. The blue line represents the unadjusted odds ratios for in-hospital mortality by age. VA-ECMO: venous–arterial extracorporeal membrane oxygenation. Figure 1: View largeDownload slide Odds ratios and probability of in-hospital mortality in patients on VA-ECMO. Purple lines represent observed probability of death based on 10-year interval age cohorts (median age of each cohort is presented above the lines). The shaded grey region represents the 95% confidence interval of the predicted probability of death model. The blue line represents the unadjusted odds ratios for in-hospital mortality by age. VA-ECMO: venous–arterial extracorporeal membrane oxygenation. There were 45 (13%) patients over 72 years old at the time of ECMO initiation (Group O), with a range of 73–90 years of age. Group O patients were more likely to have coronary disease, hypertension, hyperlipidaemia, a history of prior stroke and chronic kidney disease. They were also significantly more likely to have postcardiotomy shock as an indication for VA-ECMO with 27 (60%) patients experiencing postcardiotomy shock compared with 89 (29%) patients ≤72 years (P < 0.001). Clinical outcomes The median length of ECMO support was 4 days (interquartile range 1–6 days), with no difference between aetiologies. Weaning from ECMO was achieved in 76% of Group Y and in 47.0% of Group O (P < 0.001). The oldest patient who survived to hospital discharge was aged 84 years. The majority of the patients were cannulated peripherally using femoral or axillary arterial cannulation with femoral venous drainage. There were 50 (14%) patients who had central cannulation: 39 (13%) patients of Group Y and 11 (24%) patients of Group O (P = 0.040), with no difference in mortality based on cannulation strategy. There were no significant differences in pre-ECMO initiation labs between the age groups (Table 2). Table 2: Most recent pre-extracorporeal life support initiation labs   Overall (n = 370)  Age (years)   P-value  ≤72 (n  =310)  >72 (n = 60)  White blood cells (103/μl)  13 (8–19)  13 (8–19)  13 (7–19)  0.41  Haemoglobin (g/dl)  10.0 (9.0–12.3)  10.2 (9.0–12.4)  9.7 (8.9–11.4)  0.40  Platelets (103/μl)  144 (88–204)  150 (88–206)  128 (67–189)  0.25  Blood urea nitrogen (mg/dl)  28 (19–43)  28 (19–44)  28 (21–45)  0.74  Creatinine (mg/dl)  1.6 (1.1–2.2)  1.6 (1.1–2.2)  1.5 (1.1–2.1)  0.54  Glucose (mg/dl)  177 (130–233)  178 (133–236)  150 (113–197)  0.075  Total bilirubin (mg/dl)  1.2 (0.7–2.1)  1.3 (0.7–2.2)  1.1 (0.68–1.9)  0.39  Direct bilirubin (mg/dl)  0.5 (0.2–1.0)  0.5 (0.2–1.1)  0.4 (0.2–0.7)  0.19  AST (U/l)  80 (30–271)  80 (30–275)  76 (22–203)  0.73  ALT (U/l)  38 (20–127)  38 (22–126)  37 (14–124)  0.46  Lactate (mmol/l)  5.2 (2.7–8.8)  5.3 (2.9–9.6)  4.2 (2.3–7.7)  0.76  pH  7.34 (7.25–7.41)  7.33 (7.25–7.41)  7.34 (7.25–7.39)  0.97  INR  1.44 (1.23–1.94)  1.44 (1.24–1.94)  1.46 (1.17–1.97)  0.71    Overall (n = 370)  Age (years)   P-value  ≤72 (n  =310)  >72 (n = 60)  White blood cells (103/μl)  13 (8–19)  13 (8–19)  13 (7–19)  0.41  Haemoglobin (g/dl)  10.0 (9.0–12.3)  10.2 (9.0–12.4)  9.7 (8.9–11.4)  0.40  Platelets (103/μl)  144 (88–204)  150 (88–206)  128 (67–189)  0.25  Blood urea nitrogen (mg/dl)  28 (19–43)  28 (19–44)  28 (21–45)  0.74  Creatinine (mg/dl)  1.6 (1.1–2.2)  1.6 (1.1–2.2)  1.5 (1.1–2.1)  0.54  Glucose (mg/dl)  177 (130–233)  178 (133–236)  150 (113–197)  0.075  Total bilirubin (mg/dl)  1.2 (0.7–2.1)  1.3 (0.7–2.2)  1.1 (0.68–1.9)  0.39  Direct bilirubin (mg/dl)  0.5 (0.2–1.0)  0.5 (0.2–1.1)  0.4 (0.2–0.7)  0.19  AST (U/l)  80 (30–271)  80 (30–275)  76 (22–203)  0.73  ALT (U/l)  38 (20–127)  38 (22–126)  37 (14–124)  0.46  Lactate (mmol/l)  5.2 (2.7–8.8)  5.3 (2.9–9.6)  4.2 (2.3–7.7)  0.76  pH  7.34 (7.25–7.41)  7.33 (7.25–7.41)  7.34 (7.25–7.39)  0.97  INR  1.44 (1.23–1.94)  1.44 (1.24–1.94)  1.46 (1.17–1.97)  0.71  Data are presented as median (interquartile range) unless otherwise specified. ALT: alanine aminotransferase; AST: aspartate aminotransferase; INR: international normalized ratio. Table 2: Most recent pre-extracorporeal life support initiation labs   Overall (n = 370)  Age (years)   P-value  ≤72 (n  =310)  >72 (n = 60)  White blood cells (103/μl)  13 (8–19)  13 (8–19)  13 (7–19)  0.41  Haemoglobin (g/dl)  10.0 (9.0–12.3)  10.2 (9.0–12.4)  9.7 (8.9–11.4)  0.40  Platelets (103/μl)  144 (88–204)  150 (88–206)  128 (67–189)  0.25  Blood urea nitrogen (mg/dl)  28 (19–43)  28 (19–44)  28 (21–45)  0.74  Creatinine (mg/dl)  1.6 (1.1–2.2)  1.6 (1.1–2.2)  1.5 (1.1–2.1)  0.54  Glucose (mg/dl)  177 (130–233)  178 (133–236)  150 (113–197)  0.075  Total bilirubin (mg/dl)  1.2 (0.7–2.1)  1.3 (0.7–2.2)  1.1 (0.68–1.9)  0.39  Direct bilirubin (mg/dl)  0.5 (0.2–1.0)  0.5 (0.2–1.1)  0.4 (0.2–0.7)  0.19  AST (U/l)  80 (30–271)  80 (30–275)  76 (22–203)  0.73  ALT (U/l)  38 (20–127)  38 (22–126)  37 (14–124)  0.46  Lactate (mmol/l)  5.2 (2.7–8.8)  5.3 (2.9–9.6)  4.2 (2.3–7.7)  0.76  pH  7.34 (7.25–7.41)  7.33 (7.25–7.41)  7.34 (7.25–7.39)  0.97  INR  1.44 (1.23–1.94)  1.44 (1.24–1.94)  1.46 (1.17–1.97)  0.71    Overall (n = 370)  Age (years)   P-value  ≤72 (n  =310)  >72 (n = 60)  White blood cells (103/μl)  13 (8–19)  13 (8–19)  13 (7–19)  0.41  Haemoglobin (g/dl)  10.0 (9.0–12.3)  10.2 (9.0–12.4)  9.7 (8.9–11.4)  0.40  Platelets (103/μl)  144 (88–204)  150 (88–206)  128 (67–189)  0.25  Blood urea nitrogen (mg/dl)  28 (19–43)  28 (19–44)  28 (21–45)  0.74  Creatinine (mg/dl)  1.6 (1.1–2.2)  1.6 (1.1–2.2)  1.5 (1.1–2.1)  0.54  Glucose (mg/dl)  177 (130–233)  178 (133–236)  150 (113–197)  0.075  Total bilirubin (mg/dl)  1.2 (0.7–2.1)  1.3 (0.7–2.2)  1.1 (0.68–1.9)  0.39  Direct bilirubin (mg/dl)  0.5 (0.2–1.0)  0.5 (0.2–1.1)  0.4 (0.2–0.7)  0.19  AST (U/l)  80 (30–271)  80 (30–275)  76 (22–203)  0.73  ALT (U/l)  38 (20–127)  38 (22–126)  37 (14–124)  0.46  Lactate (mmol/l)  5.2 (2.7–8.8)  5.3 (2.9–9.6)  4.2 (2.3–7.7)  0.76  pH  7.34 (7.25–7.41)  7.33 (7.25–7.41)  7.34 (7.25–7.39)  0.97  INR  1.44 (1.23–1.94)  1.44 (1.24–1.94)  1.46 (1.17–1.97)  0.71  Data are presented as median (interquartile range) unless otherwise specified. ALT: alanine aminotransferase; AST: aspartate aminotransferase; INR: international normalized ratio. Total overall in-hospital mortality was 54%. Mortality in Group Y was 52% vs 69% in Group O (P = 0.037). Of the 12 patients over the age of 80, only 2 survived to discharge. Care was withdrawn in 69% of patients with no difference in the rates or causes of withdrawal between age groups both among those patients who died while on ECMO support and among those who died after decannulation. The most common reason for withdrawal of care was multisystem organ failure. The most common indications for support among patients of advanced age were postcardiotomy shock and acute myocardial infarction (AMI). In-hospital mortality in those patients with postcardiotomy shock occurred in 17 (63%) Group O patients and in 49 (55%) Group Y patients (P = 0.51). In patients with AMI, 8 (80%) of the 10 Group O patients and 49 (52%) of the 95 Group Y AMI patients died during their hospitalization (P = 0.075). Predictors of in-hospital mortality Based on our predefined selection criteria, age >72 years, hypertension, hyperlipidaemia, chronic kidney disease, chronic obstructive pulmonary disease, prior cerebrovascular accident, coronary artery disease, postcardiotomy shock, central cannulation, acute decompensated heart failure and immediate pre-ECMO lactate were variables that were adjusted for in a multivariable logistic regression to identify significant risk factors for in-hospital mortality. Using these risk factors, coronary artery disease, acute decompensated heart failure, baseline lactate and an age >72 were identified as independent predictors of in-hospital mortality. After adjusting for these risk factors, the odds of in-hospital mortality among patients >72 years of age was 195% higher than those ≤72 years of age (odds ratio 2.71, 95% confidence interval 1.22–6.00, P = 0.014). Table 3 presents the results for the full model and the final model. Table 3: Multivariable logistic regression of in-hospital mortality predictors   Odds ratio  95% confidence interval  P-value  Full model   Age >72 years  2.53  1.10–5.90  0.032   Postcardiotomy shock  1.35  0.70–2.60  0.37   Coronary artery disease  2.14  1.08–4.13  0.023   Hyperlipidaemia  0.70  0.33–1.49  0.35   Hypertension  0.86  0.40–1.83  0.69   Prior cerebrovascular accident  1.13  0.37–3.48  0.83   Chronic obstructive pulmonary disease  0.87  0.32–2.32  0.78   Chronic kidney disease  1.94  0.99–3.80  0.056   Acute decompensated heart failure  4.65  1.88–11.51  0.001   Central cannulation  1.68  0.74–3.82  0.22   Baseline lactate level (mmol/l)  1.08  1.01–1.15  0.024  Final model   Age >72 years  2.71  1.22–6.00  0.014   Coronary artery disease  1.67  0.94–2.98  0.081   Acute decompensated heart failure  4.27  1.69–9.95  0.001   Baseline lactate level (mmol/l)  1.09  1.02–1.15  0.011    Odds ratio  95% confidence interval  P-value  Full model   Age >72 years  2.53  1.10–5.90  0.032   Postcardiotomy shock  1.35  0.70–2.60  0.37   Coronary artery disease  2.14  1.08–4.13  0.023   Hyperlipidaemia  0.70  0.33–1.49  0.35   Hypertension  0.86  0.40–1.83  0.69   Prior cerebrovascular accident  1.13  0.37–3.48  0.83   Chronic obstructive pulmonary disease  0.87  0.32–2.32  0.78   Chronic kidney disease  1.94  0.99–3.80  0.056   Acute decompensated heart failure  4.65  1.88–11.51  0.001   Central cannulation  1.68  0.74–3.82  0.22   Baseline lactate level (mmol/l)  1.08  1.01–1.15  0.024  Final model   Age >72 years  2.71  1.22–6.00  0.014   Coronary artery disease  1.67  0.94–2.98  0.081   Acute decompensated heart failure  4.27  1.69–9.95  0.001   Baseline lactate level (mmol/l)  1.09  1.02–1.15  0.011  Table 3: Multivariable logistic regression of in-hospital mortality predictors   Odds ratio  95% confidence interval  P-value  Full model   Age >72 years  2.53  1.10–5.90  0.032   Postcardiotomy shock  1.35  0.70–2.60  0.37   Coronary artery disease  2.14  1.08–4.13  0.023   Hyperlipidaemia  0.70  0.33–1.49  0.35   Hypertension  0.86  0.40–1.83  0.69   Prior cerebrovascular accident  1.13  0.37–3.48  0.83   Chronic obstructive pulmonary disease  0.87  0.32–2.32  0.78   Chronic kidney disease  1.94  0.99–3.80  0.056   Acute decompensated heart failure  4.65  1.88–11.51  0.001   Central cannulation  1.68  0.74–3.82  0.22   Baseline lactate level (mmol/l)  1.08  1.01–1.15  0.024  Final model   Age >72 years  2.71  1.22–6.00  0.014   Coronary artery disease  1.67  0.94–2.98  0.081   Acute decompensated heart failure  4.27  1.69–9.95  0.001   Baseline lactate level (mmol/l)  1.09  1.02–1.15  0.011    Odds ratio  95% confidence interval  P-value  Full model   Age >72 years  2.53  1.10–5.90  0.032   Postcardiotomy shock  1.35  0.70–2.60  0.37   Coronary artery disease  2.14  1.08–4.13  0.023   Hyperlipidaemia  0.70  0.33–1.49  0.35   Hypertension  0.86  0.40–1.83  0.69   Prior cerebrovascular accident  1.13  0.37–3.48  0.83   Chronic obstructive pulmonary disease  0.87  0.32–2.32  0.78   Chronic kidney disease  1.94  0.99–3.80  0.056   Acute decompensated heart failure  4.65  1.88–11.51  0.001   Central cannulation  1.68  0.74–3.82  0.22   Baseline lactate level (mmol/l)  1.08  1.01–1.15  0.024  Final model   Age >72 years  2.71  1.22–6.00  0.014   Coronary artery disease  1.67  0.94–2.98  0.081   Acute decompensated heart failure  4.27  1.69–9.95  0.001   Baseline lactate level (mmol/l)  1.09  1.02–1.15  0.011  Propensity score matching was performed for Group O versus Group Y using 6 variables: shock aetiology, coronary artery disease, hyperlipidaemia, hypertension, prior cerebrovascular disease and chronic kidney disease. A total of 41 pairs were identified and compared (Table 4). Among these 2 matched groups, in-hospital mortality was significantly higher in the >72 years of age cohort at 68% (n = 28/41) compared with 43% (n = 18/41) of those ≤72 years of age (P = 0.022). Table 4: Characteristics and outcomes of propensity-matched groups   Overall (n = 82)  Age (years)   ≤ 72 (n =41)  >72 (n = 41)  P-value  Age (years)  72 (62–78)  62 (54–67)  78 (74–80)  <0.001  Female gender  24 (29)  9 (22)  15 (36)  0.22  Body mass index (kg/m2)  28.0 (24.6–33.0)  28.2 (25.0–33.9)  27.9 (23.7–31.9)  0.91  Hypertension  68 (83)  34 (41)  34 (41)  1.00  Hyperlipidaemia  54 (66)  27 (33)  27 (33)  1.00  Coronary artery disease  55 (67)  28 (34)  27 (33)  1.00  Chronic obstructive pulmonary disease  8 (10)  3 (4)  5 (6)  0.71  Diabetes mellitus  34 (41)  19 (23)  15 (18)  0.50  Prior cerebrovasculr accident  9 (11)  4 (5)  5 (6)  1.00  Chronic kidney disease  32 (39)  16 (20)  16 (20)  1.00  Smoking  9 (6)  5 (6)  4 (5)  1.00  Previous cardiac surgery  17 (21)  9 (11)  8 (10)  1.00  Aetiology   Postcardiotomy shock  53 (65)  27 (33)  26 (32)  1.00   Acute myocardial infarction  18 (22)  10 (12)  8 (10)  0.79   Primary graft dysfunction  0 (0)  0 (0)  0 (0)  1.00   Decompensated heart failure  6 (7)  3 (4)  3 (4)  1.00   Other (e.g. sepsis, myocarditis)  5 (6)  1 (1)  4 (5)  0.36  Length of support (days)  4.1 (1.9–7.3)  5.6 (3.2–7.5)  3.7 (1.1–7.1)  0.077  Intensive care unit length of stay (days)  12 (4–25)  19 (8–35)  8 (3–18)  0.020  Hospital length of stay (days)  22 (8–42)  28 (18–46)  12 (4–32)  0.059  Major bleeding event  2 (2)  1 (1)  1 (1)  1.00  In-hospital mortality  46 (56)  18 (43)  28 (68)  0.022    Overall (n = 82)  Age (years)   ≤ 72 (n =41)  >72 (n = 41)  P-value  Age (years)  72 (62–78)  62 (54–67)  78 (74–80)  <0.001  Female gender  24 (29)  9 (22)  15 (36)  0.22  Body mass index (kg/m2)  28.0 (24.6–33.0)  28.2 (25.0–33.9)  27.9 (23.7–31.9)  0.91  Hypertension  68 (83)  34 (41)  34 (41)  1.00  Hyperlipidaemia  54 (66)  27 (33)  27 (33)  1.00  Coronary artery disease  55 (67)  28 (34)  27 (33)  1.00  Chronic obstructive pulmonary disease  8 (10)  3 (4)  5 (6)  0.71  Diabetes mellitus  34 (41)  19 (23)  15 (18)  0.50  Prior cerebrovasculr accident  9 (11)  4 (5)  5 (6)  1.00  Chronic kidney disease  32 (39)  16 (20)  16 (20)  1.00  Smoking  9 (6)  5 (6)  4 (5)  1.00  Previous cardiac surgery  17 (21)  9 (11)  8 (10)  1.00  Aetiology   Postcardiotomy shock  53 (65)  27 (33)  26 (32)  1.00   Acute myocardial infarction  18 (22)  10 (12)  8 (10)  0.79   Primary graft dysfunction  0 (0)  0 (0)  0 (0)  1.00   Decompensated heart failure  6 (7)  3 (4)  3 (4)  1.00   Other (e.g. sepsis, myocarditis)  5 (6)  1 (1)  4 (5)  0.36  Length of support (days)  4.1 (1.9–7.3)  5.6 (3.2–7.5)  3.7 (1.1–7.1)  0.077  Intensive care unit length of stay (days)  12 (4–25)  19 (8–35)  8 (3–18)  0.020  Hospital length of stay (days)  22 (8–42)  28 (18–46)  12 (4–32)  0.059  Major bleeding event  2 (2)  1 (1)  1 (1)  1.00  In-hospital mortality  46 (56)  18 (43)  28 (68)  0.022  Data are presented as median (interquartile range) or frequency (%). Table 4: Characteristics and outcomes of propensity-matched groups   Overall (n = 82)  Age (years)   ≤ 72 (n =41)  >72 (n = 41)  P-value  Age (years)  72 (62–78)  62 (54–67)  78 (74–80)  <0.001  Female gender  24 (29)  9 (22)  15 (36)  0.22  Body mass index (kg/m2)  28.0 (24.6–33.0)  28.2 (25.0–33.9)  27.9 (23.7–31.9)  0.91  Hypertension  68 (83)  34 (41)  34 (41)  1.00  Hyperlipidaemia  54 (66)  27 (33)  27 (33)  1.00  Coronary artery disease  55 (67)  28 (34)  27 (33)  1.00  Chronic obstructive pulmonary disease  8 (10)  3 (4)  5 (6)  0.71  Diabetes mellitus  34 (41)  19 (23)  15 (18)  0.50  Prior cerebrovasculr accident  9 (11)  4 (5)  5 (6)  1.00  Chronic kidney disease  32 (39)  16 (20)  16 (20)  1.00  Smoking  9 (6)  5 (6)  4 (5)  1.00  Previous cardiac surgery  17 (21)  9 (11)  8 (10)  1.00  Aetiology   Postcardiotomy shock  53 (65)  27 (33)  26 (32)  1.00   Acute myocardial infarction  18 (22)  10 (12)  8 (10)  0.79   Primary graft dysfunction  0 (0)  0 (0)  0 (0)  1.00   Decompensated heart failure  6 (7)  3 (4)  3 (4)  1.00   Other (e.g. sepsis, myocarditis)  5 (6)  1 (1)  4 (5)  0.36  Length of support (days)  4.1 (1.9–7.3)  5.6 (3.2–7.5)  3.7 (1.1–7.1)  0.077  Intensive care unit length of stay (days)  12 (4–25)  19 (8–35)  8 (3–18)  0.020  Hospital length of stay (days)  22 (8–42)  28 (18–46)  12 (4–32)  0.059  Major bleeding event  2 (2)  1 (1)  1 (1)  1.00  In-hospital mortality  46 (56)  18 (43)  28 (68)  0.022    Overall (n = 82)  Age (years)   ≤ 72 (n =41)  >72 (n = 41)  P-value  Age (years)  72 (62–78)  62 (54–67)  78 (74–80)  <0.001  Female gender  24 (29)  9 (22)  15 (36)  0.22  Body mass index (kg/m2)  28.0 (24.6–33.0)  28.2 (25.0–33.9)  27.9 (23.7–31.9)  0.91  Hypertension  68 (83)  34 (41)  34 (41)  1.00  Hyperlipidaemia  54 (66)  27 (33)  27 (33)  1.00  Coronary artery disease  55 (67)  28 (34)  27 (33)  1.00  Chronic obstructive pulmonary disease  8 (10)  3 (4)  5 (6)  0.71  Diabetes mellitus  34 (41)  19 (23)  15 (18)  0.50  Prior cerebrovasculr accident  9 (11)  4 (5)  5 (6)  1.00  Chronic kidney disease  32 (39)  16 (20)  16 (20)  1.00  Smoking  9 (6)  5 (6)  4 (5)  1.00  Previous cardiac surgery  17 (21)  9 (11)  8 (10)  1.00  Aetiology   Postcardiotomy shock  53 (65)  27 (33)  26 (32)  1.00   Acute myocardial infarction  18 (22)  10 (12)  8 (10)  0.79   Primary graft dysfunction  0 (0)  0 (0)  0 (0)  1.00   Decompensated heart failure  6 (7)  3 (4)  3 (4)  1.00   Other (e.g. sepsis, myocarditis)  5 (6)  1 (1)  4 (5)  0.36  Length of support (days)  4.1 (1.9–7.3)  5.6 (3.2–7.5)  3.7 (1.1–7.1)  0.077  Intensive care unit length of stay (days)  12 (4–25)  19 (8–35)  8 (3–18)  0.020  Hospital length of stay (days)  22 (8–42)  28 (18–46)  12 (4–32)  0.059  Major bleeding event  2 (2)  1 (1)  1 (1)  1.00  In-hospital mortality  46 (56)  18 (43)  28 (68)  0.022  Data are presented as median (interquartile range) or frequency (%). DISCUSSION The findings of this study are summarized as follows: (i) age has an impact on mortality after age 63; (ii) ECLS in-hospital mortality dramatically increased for patients over 72 years old to nearly 70% and (iii) an age over 72 years was an independent predictor of in-hospital mortality when adjusted for the higher number of comorbidities in this population. ECLS is the most aggressive form of cardiopulmonary resuscitation for cardiogenic shock. Advancements in extracorporeal device technology have made this therapeutic modality more accessible and a rapidly growing body of literature has demonstrated an exponential increase in the use of VA-ECMO [15–18]. While there has been a suggestion that, in some circumstances, the use of extracorporeal support serves as a ‘bridge to nowhere’ [19], as cardiac surgical patients and the population at-large ages, so too does the potential need for VA-ECMO in postcardiotomy shock, heart failure, myocardial infarction and other aetiologies of refractory cardiac failure. In this study, we included patients requiring ECMO for a variety of reasons, most commonly for postcardiotomy shock and AMI, with a particular focus on a controversial demographic—patients of advanced age. Saxena et al. [20] described the Mayo Clinic experience with 45 postcardiotomy shock patients over the age of 70 and found an in-hospital mortality of 75.6%. Their conclusion was that perioperative cardiogenic shock in the elderly requiring extracorporeal support portends a poor prognosis but that age itself should not preclude patients from ECMO candidacy for postcardiotomy shock. Our mortality in these patients over 72 years of age was 65% and we are inclined to agree. Although postcardiotomy shock patients are acutely and severely ill due to surgical trauma and subsequent related complications, patients undergoing non-emergent cardiac surgery are presumably deemed, preoperatively, to be capable of withstanding the stress of a surgical intervention and thereby may have the reserve to tolerate extracorporeal support. Compared with postcardiotomy shock patients, elderly patients with AMI who were deemed to be suitable candidates for VA-ECMO had a worse morality rate—reaching 80%. Although subgroup analyses of shock aetiology-based outcomes are beyond the scope of this article, this observation is interesting, given that the existing literature cites postcardiotomy shock as having the worst outcome compared with cardiogenic shock from other aetiologies [14, 21, 22]. In cases of AMI, patients are often in cardiopulmonary arrest when they arrive at the hospital, and clinicians have insufficient time and clinical information about the patient’s risk factors to make well-informed decisions. Within our old age cohort, close to 40% of AMI patients underwent extracorporeal cardiopulmonary resuscitation and mortality occurred in two-thirds of these patients. An extracorporeal life support organization (ELSO) registry database investigation by Mendiratta et al. [23] demonstrated that extracorporeal cardiopulmonary resuscitation in patients of advanced age (median age 70 years) was associated with 78% mortality. Again, these results do not preclude all patients of advanced age from cannulation but may better inform emergent decision-making. Previous studies have suggested that age itself, regardless of comorbidities, should not preclude patients from being candidates for ECLS with VA-ECMO [1, 20, 24]. While we are not suggesting that advanced age be a contraindication to VA-ECMO, it is notable that mortality rates in our study accelerated in >72-year-old group to nearly 70% and reached 80% in those with AMI. Moreover, both our multivariable logistic regression and propensity-matched cohort comparisons confirmed that advanced age was an influential predictor of in-hospital mortality in patients on VA-ECMO. Advanced age patients are likely to have risk profiles that were not measured in this study such as functional capacity, physiological reserve and disqualification from more definitive cardiac replacement therapies. Many of these variables are not easily measurable in the current health care setting. However, the evaluation of frailty has become the standard practice in transcatheter aortic valve replacement, where its presence may adversely impact long-term post-procedural outcomes [25, 26]. Frailty has not been well defined in the extracorporeal support literature, but it is imperative to be able to identify such patients, particularly those of older age for whom the prospect of mortality is already high to better guide decision-making. As the field of mechanical circulatory support continues to advance, careful patient selection among the elderly population will become increasingly important. While biological age, alone, should be used cautiously—merely as one of multiple clinically relevant factors in the decision-making process—it is important to remember that advanced age is frequently accompanied by other influential comorbidities. This high level of collinearity should be taken into careful consideration at VA-ECMO consultation in which the initiation decision must be made with limited information and time. Future studies will need to develop risk stratification models to delineate those elderly patients who will benefit from ECLS and for those in whom support is futile. Limitations There are several limitations to this investigation. First, the age cut-off is data driven and has not been validated either internally or externally. In addition, using dichotomized age instead of continuous variables might result in loss of information. Second, the retrospective nature of this study subjects it to limitations inherent in all observational investigations. Third, these data were derived from a single, high-volume institution with limited sample sizes, which may limit overall generalizability and the feasibility to draw conclusions. We also lack follow-up quality of life and functional data on patients at discharge, which limits the scope of post-discharge outcomes analysis. Finally, conclusions drawn on our global in-hospital mortality data might minimize aetiology-specific differences among patients. We have tried to control for this by adjusting for the aetiologies significantly associated with mortality in our multivariable regression model. CONCLUSION In conclusion, ECLS offers a means of resuscitation for patients experiencing refractory cardiogenic shock but is associated with considerable mortality in patients of advanced age. Whether this mortality rate is prohibitive for this therapy will be better informed by future multi-institutional trials though it will, ultimately, be up to the health care team, which must consider each patient’s clinical scenario to determine how age might impact the most appropriate therapeutic approach. Further studies are needed to assess the effect of advanced age as well as the impact of cost and long-term functional outcomes. Conflict of interest: none declared. REFERENCES 1 Massetti M, Tasle M, Le Page O, Deredec R, Babatasi G, Buklas D et al.   Back from irreversibility: extracorporeal life support for prolonged cardiac arrest. Ann Thorac Surg  2005; 79: 178– 83; discussion 183–174. Google Scholar CrossRef Search ADS PubMed  2 Kolla S, Lee WA, Hirschl RB, Bartlett RH. Extracorporeal life support for cardiovascular support in adults. ASAIO J  1996; 42: M809– 19. Google Scholar CrossRef Search ADS PubMed  3 Rosenbaum AN, John R, Liao KK, Adatya S, Colvin-Adams MM, Pritzker M et al.   Survival in elderly patients supported with continuous flow left ventricular assist device as bridge to transplantation or destination therapy. J Card Fail  2014; 20: 161– 7. Google Scholar CrossRef Search ADS PubMed  4 Atluri P, Goldstone AB, Kobrin DM, Cohen JE, MacArthur JW, Howard JL et al.   Ventricular assist device implant in the elderly is associated with increased, but respectable risk: a multi-institutional study. Ann Thorac Surg  2013; 96: 141– 7. Google Scholar CrossRef Search ADS PubMed  5 Kilic A, Sultan I, Yuh DD, Shah AS, Baumgartner WA, Cameron DE et al.   Ventricular assist device implantation in the elderly: nationwide outcomes in the United States. J Card Surg  2013; 28: 183– 9. Google Scholar CrossRef Search ADS PubMed  6 Safar P, Abramson NS, Angelos M, Cantadore R, Leonov Y, Levine R et al.   Emergency cardiopulmonary bypass for resuscitation from prolonged cardiac arrest. Am J Emerg Med  1990; 8: 55– 67. Google Scholar CrossRef Search ADS PubMed  7 Hill JG, Bruhn PS, Cohen SE, Gallagher MW, Manart F, Moore CA et al.   Emergent applications of cardiopulmonary support: a multiinstitutional experience. Ann Thorac Surg  1992; 54: 699– 704. Google Scholar CrossRef Search ADS PubMed  8 Magovern GJJr, Simpson KA. Extracorporeal membrane oxygenation for adult cardiac support: the allegheny experience. Ann Thorac Surg  1999; 68: 655– 61. Google Scholar CrossRef Search ADS PubMed  9 Nelson S, Ramakrishnan V, Nietert PJ, Kamen DL, Ramos PS, Wolf BJ. An evaluation of common methods for dichotomization of continuous variables to discriminate disease status. Comm Stat Theory Methods  2016; 46: 23– 34. 10 Takayama H, Truby L, Koekort M, Uriel N, Colombo P, Mancini DM et al.   Clinical outcome of mechanical circulatory support for refractory cardiogenic shock in the current era. J Heart Lung Transplant  2013; 32: 106– 11. Google Scholar CrossRef Search ADS PubMed  11 Garan AR, Kirtane A, Takayama H. Redesigning care for patients with acute myocardial infarction complicated by cardiogenic shock: the ‘Shock Team’. JAMA Surg  2016; 151: 684– 85. Google Scholar CrossRef Search ADS PubMed  12 Truby LK, Takeda K, Mauro C, Yuzefpolskaya M, Garan AR, Kirtane AJ et al.   Incidence and implications of left ventricular distention during venoarterial extracorporeal membrane oxygenation support. ASAIO J  2017; 63: 257– 65. Google Scholar CrossRef Search ADS PubMed  13 Takeda K, Garan AR, Ando M, Han J, Topkara VK, Kurlansky P et al.   Minimally invasive CentriMag ventricular assist device support integrated with extracorporeal membrane oxygenation in cardiogenic shock patients: a comparison with conventional CentriMag biventricular support configuration. Eur J Cardiothorac Surg  2017; 52: 1055– 61. Google Scholar CrossRef Search ADS PubMed  14 Takayama H, Soni L, Kalesan B, Truby LK, Ota T, Cedola S et al.   Bridge-to-decision therapy with a continuous-flow external ventricular assist device in refractory cardiogenic shock of various causes. Circ Heart Fail  2014; 7: 799– 806. Google Scholar CrossRef Search ADS PubMed  15 Stretch R, Sauer CM, Yuh DD, Bonde P. National trends in the utilization of short-term mechanical circulatory support: incidence, outcomes, and cost analysis. J Am Coll Cardiol  2014; 64: 1407– 15. Google Scholar CrossRef Search ADS PubMed  16 Truby L, Mundy L, Kalesan B, Kirtane A, Colombo PC, Takeda K et al.   Contemporary outcomes of venoarterial extracorporeal membrane oxygenation for refractory cardiogenic shock at a large tertiary care center. ASAIO J  2015; 61: 403– 9. Google Scholar CrossRef Search ADS PubMed  17 Karagiannidis C, Brodie D, Strassmann S, Stoelben E, Philipp A, Bein T et al.   Extracorporeal membrane oxygenation: evolving epidemiology and mortality. Intensive Care Med  2016; 42: 889– 96. Google Scholar CrossRef Search ADS PubMed  18 Guenther SP, Brunner S, Born F, Fischer M, Schramm R, Pichlmaier M et al.   When all else fails: extracorporeal life support in therapy-refractory cardiogenic shock. Eur J Cardiothorac Surg  2016; 49: 802– 9. Google Scholar CrossRef Search ADS PubMed  19 Abrams DC, Prager K, Blinderman CD, Burkart KM, Brodie D. Ethical dilemmas encountered with the use of extracorporeal membrane oxygenation in adults. Chest  2014; 145: 876– 82. Google Scholar CrossRef Search ADS PubMed  20 Saxena P, Neal J, Joyce LD, Greason KL, Schaff HV, Guru P et al.   Extracorporeal membrane oxygenation support in postcardiotomy elderly patients: the Mayo Clinic experience. Ann Thorac Surg  2015; 99: 2053– 60. Google Scholar CrossRef Search ADS PubMed  21 Mohite PN, Sabashnikov A, Patil NP, Saez DG, Zych B, Popov AF et al.   Short-term ventricular assist device in post-cardiotomy cardiogenic shock: factors influencing survival. J Artif Organs  2014; 17: 228– 35. Google Scholar CrossRef Search ADS PubMed  22 Pokersnik JA, Buda T, Bashour CA, Gonzalez-Stawinski GV. Have changes in ECMO technology impacted outcomes in adult patients developing postcardiotomy cardiogenic shock? J Card Surg  2012; 27: 246– 52. Google Scholar CrossRef Search ADS PubMed  23 Mendiratta P, Wei JY, Gomez A, Podrazik P, Riggs AT, Rycus P et al.   Cardiopulmonary resuscitation requiring extracorporeal membrane oxygenation in the elderly: a review of the Extracorporeal Life Support Organization Registry. ASAIO J  2013; 59: 211– 5. Google Scholar CrossRef Search ADS PubMed  24 Saito S, Nakatani T, Kobayashi J, Tagusari O, Bando K, Niwaya K et al.   Is extracorporeal life support contraindicated in elderly patients? Ann Thorac Surg  2007; 83: 140– 5. Google Scholar CrossRef Search ADS PubMed  25 Green P, Woglom AE, Genereux P, Daneault B, Paradis JM, Schnell S et al.   The impact of frailty status on survival after transcatheter aortic valve replacement in older adults with severe aortic stenosis: a single-center experience. JACC Cardiovasc Interv  2012; 5: 974– 81. Google Scholar CrossRef Search ADS PubMed  26 Huded CP, Huded JM, Friedman JL, Benck LR, Lindquist LA, Holly TA et al.   Frailty status and outcomes after transcatheter aortic valve implantation. Am J Cardiol  2016; 117: 1966– 71. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Journal of Cardio-Thoracic Surgery Oxford University Press

Loading next page...
 
/lp/ou_press/the-influence-of-advanced-age-on-venous-arterial-extracorporeal-AOmMtefzUC
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
ISSN
1010-7940
eISSN
1873-734X
D.O.I.
10.1093/ejcts/ezx510
Publisher site
See Article on Publisher Site

Abstract

Abstract OBJECTIVES Ethical and health care economic concerns surround the use of venous–arterial extracorporeal membrane oxygenation (VA-ECMO) in elderly patients. Patients requiring VA-ECMO are often in critical condition and the decision to cannulate is time-sensitive. We investigated the relationship between age and VA-ECMO outcomes to better inform this decision. METHODS This is a retrospective study of 355 patients placed on VA-ECMO between March 2007 and August 2016 at our institution. Using piecewise modelling, age became associated with in-hospital mortality after 63 years. Based on further analysis with the χ2 statistic maximization, patients were divided into 2 age groups: ≤72 years old [Group Y (Young), n = 310] and >72 years old [Group O (Old), n = 45]. Multivariable logistic regression was performed to identify preoperative predictors of in-hospital mortality. RESULTS Patients over the age of 72 had a significantly higher prevalence of comorbidities, including coronary disease, previous strokes and chronic kidney disease. Weaning from ECMO was achieved in 76% of Group Y and 47% of Group O (P < 0.001). In-hospital mortality was 52% among Group Y and 69% among Group O (P = 0.037). Multivariable logistic regression using preoperative risk factors identified coronary artery disease, acute decompensated heart failure and an age >72 years as independent predictors of mortality (age >72 years: odds ratio 2.71, 95% confidence interval 1.22–6.00; P = 0.01). CONCLUSIONS VA-ECMO in-hospital mortality is considerable across all age groups. However, age only becomes associated with mortality after 63 years and rises dramatically after 72 years. This study provides useful insight into these time-sensitive decisions for the development of possible practice guidelines. Extracorporeal membrane oxygenation, Extracorporeal life support, Acute myocardial infarction, Cardiogenic shock INTRODUCTION Extracorporeal life support (ECLS) using extracorporeal membrane oxygenation (ECMO) is an effective therapy for refractory cardiogenic shock. As new ECMO systems are developed and resuscitation protocols refined, patients previously not considered for ECLS are now being supported. While application of this therapy in patients with advanced age was reported to result in poor outcomes in previous small series [1, 2], there is insufficient literature to answer the question of ‘how old is too old?’ Many programmes entertain somewhat arbitrary upper age limits for ECMO use, but the immediate decision as to whether or not place an older patient on mechanical support can be complicated. Elderly patients in refractory cardiac failure often have significant and complex comorbidities that can affect outcomes. However, the decision to initiate ECMO oftentimes must be made quickly. While the left ventricular assist device literature has extensively investigated outcomes in elderly patients [3–5], there remains a paucity of literature involving the use of ECLS in the elderly. Moreover, often-cited guidelines that outline futility are nearly 2–3 decades old [6–8]. As the population ages, the prevalence of cardiovascular disease continues to rise, and the number of elderly patients undergoing cardiac surgery increases, an understanding of the impact of advanced age on outcomes will become increasingly more relevant. The goal of this study was to investigate the outcomes in patients of advanced age at our institution over the last 10 years and the utility of venous–arterial extracorporeal membrane oxygenation (VA-ECMO) in this population. MATERIALS AND METHODS Patients We retrospectively reviewed all patients >18 years who were placed on VA-ECMO at the New York Presbyterian Hospital/Columbia University Medical Center from January 2007 to August 2016. This study was approved by the institutional review board of Columbia University Medical Center, and individual consent was waived. Primary end-point and predictor variables The primary end-point was in-hospital mortality, and the primary predictor of interest was patient age at cannulation. A piecewise model was initially used to evaluate the association between age and in-hospital mortality. An age cut-off was identified by calculating the probabilities of 2 × 2 tables formed by the outcome and dichotomized age for all possible thresholds of age and calculating the associated χ2 statistic. The age with the maximum χ2 statistic was identified as the cut-off point. Simulation study has shown that if a cut-off point exists, maximizing the χ2 statistic can recover a true threshold for a continuous random variable [9]. Other predictor variables considered were pre-existing comorbidities such as coronary artery disease, hypertension, hyperlipidaemia, diabetes mellitus, prior stroke, chronic kidney disease, chronic obstructive pulmonary disease, baseline haemoglobin, platelets, creatinine, liver function tests and lactate as well as prior cardiac surgery through a midline sternotomy. Indications for venous–arterial extracorporeal membrane oxygenation Our mechanical circulatory support algorithm for refractory cardiogenic shock has been previously described [10]. We characterize cardiogenic shock by a systolic blood pressure of <90 mmHg, a cardiac index of <2.0 l/min/m2, a pulmonary capillary wedge pressure of >16 mmHg (or evidence of pulmonary oedema on chest radiography in the absence of a pulmonary artery catheter) and evidence of end-organ hypoperfusion. These patients were rapidly evaluated by a multidisciplinary ‘Shock Team’ comprising cardiac surgeons, interventional and heart failure cardiologists, intensivists and nurse practitioners to determine the most suitable device for each patient [11]. ECMO initiation occurs at the bedside, in the catheterization laboratory, or in the operating room—depending on the severity of haemodynamic compromise and patient status. Venous–arterial extracorporeal membrane oxygenation circuit and on-extracorporeal membrane oxygenation patient management The ECMO circuit is composed of a Quadrox D oxygenator (Maquet, Wayne, NJ, USA), Rotaflow pump (Maquet) and SMART-coated tubing (Sorin, Italy). ECMO flow was adjusted to provide systemic perfusion, which was monitored by mixed venous saturation and serum lactate, while maintaining native left ventricular (LV) ejection through the aortic valve. If clinically significant LV distension is noted after cannulation (increasing pulmonary artery diastolic pressure, worsening oxygenation, fulminant pulmonary oedema, refractory ventricular arrhythmias with enlarging LV end-diastolic diameter or significant stagnation of blood flow within the LV), the LV was vented with percutaneous femoral placement of an Impella 2.5 or CP LV assist device [12]. Patients were heparinized with a goal partial thromboplastin time of 60–80 s. Device weaning was considered when the patient demonstrated clinical improvement, as evidenced by improved end-organ function, reduction in vasoactive medication requirements and improved respiratory status. With adequate anticoagulation, device flow was temporarily decreased to 1 l/min at the bedside. Maintenance of satisfactory haemodynamics with acceptable central venous pressure (<13 mmHg) and mean arterial pressure (>70 mmHg) as well as echocardiographic parameters confirmed appropriate biventricular function, and VA-ECMO could be discontinued and subsequently explanted. When prolonged mechanical circulatory support was anticipated (generally longer than 7–10 days), VA-ECMO was aggressively converted to a short-term ventricular assist device [13, 14]. Bridging directly to an LV assist device was performed in a highly selected small number of cases. Statistical analysis Clinical and demographic variables were presented using standard summary statistics, including, mean ± standard deviation or median and interquartile range (25th, 75th percentile) depending on normality of distribution for continuous variables and frequencies and proportions for categorical variables. The distributions of continuous variables were tested with the Shapiro–Wilk test. To assess predictor variables and major morbidity, the χ2 or the Fisher’s exact test was used for categorical variables and the Student’s t-test or Mann–Whitney U-test for continuous variables. Variables with P-values ≤0.10 in univariate analyses were analysed in multivariable logistic regression and summarized as odds ratios and 95% confidence intervals. We made an a priori decision to include baseline lactate level in the model because of its widely reported association with mortality in shock patients. Backwards selection was adopted for variable selection in the final model. All statistical analyses were performed with SAS version 9.4, and a P-value of <0.05 was considered statistically significant. A propensity score-matched analysis was performed to create comparable risk groups between the advanced age and non-advanced age groups. Patients were matched according to significant differences between the 2 groups with respect to preoperative demographic factors and comorbidities. Matching was performed using the nearest neighbour algorithm, with Group O matched to Group Y based on a 4-digit match of the propensity score. RESULTS A total of 370 patients were placed on VA-ECMO at our institution during the study period. There were 15 patients over the age of 72 actively undergoing cardiopulmonary resuscitation at the time of ECMO cannulation (extracorporeal cardiopulmonary resuscitation) with mortality occurring in 12 (80%) patients. For the purpose of a more unbiased analysis, these patients were removed from subsequent analyses, leaving 355 patients for subsequent analyses. The overall median age was 59 years (interquartile range 48–68 years) and in-hospital mortality occurred in 191 (54%) patients. Cut-off age and baseline characteristics Based on the piecewise model, the estimated change point was at the age of 63 years (95% confidence interval 62.87–63.13, P <0.001). Prior to age 63 years, the estimated probability of death was the same across all ages—with no impact of age on the probability of death. After age 63 years, with every 1 year increase in age, the odds of death increased by 6% (odds ratio 1.06, P = 0.003). To establish an age cut-off, a distribution of the χ2 statistics was calculated for all possible cut-offs of age. The maximum χ2 was 7.374 when age was dichotomized at 72 years, generating 2 patient groups: >72 years (Group O) and ≤72 years (Group Y). These data are shown in Fig. 1. Patient characteristics are summarized in Table 1. Table 1: Characteristics and outcomes stratified by age   Overall (n = 355)  Age (years)   P-value  ≤72 (n =310)  >72 (n = 45)  Age (years)  59 (48–68)  56 (47–65)  78 (74–80)  <0.001  Female gender  113 (31)  96 (31)  17 (37)  0.39  Body mass index (kg/m2)  28.0 (23.9–32.2)  28.1 (24.4–31.7)  27.9 (23.7–31.9)  0.99  Hypertension  204 (57)  167 (54)  37 (82)  <0.001  Hyperlipidaemia  151 (43)  122 (39)  29 (64)  0.002  Coronary artery disease  161 (45)  144 (46)  28 (62)  0.034  Chronic obstructive pulmonary disease  29 (8)  22 (7)  7 (16)  0.074  Diabetes mellitus  100 (28)  85 (27)  15 (33)  0.48  Prior cerebrovascular accident  30 (8)  22 (7)  8 (18)  0.024  Chronic kidney disease  93 (26)  75 (24)  17 (38)  0.042  Previous cardiac surgery  67 (19)  59 (19)  8 (18)  1.00  Aetiology   Postcardiotomy shock  116 (33)  89 (29)  27 (60)  <0.001   Acute myocardial infarction  105 (30)  95 (31)  10 (22)  0.30   Primary graft dysfunction  41 (12)  41 (13)  0 (0.0)  0.010   Decompensated heart failure  55 (16)  52 (17)  3 (7)  0.088   Other (e.g. sepsis, myocarditis)  38 (11)  33 (11)  5 (11)  1.00  Central cannulation  50 (14)  39 (13)  11 (24)  0.040  Length of support (days)  4 (2–6)  4 (2–7)  3 (1–6)  0.506  Intensive care unit length of stay (days)  15 (5–30)  16 (5–31)  8 (4–20)  0.012  Hospital length of stay (days)  26 (9–48)  29 (10–51)  14 (5–33)  0.004  Major bleeding event  92 (26)  79 (26)  13 (29)  0.72  Survival to decannulation  257 (72)  236 (76)  21 (47)  <0.001  In-hospital mortality  191 (54)  160 (52)  31 (69)  0.037   Withdrawal of care  132 (69)  108 (67)  24 (77)  0.30    Overall (n = 355)  Age (years)   P-value  ≤72 (n =310)  >72 (n = 45)  Age (years)  59 (48–68)  56 (47–65)  78 (74–80)  <0.001  Female gender  113 (31)  96 (31)  17 (37)  0.39  Body mass index (kg/m2)  28.0 (23.9–32.2)  28.1 (24.4–31.7)  27.9 (23.7–31.9)  0.99  Hypertension  204 (57)  167 (54)  37 (82)  <0.001  Hyperlipidaemia  151 (43)  122 (39)  29 (64)  0.002  Coronary artery disease  161 (45)  144 (46)  28 (62)  0.034  Chronic obstructive pulmonary disease  29 (8)  22 (7)  7 (16)  0.074  Diabetes mellitus  100 (28)  85 (27)  15 (33)  0.48  Prior cerebrovascular accident  30 (8)  22 (7)  8 (18)  0.024  Chronic kidney disease  93 (26)  75 (24)  17 (38)  0.042  Previous cardiac surgery  67 (19)  59 (19)  8 (18)  1.00  Aetiology   Postcardiotomy shock  116 (33)  89 (29)  27 (60)  <0.001   Acute myocardial infarction  105 (30)  95 (31)  10 (22)  0.30   Primary graft dysfunction  41 (12)  41 (13)  0 (0.0)  0.010   Decompensated heart failure  55 (16)  52 (17)  3 (7)  0.088   Other (e.g. sepsis, myocarditis)  38 (11)  33 (11)  5 (11)  1.00  Central cannulation  50 (14)  39 (13)  11 (24)  0.040  Length of support (days)  4 (2–6)  4 (2–7)  3 (1–6)  0.506  Intensive care unit length of stay (days)  15 (5–30)  16 (5–31)  8 (4–20)  0.012  Hospital length of stay (days)  26 (9–48)  29 (10–51)  14 (5–33)  0.004  Major bleeding event  92 (26)  79 (26)  13 (29)  0.72  Survival to decannulation  257 (72)  236 (76)  21 (47)  <0.001  In-hospital mortality  191 (54)  160 (52)  31 (69)  0.037   Withdrawal of care  132 (69)  108 (67)  24 (77)  0.30  Data are presented as median (interquartile range) or frequency (%). Table 1: Characteristics and outcomes stratified by age   Overall (n = 355)  Age (years)   P-value  ≤72 (n =310)  >72 (n = 45)  Age (years)  59 (48–68)  56 (47–65)  78 (74–80)  <0.001  Female gender  113 (31)  96 (31)  17 (37)  0.39  Body mass index (kg/m2)  28.0 (23.9–32.2)  28.1 (24.4–31.7)  27.9 (23.7–31.9)  0.99  Hypertension  204 (57)  167 (54)  37 (82)  <0.001  Hyperlipidaemia  151 (43)  122 (39)  29 (64)  0.002  Coronary artery disease  161 (45)  144 (46)  28 (62)  0.034  Chronic obstructive pulmonary disease  29 (8)  22 (7)  7 (16)  0.074  Diabetes mellitus  100 (28)  85 (27)  15 (33)  0.48  Prior cerebrovascular accident  30 (8)  22 (7)  8 (18)  0.024  Chronic kidney disease  93 (26)  75 (24)  17 (38)  0.042  Previous cardiac surgery  67 (19)  59 (19)  8 (18)  1.00  Aetiology   Postcardiotomy shock  116 (33)  89 (29)  27 (60)  <0.001   Acute myocardial infarction  105 (30)  95 (31)  10 (22)  0.30   Primary graft dysfunction  41 (12)  41 (13)  0 (0.0)  0.010   Decompensated heart failure  55 (16)  52 (17)  3 (7)  0.088   Other (e.g. sepsis, myocarditis)  38 (11)  33 (11)  5 (11)  1.00  Central cannulation  50 (14)  39 (13)  11 (24)  0.040  Length of support (days)  4 (2–6)  4 (2–7)  3 (1–6)  0.506  Intensive care unit length of stay (days)  15 (5–30)  16 (5–31)  8 (4–20)  0.012  Hospital length of stay (days)  26 (9–48)  29 (10–51)  14 (5–33)  0.004  Major bleeding event  92 (26)  79 (26)  13 (29)  0.72  Survival to decannulation  257 (72)  236 (76)  21 (47)  <0.001  In-hospital mortality  191 (54)  160 (52)  31 (69)  0.037   Withdrawal of care  132 (69)  108 (67)  24 (77)  0.30    Overall (n = 355)  Age (years)   P-value  ≤72 (n =310)  >72 (n = 45)  Age (years)  59 (48–68)  56 (47–65)  78 (74–80)  <0.001  Female gender  113 (31)  96 (31)  17 (37)  0.39  Body mass index (kg/m2)  28.0 (23.9–32.2)  28.1 (24.4–31.7)  27.9 (23.7–31.9)  0.99  Hypertension  204 (57)  167 (54)  37 (82)  <0.001  Hyperlipidaemia  151 (43)  122 (39)  29 (64)  0.002  Coronary artery disease  161 (45)  144 (46)  28 (62)  0.034  Chronic obstructive pulmonary disease  29 (8)  22 (7)  7 (16)  0.074  Diabetes mellitus  100 (28)  85 (27)  15 (33)  0.48  Prior cerebrovascular accident  30 (8)  22 (7)  8 (18)  0.024  Chronic kidney disease  93 (26)  75 (24)  17 (38)  0.042  Previous cardiac surgery  67 (19)  59 (19)  8 (18)  1.00  Aetiology   Postcardiotomy shock  116 (33)  89 (29)  27 (60)  <0.001   Acute myocardial infarction  105 (30)  95 (31)  10 (22)  0.30   Primary graft dysfunction  41 (12)  41 (13)  0 (0.0)  0.010   Decompensated heart failure  55 (16)  52 (17)  3 (7)  0.088   Other (e.g. sepsis, myocarditis)  38 (11)  33 (11)  5 (11)  1.00  Central cannulation  50 (14)  39 (13)  11 (24)  0.040  Length of support (days)  4 (2–6)  4 (2–7)  3 (1–6)  0.506  Intensive care unit length of stay (days)  15 (5–30)  16 (5–31)  8 (4–20)  0.012  Hospital length of stay (days)  26 (9–48)  29 (10–51)  14 (5–33)  0.004  Major bleeding event  92 (26)  79 (26)  13 (29)  0.72  Survival to decannulation  257 (72)  236 (76)  21 (47)  <0.001  In-hospital mortality  191 (54)  160 (52)  31 (69)  0.037   Withdrawal of care  132 (69)  108 (67)  24 (77)  0.30  Data are presented as median (interquartile range) or frequency (%). Figure 1: View largeDownload slide Odds ratios and probability of in-hospital mortality in patients on VA-ECMO. Purple lines represent observed probability of death based on 10-year interval age cohorts (median age of each cohort is presented above the lines). The shaded grey region represents the 95% confidence interval of the predicted probability of death model. The blue line represents the unadjusted odds ratios for in-hospital mortality by age. VA-ECMO: venous–arterial extracorporeal membrane oxygenation. Figure 1: View largeDownload slide Odds ratios and probability of in-hospital mortality in patients on VA-ECMO. Purple lines represent observed probability of death based on 10-year interval age cohorts (median age of each cohort is presented above the lines). The shaded grey region represents the 95% confidence interval of the predicted probability of death model. The blue line represents the unadjusted odds ratios for in-hospital mortality by age. VA-ECMO: venous–arterial extracorporeal membrane oxygenation. There were 45 (13%) patients over 72 years old at the time of ECMO initiation (Group O), with a range of 73–90 years of age. Group O patients were more likely to have coronary disease, hypertension, hyperlipidaemia, a history of prior stroke and chronic kidney disease. They were also significantly more likely to have postcardiotomy shock as an indication for VA-ECMO with 27 (60%) patients experiencing postcardiotomy shock compared with 89 (29%) patients ≤72 years (P < 0.001). Clinical outcomes The median length of ECMO support was 4 days (interquartile range 1–6 days), with no difference between aetiologies. Weaning from ECMO was achieved in 76% of Group Y and in 47.0% of Group O (P < 0.001). The oldest patient who survived to hospital discharge was aged 84 years. The majority of the patients were cannulated peripherally using femoral or axillary arterial cannulation with femoral venous drainage. There were 50 (14%) patients who had central cannulation: 39 (13%) patients of Group Y and 11 (24%) patients of Group O (P = 0.040), with no difference in mortality based on cannulation strategy. There were no significant differences in pre-ECMO initiation labs between the age groups (Table 2). Table 2: Most recent pre-extracorporeal life support initiation labs   Overall (n = 370)  Age (years)   P-value  ≤72 (n  =310)  >72 (n = 60)  White blood cells (103/μl)  13 (8–19)  13 (8–19)  13 (7–19)  0.41  Haemoglobin (g/dl)  10.0 (9.0–12.3)  10.2 (9.0–12.4)  9.7 (8.9–11.4)  0.40  Platelets (103/μl)  144 (88–204)  150 (88–206)  128 (67–189)  0.25  Blood urea nitrogen (mg/dl)  28 (19–43)  28 (19–44)  28 (21–45)  0.74  Creatinine (mg/dl)  1.6 (1.1–2.2)  1.6 (1.1–2.2)  1.5 (1.1–2.1)  0.54  Glucose (mg/dl)  177 (130–233)  178 (133–236)  150 (113–197)  0.075  Total bilirubin (mg/dl)  1.2 (0.7–2.1)  1.3 (0.7–2.2)  1.1 (0.68–1.9)  0.39  Direct bilirubin (mg/dl)  0.5 (0.2–1.0)  0.5 (0.2–1.1)  0.4 (0.2–0.7)  0.19  AST (U/l)  80 (30–271)  80 (30–275)  76 (22–203)  0.73  ALT (U/l)  38 (20–127)  38 (22–126)  37 (14–124)  0.46  Lactate (mmol/l)  5.2 (2.7–8.8)  5.3 (2.9–9.6)  4.2 (2.3–7.7)  0.76  pH  7.34 (7.25–7.41)  7.33 (7.25–7.41)  7.34 (7.25–7.39)  0.97  INR  1.44 (1.23–1.94)  1.44 (1.24–1.94)  1.46 (1.17–1.97)  0.71    Overall (n = 370)  Age (years)   P-value  ≤72 (n  =310)  >72 (n = 60)  White blood cells (103/μl)  13 (8–19)  13 (8–19)  13 (7–19)  0.41  Haemoglobin (g/dl)  10.0 (9.0–12.3)  10.2 (9.0–12.4)  9.7 (8.9–11.4)  0.40  Platelets (103/μl)  144 (88–204)  150 (88–206)  128 (67–189)  0.25  Blood urea nitrogen (mg/dl)  28 (19–43)  28 (19–44)  28 (21–45)  0.74  Creatinine (mg/dl)  1.6 (1.1–2.2)  1.6 (1.1–2.2)  1.5 (1.1–2.1)  0.54  Glucose (mg/dl)  177 (130–233)  178 (133–236)  150 (113–197)  0.075  Total bilirubin (mg/dl)  1.2 (0.7–2.1)  1.3 (0.7–2.2)  1.1 (0.68–1.9)  0.39  Direct bilirubin (mg/dl)  0.5 (0.2–1.0)  0.5 (0.2–1.1)  0.4 (0.2–0.7)  0.19  AST (U/l)  80 (30–271)  80 (30–275)  76 (22–203)  0.73  ALT (U/l)  38 (20–127)  38 (22–126)  37 (14–124)  0.46  Lactate (mmol/l)  5.2 (2.7–8.8)  5.3 (2.9–9.6)  4.2 (2.3–7.7)  0.76  pH  7.34 (7.25–7.41)  7.33 (7.25–7.41)  7.34 (7.25–7.39)  0.97  INR  1.44 (1.23–1.94)  1.44 (1.24–1.94)  1.46 (1.17–1.97)  0.71  Data are presented as median (interquartile range) unless otherwise specified. ALT: alanine aminotransferase; AST: aspartate aminotransferase; INR: international normalized ratio. Table 2: Most recent pre-extracorporeal life support initiation labs   Overall (n = 370)  Age (years)   P-value  ≤72 (n  =310)  >72 (n = 60)  White blood cells (103/μl)  13 (8–19)  13 (8–19)  13 (7–19)  0.41  Haemoglobin (g/dl)  10.0 (9.0–12.3)  10.2 (9.0–12.4)  9.7 (8.9–11.4)  0.40  Platelets (103/μl)  144 (88–204)  150 (88–206)  128 (67–189)  0.25  Blood urea nitrogen (mg/dl)  28 (19–43)  28 (19–44)  28 (21–45)  0.74  Creatinine (mg/dl)  1.6 (1.1–2.2)  1.6 (1.1–2.2)  1.5 (1.1–2.1)  0.54  Glucose (mg/dl)  177 (130–233)  178 (133–236)  150 (113–197)  0.075  Total bilirubin (mg/dl)  1.2 (0.7–2.1)  1.3 (0.7–2.2)  1.1 (0.68–1.9)  0.39  Direct bilirubin (mg/dl)  0.5 (0.2–1.0)  0.5 (0.2–1.1)  0.4 (0.2–0.7)  0.19  AST (U/l)  80 (30–271)  80 (30–275)  76 (22–203)  0.73  ALT (U/l)  38 (20–127)  38 (22–126)  37 (14–124)  0.46  Lactate (mmol/l)  5.2 (2.7–8.8)  5.3 (2.9–9.6)  4.2 (2.3–7.7)  0.76  pH  7.34 (7.25–7.41)  7.33 (7.25–7.41)  7.34 (7.25–7.39)  0.97  INR  1.44 (1.23–1.94)  1.44 (1.24–1.94)  1.46 (1.17–1.97)  0.71    Overall (n = 370)  Age (years)   P-value  ≤72 (n  =310)  >72 (n = 60)  White blood cells (103/μl)  13 (8–19)  13 (8–19)  13 (7–19)  0.41  Haemoglobin (g/dl)  10.0 (9.0–12.3)  10.2 (9.0–12.4)  9.7 (8.9–11.4)  0.40  Platelets (103/μl)  144 (88–204)  150 (88–206)  128 (67–189)  0.25  Blood urea nitrogen (mg/dl)  28 (19–43)  28 (19–44)  28 (21–45)  0.74  Creatinine (mg/dl)  1.6 (1.1–2.2)  1.6 (1.1–2.2)  1.5 (1.1–2.1)  0.54  Glucose (mg/dl)  177 (130–233)  178 (133–236)  150 (113–197)  0.075  Total bilirubin (mg/dl)  1.2 (0.7–2.1)  1.3 (0.7–2.2)  1.1 (0.68–1.9)  0.39  Direct bilirubin (mg/dl)  0.5 (0.2–1.0)  0.5 (0.2–1.1)  0.4 (0.2–0.7)  0.19  AST (U/l)  80 (30–271)  80 (30–275)  76 (22–203)  0.73  ALT (U/l)  38 (20–127)  38 (22–126)  37 (14–124)  0.46  Lactate (mmol/l)  5.2 (2.7–8.8)  5.3 (2.9–9.6)  4.2 (2.3–7.7)  0.76  pH  7.34 (7.25–7.41)  7.33 (7.25–7.41)  7.34 (7.25–7.39)  0.97  INR  1.44 (1.23–1.94)  1.44 (1.24–1.94)  1.46 (1.17–1.97)  0.71  Data are presented as median (interquartile range) unless otherwise specified. ALT: alanine aminotransferase; AST: aspartate aminotransferase; INR: international normalized ratio. Total overall in-hospital mortality was 54%. Mortality in Group Y was 52% vs 69% in Group O (P = 0.037). Of the 12 patients over the age of 80, only 2 survived to discharge. Care was withdrawn in 69% of patients with no difference in the rates or causes of withdrawal between age groups both among those patients who died while on ECMO support and among those who died after decannulation. The most common reason for withdrawal of care was multisystem organ failure. The most common indications for support among patients of advanced age were postcardiotomy shock and acute myocardial infarction (AMI). In-hospital mortality in those patients with postcardiotomy shock occurred in 17 (63%) Group O patients and in 49 (55%) Group Y patients (P = 0.51). In patients with AMI, 8 (80%) of the 10 Group O patients and 49 (52%) of the 95 Group Y AMI patients died during their hospitalization (P = 0.075). Predictors of in-hospital mortality Based on our predefined selection criteria, age >72 years, hypertension, hyperlipidaemia, chronic kidney disease, chronic obstructive pulmonary disease, prior cerebrovascular accident, coronary artery disease, postcardiotomy shock, central cannulation, acute decompensated heart failure and immediate pre-ECMO lactate were variables that were adjusted for in a multivariable logistic regression to identify significant risk factors for in-hospital mortality. Using these risk factors, coronary artery disease, acute decompensated heart failure, baseline lactate and an age >72 were identified as independent predictors of in-hospital mortality. After adjusting for these risk factors, the odds of in-hospital mortality among patients >72 years of age was 195% higher than those ≤72 years of age (odds ratio 2.71, 95% confidence interval 1.22–6.00, P = 0.014). Table 3 presents the results for the full model and the final model. Table 3: Multivariable logistic regression of in-hospital mortality predictors   Odds ratio  95% confidence interval  P-value  Full model   Age >72 years  2.53  1.10–5.90  0.032   Postcardiotomy shock  1.35  0.70–2.60  0.37   Coronary artery disease  2.14  1.08–4.13  0.023   Hyperlipidaemia  0.70  0.33–1.49  0.35   Hypertension  0.86  0.40–1.83  0.69   Prior cerebrovascular accident  1.13  0.37–3.48  0.83   Chronic obstructive pulmonary disease  0.87  0.32–2.32  0.78   Chronic kidney disease  1.94  0.99–3.80  0.056   Acute decompensated heart failure  4.65  1.88–11.51  0.001   Central cannulation  1.68  0.74–3.82  0.22   Baseline lactate level (mmol/l)  1.08  1.01–1.15  0.024  Final model   Age >72 years  2.71  1.22–6.00  0.014   Coronary artery disease  1.67  0.94–2.98  0.081   Acute decompensated heart failure  4.27  1.69–9.95  0.001   Baseline lactate level (mmol/l)  1.09  1.02–1.15  0.011    Odds ratio  95% confidence interval  P-value  Full model   Age >72 years  2.53  1.10–5.90  0.032   Postcardiotomy shock  1.35  0.70–2.60  0.37   Coronary artery disease  2.14  1.08–4.13  0.023   Hyperlipidaemia  0.70  0.33–1.49  0.35   Hypertension  0.86  0.40–1.83  0.69   Prior cerebrovascular accident  1.13  0.37–3.48  0.83   Chronic obstructive pulmonary disease  0.87  0.32–2.32  0.78   Chronic kidney disease  1.94  0.99–3.80  0.056   Acute decompensated heart failure  4.65  1.88–11.51  0.001   Central cannulation  1.68  0.74–3.82  0.22   Baseline lactate level (mmol/l)  1.08  1.01–1.15  0.024  Final model   Age >72 years  2.71  1.22–6.00  0.014   Coronary artery disease  1.67  0.94–2.98  0.081   Acute decompensated heart failure  4.27  1.69–9.95  0.001   Baseline lactate level (mmol/l)  1.09  1.02–1.15  0.011  Table 3: Multivariable logistic regression of in-hospital mortality predictors   Odds ratio  95% confidence interval  P-value  Full model   Age >72 years  2.53  1.10–5.90  0.032   Postcardiotomy shock  1.35  0.70–2.60  0.37   Coronary artery disease  2.14  1.08–4.13  0.023   Hyperlipidaemia  0.70  0.33–1.49  0.35   Hypertension  0.86  0.40–1.83  0.69   Prior cerebrovascular accident  1.13  0.37–3.48  0.83   Chronic obstructive pulmonary disease  0.87  0.32–2.32  0.78   Chronic kidney disease  1.94  0.99–3.80  0.056   Acute decompensated heart failure  4.65  1.88–11.51  0.001   Central cannulation  1.68  0.74–3.82  0.22   Baseline lactate level (mmol/l)  1.08  1.01–1.15  0.024  Final model   Age >72 years  2.71  1.22–6.00  0.014   Coronary artery disease  1.67  0.94–2.98  0.081   Acute decompensated heart failure  4.27  1.69–9.95  0.001   Baseline lactate level (mmol/l)  1.09  1.02–1.15  0.011    Odds ratio  95% confidence interval  P-value  Full model   Age >72 years  2.53  1.10–5.90  0.032   Postcardiotomy shock  1.35  0.70–2.60  0.37   Coronary artery disease  2.14  1.08–4.13  0.023   Hyperlipidaemia  0.70  0.33–1.49  0.35   Hypertension  0.86  0.40–1.83  0.69   Prior cerebrovascular accident  1.13  0.37–3.48  0.83   Chronic obstructive pulmonary disease  0.87  0.32–2.32  0.78   Chronic kidney disease  1.94  0.99–3.80  0.056   Acute decompensated heart failure  4.65  1.88–11.51  0.001   Central cannulation  1.68  0.74–3.82  0.22   Baseline lactate level (mmol/l)  1.08  1.01–1.15  0.024  Final model   Age >72 years  2.71  1.22–6.00  0.014   Coronary artery disease  1.67  0.94–2.98  0.081   Acute decompensated heart failure  4.27  1.69–9.95  0.001   Baseline lactate level (mmol/l)  1.09  1.02–1.15  0.011  Propensity score matching was performed for Group O versus Group Y using 6 variables: shock aetiology, coronary artery disease, hyperlipidaemia, hypertension, prior cerebrovascular disease and chronic kidney disease. A total of 41 pairs were identified and compared (Table 4). Among these 2 matched groups, in-hospital mortality was significantly higher in the >72 years of age cohort at 68% (n = 28/41) compared with 43% (n = 18/41) of those ≤72 years of age (P = 0.022). Table 4: Characteristics and outcomes of propensity-matched groups   Overall (n = 82)  Age (years)   ≤ 72 (n =41)  >72 (n = 41)  P-value  Age (years)  72 (62–78)  62 (54–67)  78 (74–80)  <0.001  Female gender  24 (29)  9 (22)  15 (36)  0.22  Body mass index (kg/m2)  28.0 (24.6–33.0)  28.2 (25.0–33.9)  27.9 (23.7–31.9)  0.91  Hypertension  68 (83)  34 (41)  34 (41)  1.00  Hyperlipidaemia  54 (66)  27 (33)  27 (33)  1.00  Coronary artery disease  55 (67)  28 (34)  27 (33)  1.00  Chronic obstructive pulmonary disease  8 (10)  3 (4)  5 (6)  0.71  Diabetes mellitus  34 (41)  19 (23)  15 (18)  0.50  Prior cerebrovasculr accident  9 (11)  4 (5)  5 (6)  1.00  Chronic kidney disease  32 (39)  16 (20)  16 (20)  1.00  Smoking  9 (6)  5 (6)  4 (5)  1.00  Previous cardiac surgery  17 (21)  9 (11)  8 (10)  1.00  Aetiology   Postcardiotomy shock  53 (65)  27 (33)  26 (32)  1.00   Acute myocardial infarction  18 (22)  10 (12)  8 (10)  0.79   Primary graft dysfunction  0 (0)  0 (0)  0 (0)  1.00   Decompensated heart failure  6 (7)  3 (4)  3 (4)  1.00   Other (e.g. sepsis, myocarditis)  5 (6)  1 (1)  4 (5)  0.36  Length of support (days)  4.1 (1.9–7.3)  5.6 (3.2–7.5)  3.7 (1.1–7.1)  0.077  Intensive care unit length of stay (days)  12 (4–25)  19 (8–35)  8 (3–18)  0.020  Hospital length of stay (days)  22 (8–42)  28 (18–46)  12 (4–32)  0.059  Major bleeding event  2 (2)  1 (1)  1 (1)  1.00  In-hospital mortality  46 (56)  18 (43)  28 (68)  0.022    Overall (n = 82)  Age (years)   ≤ 72 (n =41)  >72 (n = 41)  P-value  Age (years)  72 (62–78)  62 (54–67)  78 (74–80)  <0.001  Female gender  24 (29)  9 (22)  15 (36)  0.22  Body mass index (kg/m2)  28.0 (24.6–33.0)  28.2 (25.0–33.9)  27.9 (23.7–31.9)  0.91  Hypertension  68 (83)  34 (41)  34 (41)  1.00  Hyperlipidaemia  54 (66)  27 (33)  27 (33)  1.00  Coronary artery disease  55 (67)  28 (34)  27 (33)  1.00  Chronic obstructive pulmonary disease  8 (10)  3 (4)  5 (6)  0.71  Diabetes mellitus  34 (41)  19 (23)  15 (18)  0.50  Prior cerebrovasculr accident  9 (11)  4 (5)  5 (6)  1.00  Chronic kidney disease  32 (39)  16 (20)  16 (20)  1.00  Smoking  9 (6)  5 (6)  4 (5)  1.00  Previous cardiac surgery  17 (21)  9 (11)  8 (10)  1.00  Aetiology   Postcardiotomy shock  53 (65)  27 (33)  26 (32)  1.00   Acute myocardial infarction  18 (22)  10 (12)  8 (10)  0.79   Primary graft dysfunction  0 (0)  0 (0)  0 (0)  1.00   Decompensated heart failure  6 (7)  3 (4)  3 (4)  1.00   Other (e.g. sepsis, myocarditis)  5 (6)  1 (1)  4 (5)  0.36  Length of support (days)  4.1 (1.9–7.3)  5.6 (3.2–7.5)  3.7 (1.1–7.1)  0.077  Intensive care unit length of stay (days)  12 (4–25)  19 (8–35)  8 (3–18)  0.020  Hospital length of stay (days)  22 (8–42)  28 (18–46)  12 (4–32)  0.059  Major bleeding event  2 (2)  1 (1)  1 (1)  1.00  In-hospital mortality  46 (56)  18 (43)  28 (68)  0.022  Data are presented as median (interquartile range) or frequency (%). Table 4: Characteristics and outcomes of propensity-matched groups   Overall (n = 82)  Age (years)   ≤ 72 (n =41)  >72 (n = 41)  P-value  Age (years)  72 (62–78)  62 (54–67)  78 (74–80)  <0.001  Female gender  24 (29)  9 (22)  15 (36)  0.22  Body mass index (kg/m2)  28.0 (24.6–33.0)  28.2 (25.0–33.9)  27.9 (23.7–31.9)  0.91  Hypertension  68 (83)  34 (41)  34 (41)  1.00  Hyperlipidaemia  54 (66)  27 (33)  27 (33)  1.00  Coronary artery disease  55 (67)  28 (34)  27 (33)  1.00  Chronic obstructive pulmonary disease  8 (10)  3 (4)  5 (6)  0.71  Diabetes mellitus  34 (41)  19 (23)  15 (18)  0.50  Prior cerebrovasculr accident  9 (11)  4 (5)  5 (6)  1.00  Chronic kidney disease  32 (39)  16 (20)  16 (20)  1.00  Smoking  9 (6)  5 (6)  4 (5)  1.00  Previous cardiac surgery  17 (21)  9 (11)  8 (10)  1.00  Aetiology   Postcardiotomy shock  53 (65)  27 (33)  26 (32)  1.00   Acute myocardial infarction  18 (22)  10 (12)  8 (10)  0.79   Primary graft dysfunction  0 (0)  0 (0)  0 (0)  1.00   Decompensated heart failure  6 (7)  3 (4)  3 (4)  1.00   Other (e.g. sepsis, myocarditis)  5 (6)  1 (1)  4 (5)  0.36  Length of support (days)  4.1 (1.9–7.3)  5.6 (3.2–7.5)  3.7 (1.1–7.1)  0.077  Intensive care unit length of stay (days)  12 (4–25)  19 (8–35)  8 (3–18)  0.020  Hospital length of stay (days)  22 (8–42)  28 (18–46)  12 (4–32)  0.059  Major bleeding event  2 (2)  1 (1)  1 (1)  1.00  In-hospital mortality  46 (56)  18 (43)  28 (68)  0.022    Overall (n = 82)  Age (years)   ≤ 72 (n =41)  >72 (n = 41)  P-value  Age (years)  72 (62–78)  62 (54–67)  78 (74–80)  <0.001  Female gender  24 (29)  9 (22)  15 (36)  0.22  Body mass index (kg/m2)  28.0 (24.6–33.0)  28.2 (25.0–33.9)  27.9 (23.7–31.9)  0.91  Hypertension  68 (83)  34 (41)  34 (41)  1.00  Hyperlipidaemia  54 (66)  27 (33)  27 (33)  1.00  Coronary artery disease  55 (67)  28 (34)  27 (33)  1.00  Chronic obstructive pulmonary disease  8 (10)  3 (4)  5 (6)  0.71  Diabetes mellitus  34 (41)  19 (23)  15 (18)  0.50  Prior cerebrovasculr accident  9 (11)  4 (5)  5 (6)  1.00  Chronic kidney disease  32 (39)  16 (20)  16 (20)  1.00  Smoking  9 (6)  5 (6)  4 (5)  1.00  Previous cardiac surgery  17 (21)  9 (11)  8 (10)  1.00  Aetiology   Postcardiotomy shock  53 (65)  27 (33)  26 (32)  1.00   Acute myocardial infarction  18 (22)  10 (12)  8 (10)  0.79   Primary graft dysfunction  0 (0)  0 (0)  0 (0)  1.00   Decompensated heart failure  6 (7)  3 (4)  3 (4)  1.00   Other (e.g. sepsis, myocarditis)  5 (6)  1 (1)  4 (5)  0.36  Length of support (days)  4.1 (1.9–7.3)  5.6 (3.2–7.5)  3.7 (1.1–7.1)  0.077  Intensive care unit length of stay (days)  12 (4–25)  19 (8–35)  8 (3–18)  0.020  Hospital length of stay (days)  22 (8–42)  28 (18–46)  12 (4–32)  0.059  Major bleeding event  2 (2)  1 (1)  1 (1)  1.00  In-hospital mortality  46 (56)  18 (43)  28 (68)  0.022  Data are presented as median (interquartile range) or frequency (%). DISCUSSION The findings of this study are summarized as follows: (i) age has an impact on mortality after age 63; (ii) ECLS in-hospital mortality dramatically increased for patients over 72 years old to nearly 70% and (iii) an age over 72 years was an independent predictor of in-hospital mortality when adjusted for the higher number of comorbidities in this population. ECLS is the most aggressive form of cardiopulmonary resuscitation for cardiogenic shock. Advancements in extracorporeal device technology have made this therapeutic modality more accessible and a rapidly growing body of literature has demonstrated an exponential increase in the use of VA-ECMO [15–18]. While there has been a suggestion that, in some circumstances, the use of extracorporeal support serves as a ‘bridge to nowhere’ [19], as cardiac surgical patients and the population at-large ages, so too does the potential need for VA-ECMO in postcardiotomy shock, heart failure, myocardial infarction and other aetiologies of refractory cardiac failure. In this study, we included patients requiring ECMO for a variety of reasons, most commonly for postcardiotomy shock and AMI, with a particular focus on a controversial demographic—patients of advanced age. Saxena et al. [20] described the Mayo Clinic experience with 45 postcardiotomy shock patients over the age of 70 and found an in-hospital mortality of 75.6%. Their conclusion was that perioperative cardiogenic shock in the elderly requiring extracorporeal support portends a poor prognosis but that age itself should not preclude patients from ECMO candidacy for postcardiotomy shock. Our mortality in these patients over 72 years of age was 65% and we are inclined to agree. Although postcardiotomy shock patients are acutely and severely ill due to surgical trauma and subsequent related complications, patients undergoing non-emergent cardiac surgery are presumably deemed, preoperatively, to be capable of withstanding the stress of a surgical intervention and thereby may have the reserve to tolerate extracorporeal support. Compared with postcardiotomy shock patients, elderly patients with AMI who were deemed to be suitable candidates for VA-ECMO had a worse morality rate—reaching 80%. Although subgroup analyses of shock aetiology-based outcomes are beyond the scope of this article, this observation is interesting, given that the existing literature cites postcardiotomy shock as having the worst outcome compared with cardiogenic shock from other aetiologies [14, 21, 22]. In cases of AMI, patients are often in cardiopulmonary arrest when they arrive at the hospital, and clinicians have insufficient time and clinical information about the patient’s risk factors to make well-informed decisions. Within our old age cohort, close to 40% of AMI patients underwent extracorporeal cardiopulmonary resuscitation and mortality occurred in two-thirds of these patients. An extracorporeal life support organization (ELSO) registry database investigation by Mendiratta et al. [23] demonstrated that extracorporeal cardiopulmonary resuscitation in patients of advanced age (median age 70 years) was associated with 78% mortality. Again, these results do not preclude all patients of advanced age from cannulation but may better inform emergent decision-making. Previous studies have suggested that age itself, regardless of comorbidities, should not preclude patients from being candidates for ECLS with VA-ECMO [1, 20, 24]. While we are not suggesting that advanced age be a contraindication to VA-ECMO, it is notable that mortality rates in our study accelerated in >72-year-old group to nearly 70% and reached 80% in those with AMI. Moreover, both our multivariable logistic regression and propensity-matched cohort comparisons confirmed that advanced age was an influential predictor of in-hospital mortality in patients on VA-ECMO. Advanced age patients are likely to have risk profiles that were not measured in this study such as functional capacity, physiological reserve and disqualification from more definitive cardiac replacement therapies. Many of these variables are not easily measurable in the current health care setting. However, the evaluation of frailty has become the standard practice in transcatheter aortic valve replacement, where its presence may adversely impact long-term post-procedural outcomes [25, 26]. Frailty has not been well defined in the extracorporeal support literature, but it is imperative to be able to identify such patients, particularly those of older age for whom the prospect of mortality is already high to better guide decision-making. As the field of mechanical circulatory support continues to advance, careful patient selection among the elderly population will become increasingly important. While biological age, alone, should be used cautiously—merely as one of multiple clinically relevant factors in the decision-making process—it is important to remember that advanced age is frequently accompanied by other influential comorbidities. This high level of collinearity should be taken into careful consideration at VA-ECMO consultation in which the initiation decision must be made with limited information and time. Future studies will need to develop risk stratification models to delineate those elderly patients who will benefit from ECLS and for those in whom support is futile. Limitations There are several limitations to this investigation. First, the age cut-off is data driven and has not been validated either internally or externally. In addition, using dichotomized age instead of continuous variables might result in loss of information. Second, the retrospective nature of this study subjects it to limitations inherent in all observational investigations. Third, these data were derived from a single, high-volume institution with limited sample sizes, which may limit overall generalizability and the feasibility to draw conclusions. We also lack follow-up quality of life and functional data on patients at discharge, which limits the scope of post-discharge outcomes analysis. Finally, conclusions drawn on our global in-hospital mortality data might minimize aetiology-specific differences among patients. We have tried to control for this by adjusting for the aetiologies significantly associated with mortality in our multivariable regression model. CONCLUSION In conclusion, ECLS offers a means of resuscitation for patients experiencing refractory cardiogenic shock but is associated with considerable mortality in patients of advanced age. Whether this mortality rate is prohibitive for this therapy will be better informed by future multi-institutional trials though it will, ultimately, be up to the health care team, which must consider each patient’s clinical scenario to determine how age might impact the most appropriate therapeutic approach. Further studies are needed to assess the effect of advanced age as well as the impact of cost and long-term functional outcomes. Conflict of interest: none declared. REFERENCES 1 Massetti M, Tasle M, Le Page O, Deredec R, Babatasi G, Buklas D et al.   Back from irreversibility: extracorporeal life support for prolonged cardiac arrest. Ann Thorac Surg  2005; 79: 178– 83; discussion 183–174. Google Scholar CrossRef Search ADS PubMed  2 Kolla S, Lee WA, Hirschl RB, Bartlett RH. Extracorporeal life support for cardiovascular support in adults. ASAIO J  1996; 42: M809– 19. Google Scholar CrossRef Search ADS PubMed  3 Rosenbaum AN, John R, Liao KK, Adatya S, Colvin-Adams MM, Pritzker M et al.   Survival in elderly patients supported with continuous flow left ventricular assist device as bridge to transplantation or destination therapy. J Card Fail  2014; 20: 161– 7. Google Scholar CrossRef Search ADS PubMed  4 Atluri P, Goldstone AB, Kobrin DM, Cohen JE, MacArthur JW, Howard JL et al.   Ventricular assist device implant in the elderly is associated with increased, but respectable risk: a multi-institutional study. Ann Thorac Surg  2013; 96: 141– 7. Google Scholar CrossRef Search ADS PubMed  5 Kilic A, Sultan I, Yuh DD, Shah AS, Baumgartner WA, Cameron DE et al.   Ventricular assist device implantation in the elderly: nationwide outcomes in the United States. J Card Surg  2013; 28: 183– 9. Google Scholar CrossRef Search ADS PubMed  6 Safar P, Abramson NS, Angelos M, Cantadore R, Leonov Y, Levine R et al.   Emergency cardiopulmonary bypass for resuscitation from prolonged cardiac arrest. Am J Emerg Med  1990; 8: 55– 67. Google Scholar CrossRef Search ADS PubMed  7 Hill JG, Bruhn PS, Cohen SE, Gallagher MW, Manart F, Moore CA et al.   Emergent applications of cardiopulmonary support: a multiinstitutional experience. Ann Thorac Surg  1992; 54: 699– 704. Google Scholar CrossRef Search ADS PubMed  8 Magovern GJJr, Simpson KA. Extracorporeal membrane oxygenation for adult cardiac support: the allegheny experience. Ann Thorac Surg  1999; 68: 655– 61. Google Scholar CrossRef Search ADS PubMed  9 Nelson S, Ramakrishnan V, Nietert PJ, Kamen DL, Ramos PS, Wolf BJ. An evaluation of common methods for dichotomization of continuous variables to discriminate disease status. Comm Stat Theory Methods  2016; 46: 23– 34. 10 Takayama H, Truby L, Koekort M, Uriel N, Colombo P, Mancini DM et al.   Clinical outcome of mechanical circulatory support for refractory cardiogenic shock in the current era. J Heart Lung Transplant  2013; 32: 106– 11. Google Scholar CrossRef Search ADS PubMed  11 Garan AR, Kirtane A, Takayama H. Redesigning care for patients with acute myocardial infarction complicated by cardiogenic shock: the ‘Shock Team’. JAMA Surg  2016; 151: 684– 85. Google Scholar CrossRef Search ADS PubMed  12 Truby LK, Takeda K, Mauro C, Yuzefpolskaya M, Garan AR, Kirtane AJ et al.   Incidence and implications of left ventricular distention during venoarterial extracorporeal membrane oxygenation support. ASAIO J  2017; 63: 257– 65. Google Scholar CrossRef Search ADS PubMed  13 Takeda K, Garan AR, Ando M, Han J, Topkara VK, Kurlansky P et al.   Minimally invasive CentriMag ventricular assist device support integrated with extracorporeal membrane oxygenation in cardiogenic shock patients: a comparison with conventional CentriMag biventricular support configuration. Eur J Cardiothorac Surg  2017; 52: 1055– 61. Google Scholar CrossRef Search ADS PubMed  14 Takayama H, Soni L, Kalesan B, Truby LK, Ota T, Cedola S et al.   Bridge-to-decision therapy with a continuous-flow external ventricular assist device in refractory cardiogenic shock of various causes. Circ Heart Fail  2014; 7: 799– 806. Google Scholar CrossRef Search ADS PubMed  15 Stretch R, Sauer CM, Yuh DD, Bonde P. National trends in the utilization of short-term mechanical circulatory support: incidence, outcomes, and cost analysis. J Am Coll Cardiol  2014; 64: 1407– 15. Google Scholar CrossRef Search ADS PubMed  16 Truby L, Mundy L, Kalesan B, Kirtane A, Colombo PC, Takeda K et al.   Contemporary outcomes of venoarterial extracorporeal membrane oxygenation for refractory cardiogenic shock at a large tertiary care center. ASAIO J  2015; 61: 403– 9. Google Scholar CrossRef Search ADS PubMed  17 Karagiannidis C, Brodie D, Strassmann S, Stoelben E, Philipp A, Bein T et al.   Extracorporeal membrane oxygenation: evolving epidemiology and mortality. Intensive Care Med  2016; 42: 889– 96. Google Scholar CrossRef Search ADS PubMed  18 Guenther SP, Brunner S, Born F, Fischer M, Schramm R, Pichlmaier M et al.   When all else fails: extracorporeal life support in therapy-refractory cardiogenic shock. Eur J Cardiothorac Surg  2016; 49: 802– 9. Google Scholar CrossRef Search ADS PubMed  19 Abrams DC, Prager K, Blinderman CD, Burkart KM, Brodie D. Ethical dilemmas encountered with the use of extracorporeal membrane oxygenation in adults. Chest  2014; 145: 876– 82. Google Scholar CrossRef Search ADS PubMed  20 Saxena P, Neal J, Joyce LD, Greason KL, Schaff HV, Guru P et al.   Extracorporeal membrane oxygenation support in postcardiotomy elderly patients: the Mayo Clinic experience. Ann Thorac Surg  2015; 99: 2053– 60. Google Scholar CrossRef Search ADS PubMed  21 Mohite PN, Sabashnikov A, Patil NP, Saez DG, Zych B, Popov AF et al.   Short-term ventricular assist device in post-cardiotomy cardiogenic shock: factors influencing survival. J Artif Organs  2014; 17: 228– 35. Google Scholar CrossRef Search ADS PubMed  22 Pokersnik JA, Buda T, Bashour CA, Gonzalez-Stawinski GV. Have changes in ECMO technology impacted outcomes in adult patients developing postcardiotomy cardiogenic shock? J Card Surg  2012; 27: 246– 52. Google Scholar CrossRef Search ADS PubMed  23 Mendiratta P, Wei JY, Gomez A, Podrazik P, Riggs AT, Rycus P et al.   Cardiopulmonary resuscitation requiring extracorporeal membrane oxygenation in the elderly: a review of the Extracorporeal Life Support Organization Registry. ASAIO J  2013; 59: 211– 5. Google Scholar CrossRef Search ADS PubMed  24 Saito S, Nakatani T, Kobayashi J, Tagusari O, Bando K, Niwaya K et al.   Is extracorporeal life support contraindicated in elderly patients? Ann Thorac Surg  2007; 83: 140– 5. Google Scholar CrossRef Search ADS PubMed  25 Green P, Woglom AE, Genereux P, Daneault B, Paradis JM, Schnell S et al.   The impact of frailty status on survival after transcatheter aortic valve replacement in older adults with severe aortic stenosis: a single-center experience. JACC Cardiovasc Interv  2012; 5: 974– 81. Google Scholar CrossRef Search ADS PubMed  26 Huded CP, Huded JM, Friedman JL, Benck LR, Lindquist LA, Holly TA et al.   Frailty status and outcomes after transcatheter aortic valve implantation. Am J Cardiol  2016; 117: 1966– 71. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

European Journal of Cardio-Thoracic SurgeryOxford University Press

Published: Jan 22, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off