ASAIO Journal 2014
Adult Circulatory Support
Noncardiac Surgery in Patients on Mechanical Circulatory Support Sharven Taghavi,* Carl Beyer,† Halley Vora,† Senthil N. Jayarajan,* Yoshiya Toyoda,‡ Jay Dujon,* Lars O. Sjoholm,* Abhijit Pathak,* Thomas A. Santora,* Amy J. Goldberg,* and Joseph F. Rappold*
This study examined outcomes in patients with left ventricular assist device (LVAD) and extracorporeal membrane oxygenation (ECMO) requiring noncardiac surgical procedures and identified factors that influence outcomes. All patients with mechanical circulatory support (MCS) devices at our institution from 2002 to 2013 undergoing noncardiac surgical procedures were reviewed. There were 148 patients requiring MCS during the study period, with 40 (27.0%) requiring 62 noncardiac surgical procedures. Of these, 29 (72.5%) had implantable LVAD and 11 (27.5%) were supported with ECMO. The two groups were evenly matched with regard to age (53.6 vs. 54.5 years, p = 0.87), male sex (71.4 vs. 45.5%, p = 0.16), and baseline creatinine (1.55 vs. 1.43 mg/dl, p = 0.76). Patients on ECMO had greater demand for postoperative blood products (0.8 vs. 2.8 units of packed red blood cells, p = 0.002) and greater postoperative increase in creatinine (0.07 vs. 0.44 mg/dl, p = 0.047). Median survival was markedly worse in ECMO patients. Factors associated with mortality included ECMO support, history of biventricular assist device, and postoperative blood transfusion. Preoperative aspirin was associated with survival. These findings demonstrate the importance of careful surgical hemostasis and minimizing perioperative blood transfusions in patients on MCS undergoing noncardiac surgical procedures. In addition, low-dose antiplatelet therapy should be continued perioperatively. ASAIO Journal 2014; 60:670–674.
long-term left ventricular assist devices (LVADs) can be used as a bridge to transplantation (BTT) or for destination therapy.1,4,5 As more patients are supported with MCS devices, many will go on to develop noncardiac surgical problems that require surgical intervention.6,7 No studies to date have examined noncardiac surgery (NCS) in patients supported with temporary ECMO. Furthermore, studies to guide the appropriate perioperative management of LVAD patients undergoing noncardiac surgical procedures are lacking.8 The goal of this study was to identify variables that influence outcomes in patients undergoing NCS supported with ECMO or LVAD. Materials and Methods After approval from the local institutional review board, all patients receiving MCS at Temple University Hospital from January 2002 to December 2012 were identified using our prospectively collected MCS database. Medical records of all patients were identified and retrospectively reviewed. Data collected included patient demographics, patient medical history, baseline laboratory values, preimplantation treatment strategy (BTT or destination therapy), and duration of preoperative MCS. Postoperative end points included survival, perioperative mortality (within 30 days), total length of stay, intensive care unit (ICU) length of stay, duration of mechanical ventilation and vasopressor support, change in creatinine, and requirement of perioperative blood transfusion.
Key Words: mechanical circulatory support, extracorporeal membrane oxygenation, left ventricular assist device, surgery
Statistical Analysis
Mechanical circulatory support (MCS) devices are increas-
All data analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC). Continuous variables are presented as mean ±- standard deviation or median and categorical variables are reported as percentages of the total number of data points available for that field. Student’s t-test, Wilcoxon rank sum test, and Fisher’s exact test were used to examine continuous and categorical variables as appropriate. Kaplan–Meier analysis was used to estimate survival and compared with the log-rank test. A p value less than 0.05 was considered significant.
ingly used to treat patients with end-stage heart failure. Temporary devices, such as right ventricular assist devices, biventricular assist devices, and extracorporeal membrane oxygenation (ECMO), can be used for support of critically ill patients in cardiogenic shock.2,3 In addition, fully implantable, 1
From the *Department of Surgery, Temple University School of Medicine, Philadelphia, Pennsylvania; †Temple University School of Medicine, Philadelphia, Pennsylvania; and ‡Division of Cardiac Surgery, Temple University School of Medicine, Philadelphia, Pennsylvania. Submitted for consideration May 2014; accepted for publication in revised form August 2014. Disclosure: The authors have no conflicts of interest to report. Presented at the 9th Annual Academic Surgical Congress; February 4–6, 2014; San Diego, CA. Correspondence: Joseph F. Rappold, MD, FACS, Temple University Hospital, 3401 N. Broad Street, Parkinson Pavilion, Suite 400, Philadelphia, PA 19140. Email: [email protected]. Copyright © 2014 by the American Society for Artificial Internal Organs
Results Baseline Patient Characteristics There were 148 consecutive patients requiring MCS during the study period. Of these, 40 (27.0%) required 62 noncardiac surgical procedures. There were 29 (72.5%) patients that had fully implantable LVAD and 11 (27.5%) patients were supported with temporary ECMO. For patients with implantable LVAD, 24 (82.8%) had newer generation, continuous-flow devices, whereas 5 (17.2%) had first-generation, pulsatile flow devices.
DOI: 10.1097/MAT.0000000000000140
670
671
PROCEDURES IN ECMO AND LVAD PATIENTS
All of the continuous-flow devices used were Heartmate-2 (Thoratec Corporation, Pleasanton, CA). All of the older generation devices implanted were the Heartmate XVE (Thoratec Corporation). Clinical characteristics of the two cohorts are shown in Tables 1 and 2. The two groups were evenly matched with respect to age, sex, white race, and baseline creatinine. The LVAD cohort was more likely to be black (50.0 vs. 0.0%, p = 0.003) and be on preoperative aspirin (71.4 vs. 9.1%, p = 0.001). The ECMO group had significantly shorter median duration of preoperative MCS in the ECMO group (20.0 vs. 1.0 days, p = 0.002). Eight (72.3%) patients supported by ECMO were in fulminant cardiac and pulmonary failure, requiring veno-arterial ECMO. Patients in the LVAD cohort were equally spaced out during the 10 year study period. However, the vast majority of patients requiring ECMO (n = 9, 81.8%) were supported in 2009 or later. Types of Surgical Procedures Types of noncardiac surgical procedures are shown in Table 3. The most common operations performed in the LVAD group were abdominal exploration/bowel resection (n = 18, 38.3%) and tracheostomy (n = 14, 29.8%). The most common procedure performed on patients requiring ECMO were tracheostomy (n = 5, 33.3%), extremity/vascular operations (n = 4, 26.7%), and abdominal exploration or bowel resection (n = 4, 26.7%). All patients requiring abdominal exploration presented with acute abdominal pain, nausea and emesis, or intolerance to oral intake. For patients that were intubated or sedated, this manifested as bloody bowel movements or intolerance to tube feeds with or without hemodynamic compromise. In addition, we observed elevated serum lactate levels in all patients requiring bowel resection due to infarcted bowel (n = 9) or hernia repair due to strangulated bowel (n = 3). Management of Anticoagulation In the LVAD cohort, there were five (17.9%) patients that were actively on Coumadin at the time of surgery. These patients were operated on emergently and received fresh frozen plasma and vitamin K for a goal International normalized ratio (INR) < 1.5 before going to the operating room. The remaining patients were operated on electively and were started on a heparin infusion with a partial thromboplastin time (PTT) of 60–80 seconds. The heparin infusion was continued until INR normalized. The heparin infusion was stopped approximately 6 hours Table 1. Baseline Clinical Characteristics of Left Ventricular Assist Device Cohort (N = 28) Mean age (years) Male, n (%) White, n (%) Black, n (%) Hispanic, n (%) Inpatient at time of surgery, n (%) Median days of mechanical circulatory support before surgery, n (%) Preoperative coumadin, n (%) Mean preoperative INR, n (%) Preoperative aspirin, n (%) Baseline creatinine (mg/dl) INR, International normalized ratio.
53.6 ± 14.3 20 (71.4) 9 (32.1) 14 (50.0) 0 (0.0) 27 (96.4) 20.0 (9.5–75.0) 5 (17.9) 1.50 ± 0.46 20 (71.4) 1.55 ± 0.93
Table 2. Baseline Clinical Characteristics for Extracorporeal Membrane Oxygenation Cohort (N = 11) Mean age (y) Male, n (%) White, n (%) Black, n (%) Hispanic, n (%) Inpatient at time of surgery, n (%) Median days of mechanical circulatory support before surgery, n (%) Preoperative Coumadin, n (%) Mean preoperative INR, n (%) Preoperative aspirin, n (%) Baseline creatinine (mg/dl)
54.5 ± 13.8 5 (45.5) 7 (63.6) 0 (0.0) 1 (9.1) 11 (100.0) 1.0 (1.0–8.0) 0 (0.0) 1.22 ± 0.33 1 (9.1) 1.43 ± 1.46
before incision and restarted 6 hours after surgery. There was no incidence of pump thrombosis in the LVAD cohort. There were no patients on Coumadin at the time of surgery in the ECMO cohort. All patients were on a heparin infusion before surgery. The heparin infusion was held immediately before surgery and resumed immediately after surgery. Postoperative Outcomes Postoperative outcomes are summarized in Table 4. There was no significant difference in total hospital or ICU lengths of stay. Patients in the ECMO cohort were more likely to require postoperative mechanical ventilation (68.1 vs. 100.0%, p = 0.01). In addition, patients on ECMO had a greater increase in postoperative creatinine (0.07 vs. 0.44, p = 0.047). The ECMO cohort also had a higher requirement for blood transfusion within 24 hours of the procedure (25.5 vs. 73.3%, p = 0.002) and received more packed red blood cells within 24 hours of surgery (0.8 vs. 2.8 units, p = 0.003). There was no significant difference in the requirement of postoperative vasopressor support or its duration between the two groups. Perioperative mortality was also markedly worse in the ECMO supported cohort (6.4 vs. 46.7%, p = 0.001). Survival Univariate logistic regression survival is shown in Table 5. Factors associated with mortality were support by ECMO (hazard ratio [HR]: 2.90, 95% confidence interval [CI] 1.46–5.78, p = 0.002), history of biventricular support (HR 2.02, 95% CI 1.01–4.04, p = 0.049), and requirement for postoperative blood transfusion (HR 3.14, 95% CI 1.66–5.93, p < 0.001). The use of preoperative aspirin was associated with survival (HR 0.48, 95% CI 0.26–0.89, p = 0.02). Age, sex, race, history of ischemic cardiomyopathy, diabetes, hyperlipidemia, chronic lung disease, and chronic renal insufficiency were not associated with increased mortality. In addition, preoperative warfarin use, preoperative INR, preoperative creatinine, and postoperative creatinine were not associated with mortality risk. A comparison of median survival by Kaplan–Meier analysis is shown in Figure 1. Survival was significantly higher in the LVAD group as compared with the ECMO group (142.5 vs. 6.0 days, p = 0.002). Discussion The advent of newer generation, continuous-flow LVADs has led to significant decrease in morbidity and mortality for
672 TAGHAVI et al. Table 3. Type of Noncardiac Surgical Procedures Procedure, n (%) Abdominal exploration/bowel resection Tracheostomy Extremity/vascular surgery Urological procedure Gynecological surgery Oral surgery Other surgery
Total Number (N = 62)
Left Ventricular Assist Device Cohort (N = 47)
Extracorporeal Membrane Oxygenation Cohort (N = 15)
22 (35.5) 19 (30.6) 9 (14.5) 2 (3.2) 1 (1.6) 2 (3.2) 7 (11.3)
18 (38.3) 14 (29.8) 5 (10.6) 2 (4.3) 0 (0.0) 2 (4.3) 6 (12.8)
4 (26.7) 5 (33.3) 4 (26.7) 0 (0.0) 1 (6.7) 0 (0.0) 1 (6.7)
patients with end-stage heart failure requiring MCS.9,10 With improved device reliability and durability, the number of noncardiac surgical procedures in patients with LVAD is on the rise.8,11 Studies on noncardiac surgical procedures in patients with implantable LVAD have been limited. In addition, no study to date has examined noncardiac surgical procedures on patients requiring temporary ECMO. Although patients requiring ECMO are a completely different patient population when compared with those requiring LVAD, our goal was not to compare the two patient populations. Rather, the goal of this study was to examine clinical factors that influence outcomes in patients with end-stage heart failure on MCS that required noncardiac surgical procedures. This study demonstrates that noncardiac surgical procedures can be safely performed in patients with implantable LVAD as we found a low perioperative mortality rate (6.2%) in the LVAD cohort. This finding has been supported by previous studies that have shown that noncardiac surgical procedures can be performed in this group of patients with good outcomes.6,8,11–13 All three perioperative deaths observed in this study occurred in critically ill and ventilator-dependent ICU patients. Two of these patients received urological procedures and one received a tracheostomy. As seen in the current study and corroborated by others,7,12 abdominal operations are the most frequent noncardiac surgical procedure seen in LVAD patients. For patients with implantable LVAD, indications for abdominal operations are usually unrelated to the LVAD and are more likely to be carried out because of unrelated conditions such as cholecystitis or diverticulitis.7,12 Abdominal exploration was the most frequent procedure seen in our study and clinicians must pay close attention to change in abdominal exam and the development of acute pain or nausea and vomiting. There must be an especially high clinical index of suspicion when there is hemodynamic compromise and an elevated serum lactate level in addition to worsening abdominal exam. Driveline infection is
also a known problem in patients with implantable LVAD.14,15 In this study, there were two driveline infections requiring surgical intervention; however, these infections did not result in intraabdominal procedures. However, this study represents the first to examine outcomes with noncardiac surgical procedures in patients supported with temporary ECMO. Although results for noncardiac surgical procedures in patients with implantable LVAD appear to be good, outcomes for patients on ECMO are poor. In the current study, almost half of patients in the ECMO cohort requiring noncardiac surgical procedures died within 30 days. This is not unexpected, because patients requiring temporary ECMO are known to have high mortality risk.16,17 The most common surgical procedure in the ECMO group was placement of a tracheostomy (n = 5), extremity/vascular procedures (n = 4), and abdominal exploration/bowel resection (n = 4). Vascular complications are a known risk of arterial cannulation in ECMO and can often require surgical intervention.17 In the current study, all patients requiring vascular/extremity procedures received peripheral femoral artery cannulation. As experience with ECMO increases and outcomes improve,19 more and more patients on ECMO will require noncardiac surgical procedures. Further studies are needed to improve outcomes in this very high-risk group. Because of the risk of thromboembolic complications,20 patients with LVAD and ECMO are usually anticoagulated and on aspirin.10,21 This presents a conundrum to surgeons as they are faced with high risk of bleeding versus inducing a thromboembolic complication. Current guidelines for patients with LVAD recommend maintaining an INR between 1.5 and 2.5 to minimize pump thrombosis and thromboembolic complications.10,21 Because of the need for anticoagulation in this patient population, perioperative bleeding is a known complication during noncardiac surgical procedures11,23 and was found to be problematic in the current study. In the current study, 25.5%
Table 4. Postoperative Outcomes
Mean total length of stay (d) Mean intensive care unit stay (hr) Requirement of postoperative mechanical ventilation, n (%) Duration mechanical ventilation (hr) Requirement of postoperative vasopressor support, n (%) Time requiring vasopressor support (hr) Increase in creatinine (mg/dl) Requirement of blood transfusion within 24 hours of surgery, n (%) Average packed red blood cells transfused within 24 hours Perioperative mortality, n (%)
Left Ventricular Assist Device (N = 47)
Extracorporeal Membrane Oxygenation (N = 15)
p Value
37.0 ± 33.6 28.7 ± 33.2 32 (68.1) 216.7 ± 277.0 19 (36.2) 228.6 ± 653.1 0.07 ± 0.57 12 (25.5) 0.8 ± 1.7 3 (6.4)
30.1 ± 42.2 24.9 ± 38.8 15 (100.0) 347.1 ± 447.8 9 (19.1) 48.7 ± 78.8 0.44 ± 0.67 11 (73.3) 2.8 ± 3.2 7 (46.7)
0.52 0.71 0.01 0.19 0.24 0.29 0.047 0.002 0.003 0.001
673
PROCEDURES IN ECMO AND LVAD PATIENTS Table 5. Univariate Regression Analysis for Survival
Age Male White Black Hispanic Extracorporeal membrane oxygenation History of biventricular assist device Ischemic cardiomyopathy Preoperative coumadin Preoperative INR Preoperative aspirin History of hyperlipidemia History of diabetes History of hyperlipidemia History of chronic lung disease History of chronic renal insufficiency Postoperative blood transfusion Preoperative creatinine Postoperative creatinine
Hazard Ratio
95% Confidence Interval
p Value
0.99 0.74 0.80 0.81 1.07 2.90 2.02 0.75 0.59 0.85 0.48 0.65 0.97 0.65 1.41 0.85 3.14 1.00 0.95
0.97–1.01 0.41–1.35 0.42–1.54 0.44–1.49 0.33–3.51 1.46–5.78 1.01–4.04 0.40–1.41 0.21–1.66 0.53–1.38 0.26–0.89 0.35–1.20 0.53–1.76 0.35–1.20 0.70–2.87 0.42–1.69 1.66–5.93 0.76–1.34 0.75–1.22
0.74 0.33 0.51 0.50 0.91 0.002 0.049 0.78 0.32 0.52 0.02 0.17 0.91 0.17 0.34 0.63
Adult Circulatory Support
Noncardiac Surgery in Patients on Mechanical Circulatory Support Sharven Taghavi,* Carl Beyer,† Halley Vora,† Senthil N. Jayarajan,* Yoshiya Toyoda,‡ Jay Dujon,* Lars O. Sjoholm,* Abhijit Pathak,* Thomas A. Santora,* Amy J. Goldberg,* and Joseph F. Rappold*
This study examined outcomes in patients with left ventricular assist device (LVAD) and extracorporeal membrane oxygenation (ECMO) requiring noncardiac surgical procedures and identified factors that influence outcomes. All patients with mechanical circulatory support (MCS) devices at our institution from 2002 to 2013 undergoing noncardiac surgical procedures were reviewed. There were 148 patients requiring MCS during the study period, with 40 (27.0%) requiring 62 noncardiac surgical procedures. Of these, 29 (72.5%) had implantable LVAD and 11 (27.5%) were supported with ECMO. The two groups were evenly matched with regard to age (53.6 vs. 54.5 years, p = 0.87), male sex (71.4 vs. 45.5%, p = 0.16), and baseline creatinine (1.55 vs. 1.43 mg/dl, p = 0.76). Patients on ECMO had greater demand for postoperative blood products (0.8 vs. 2.8 units of packed red blood cells, p = 0.002) and greater postoperative increase in creatinine (0.07 vs. 0.44 mg/dl, p = 0.047). Median survival was markedly worse in ECMO patients. Factors associated with mortality included ECMO support, history of biventricular assist device, and postoperative blood transfusion. Preoperative aspirin was associated with survival. These findings demonstrate the importance of careful surgical hemostasis and minimizing perioperative blood transfusions in patients on MCS undergoing noncardiac surgical procedures. In addition, low-dose antiplatelet therapy should be continued perioperatively. ASAIO Journal 2014; 60:670–674.
long-term left ventricular assist devices (LVADs) can be used as a bridge to transplantation (BTT) or for destination therapy.1,4,5 As more patients are supported with MCS devices, many will go on to develop noncardiac surgical problems that require surgical intervention.6,7 No studies to date have examined noncardiac surgery (NCS) in patients supported with temporary ECMO. Furthermore, studies to guide the appropriate perioperative management of LVAD patients undergoing noncardiac surgical procedures are lacking.8 The goal of this study was to identify variables that influence outcomes in patients undergoing NCS supported with ECMO or LVAD. Materials and Methods After approval from the local institutional review board, all patients receiving MCS at Temple University Hospital from January 2002 to December 2012 were identified using our prospectively collected MCS database. Medical records of all patients were identified and retrospectively reviewed. Data collected included patient demographics, patient medical history, baseline laboratory values, preimplantation treatment strategy (BTT or destination therapy), and duration of preoperative MCS. Postoperative end points included survival, perioperative mortality (within 30 days), total length of stay, intensive care unit (ICU) length of stay, duration of mechanical ventilation and vasopressor support, change in creatinine, and requirement of perioperative blood transfusion.
Key Words: mechanical circulatory support, extracorporeal membrane oxygenation, left ventricular assist device, surgery
Statistical Analysis
Mechanical circulatory support (MCS) devices are increas-
All data analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC). Continuous variables are presented as mean ±- standard deviation or median and categorical variables are reported as percentages of the total number of data points available for that field. Student’s t-test, Wilcoxon rank sum test, and Fisher’s exact test were used to examine continuous and categorical variables as appropriate. Kaplan–Meier analysis was used to estimate survival and compared with the log-rank test. A p value less than 0.05 was considered significant.
ingly used to treat patients with end-stage heart failure. Temporary devices, such as right ventricular assist devices, biventricular assist devices, and extracorporeal membrane oxygenation (ECMO), can be used for support of critically ill patients in cardiogenic shock.2,3 In addition, fully implantable, 1
From the *Department of Surgery, Temple University School of Medicine, Philadelphia, Pennsylvania; †Temple University School of Medicine, Philadelphia, Pennsylvania; and ‡Division of Cardiac Surgery, Temple University School of Medicine, Philadelphia, Pennsylvania. Submitted for consideration May 2014; accepted for publication in revised form August 2014. Disclosure: The authors have no conflicts of interest to report. Presented at the 9th Annual Academic Surgical Congress; February 4–6, 2014; San Diego, CA. Correspondence: Joseph F. Rappold, MD, FACS, Temple University Hospital, 3401 N. Broad Street, Parkinson Pavilion, Suite 400, Philadelphia, PA 19140. Email: [email protected]. Copyright © 2014 by the American Society for Artificial Internal Organs
Results Baseline Patient Characteristics There were 148 consecutive patients requiring MCS during the study period. Of these, 40 (27.0%) required 62 noncardiac surgical procedures. There were 29 (72.5%) patients that had fully implantable LVAD and 11 (27.5%) patients were supported with temporary ECMO. For patients with implantable LVAD, 24 (82.8%) had newer generation, continuous-flow devices, whereas 5 (17.2%) had first-generation, pulsatile flow devices.
DOI: 10.1097/MAT.0000000000000140
670
671
PROCEDURES IN ECMO AND LVAD PATIENTS
All of the continuous-flow devices used were Heartmate-2 (Thoratec Corporation, Pleasanton, CA). All of the older generation devices implanted were the Heartmate XVE (Thoratec Corporation). Clinical characteristics of the two cohorts are shown in Tables 1 and 2. The two groups were evenly matched with respect to age, sex, white race, and baseline creatinine. The LVAD cohort was more likely to be black (50.0 vs. 0.0%, p = 0.003) and be on preoperative aspirin (71.4 vs. 9.1%, p = 0.001). The ECMO group had significantly shorter median duration of preoperative MCS in the ECMO group (20.0 vs. 1.0 days, p = 0.002). Eight (72.3%) patients supported by ECMO were in fulminant cardiac and pulmonary failure, requiring veno-arterial ECMO. Patients in the LVAD cohort were equally spaced out during the 10 year study period. However, the vast majority of patients requiring ECMO (n = 9, 81.8%) were supported in 2009 or later. Types of Surgical Procedures Types of noncardiac surgical procedures are shown in Table 3. The most common operations performed in the LVAD group were abdominal exploration/bowel resection (n = 18, 38.3%) and tracheostomy (n = 14, 29.8%). The most common procedure performed on patients requiring ECMO were tracheostomy (n = 5, 33.3%), extremity/vascular operations (n = 4, 26.7%), and abdominal exploration or bowel resection (n = 4, 26.7%). All patients requiring abdominal exploration presented with acute abdominal pain, nausea and emesis, or intolerance to oral intake. For patients that were intubated or sedated, this manifested as bloody bowel movements or intolerance to tube feeds with or without hemodynamic compromise. In addition, we observed elevated serum lactate levels in all patients requiring bowel resection due to infarcted bowel (n = 9) or hernia repair due to strangulated bowel (n = 3). Management of Anticoagulation In the LVAD cohort, there were five (17.9%) patients that were actively on Coumadin at the time of surgery. These patients were operated on emergently and received fresh frozen plasma and vitamin K for a goal International normalized ratio (INR) < 1.5 before going to the operating room. The remaining patients were operated on electively and were started on a heparin infusion with a partial thromboplastin time (PTT) of 60–80 seconds. The heparin infusion was continued until INR normalized. The heparin infusion was stopped approximately 6 hours Table 1. Baseline Clinical Characteristics of Left Ventricular Assist Device Cohort (N = 28) Mean age (years) Male, n (%) White, n (%) Black, n (%) Hispanic, n (%) Inpatient at time of surgery, n (%) Median days of mechanical circulatory support before surgery, n (%) Preoperative coumadin, n (%) Mean preoperative INR, n (%) Preoperative aspirin, n (%) Baseline creatinine (mg/dl) INR, International normalized ratio.
53.6 ± 14.3 20 (71.4) 9 (32.1) 14 (50.0) 0 (0.0) 27 (96.4) 20.0 (9.5–75.0) 5 (17.9) 1.50 ± 0.46 20 (71.4) 1.55 ± 0.93
Table 2. Baseline Clinical Characteristics for Extracorporeal Membrane Oxygenation Cohort (N = 11) Mean age (y) Male, n (%) White, n (%) Black, n (%) Hispanic, n (%) Inpatient at time of surgery, n (%) Median days of mechanical circulatory support before surgery, n (%) Preoperative Coumadin, n (%) Mean preoperative INR, n (%) Preoperative aspirin, n (%) Baseline creatinine (mg/dl)
54.5 ± 13.8 5 (45.5) 7 (63.6) 0 (0.0) 1 (9.1) 11 (100.0) 1.0 (1.0–8.0) 0 (0.0) 1.22 ± 0.33 1 (9.1) 1.43 ± 1.46
before incision and restarted 6 hours after surgery. There was no incidence of pump thrombosis in the LVAD cohort. There were no patients on Coumadin at the time of surgery in the ECMO cohort. All patients were on a heparin infusion before surgery. The heparin infusion was held immediately before surgery and resumed immediately after surgery. Postoperative Outcomes Postoperative outcomes are summarized in Table 4. There was no significant difference in total hospital or ICU lengths of stay. Patients in the ECMO cohort were more likely to require postoperative mechanical ventilation (68.1 vs. 100.0%, p = 0.01). In addition, patients on ECMO had a greater increase in postoperative creatinine (0.07 vs. 0.44, p = 0.047). The ECMO cohort also had a higher requirement for blood transfusion within 24 hours of the procedure (25.5 vs. 73.3%, p = 0.002) and received more packed red blood cells within 24 hours of surgery (0.8 vs. 2.8 units, p = 0.003). There was no significant difference in the requirement of postoperative vasopressor support or its duration between the two groups. Perioperative mortality was also markedly worse in the ECMO supported cohort (6.4 vs. 46.7%, p = 0.001). Survival Univariate logistic regression survival is shown in Table 5. Factors associated with mortality were support by ECMO (hazard ratio [HR]: 2.90, 95% confidence interval [CI] 1.46–5.78, p = 0.002), history of biventricular support (HR 2.02, 95% CI 1.01–4.04, p = 0.049), and requirement for postoperative blood transfusion (HR 3.14, 95% CI 1.66–5.93, p < 0.001). The use of preoperative aspirin was associated with survival (HR 0.48, 95% CI 0.26–0.89, p = 0.02). Age, sex, race, history of ischemic cardiomyopathy, diabetes, hyperlipidemia, chronic lung disease, and chronic renal insufficiency were not associated with increased mortality. In addition, preoperative warfarin use, preoperative INR, preoperative creatinine, and postoperative creatinine were not associated with mortality risk. A comparison of median survival by Kaplan–Meier analysis is shown in Figure 1. Survival was significantly higher in the LVAD group as compared with the ECMO group (142.5 vs. 6.0 days, p = 0.002). Discussion The advent of newer generation, continuous-flow LVADs has led to significant decrease in morbidity and mortality for
672 TAGHAVI et al. Table 3. Type of Noncardiac Surgical Procedures Procedure, n (%) Abdominal exploration/bowel resection Tracheostomy Extremity/vascular surgery Urological procedure Gynecological surgery Oral surgery Other surgery
Total Number (N = 62)
Left Ventricular Assist Device Cohort (N = 47)
Extracorporeal Membrane Oxygenation Cohort (N = 15)
22 (35.5) 19 (30.6) 9 (14.5) 2 (3.2) 1 (1.6) 2 (3.2) 7 (11.3)
18 (38.3) 14 (29.8) 5 (10.6) 2 (4.3) 0 (0.0) 2 (4.3) 6 (12.8)
4 (26.7) 5 (33.3) 4 (26.7) 0 (0.0) 1 (6.7) 0 (0.0) 1 (6.7)
patients with end-stage heart failure requiring MCS.9,10 With improved device reliability and durability, the number of noncardiac surgical procedures in patients with LVAD is on the rise.8,11 Studies on noncardiac surgical procedures in patients with implantable LVAD have been limited. In addition, no study to date has examined noncardiac surgical procedures on patients requiring temporary ECMO. Although patients requiring ECMO are a completely different patient population when compared with those requiring LVAD, our goal was not to compare the two patient populations. Rather, the goal of this study was to examine clinical factors that influence outcomes in patients with end-stage heart failure on MCS that required noncardiac surgical procedures. This study demonstrates that noncardiac surgical procedures can be safely performed in patients with implantable LVAD as we found a low perioperative mortality rate (6.2%) in the LVAD cohort. This finding has been supported by previous studies that have shown that noncardiac surgical procedures can be performed in this group of patients with good outcomes.6,8,11–13 All three perioperative deaths observed in this study occurred in critically ill and ventilator-dependent ICU patients. Two of these patients received urological procedures and one received a tracheostomy. As seen in the current study and corroborated by others,7,12 abdominal operations are the most frequent noncardiac surgical procedure seen in LVAD patients. For patients with implantable LVAD, indications for abdominal operations are usually unrelated to the LVAD and are more likely to be carried out because of unrelated conditions such as cholecystitis or diverticulitis.7,12 Abdominal exploration was the most frequent procedure seen in our study and clinicians must pay close attention to change in abdominal exam and the development of acute pain or nausea and vomiting. There must be an especially high clinical index of suspicion when there is hemodynamic compromise and an elevated serum lactate level in addition to worsening abdominal exam. Driveline infection is
also a known problem in patients with implantable LVAD.14,15 In this study, there were two driveline infections requiring surgical intervention; however, these infections did not result in intraabdominal procedures. However, this study represents the first to examine outcomes with noncardiac surgical procedures in patients supported with temporary ECMO. Although results for noncardiac surgical procedures in patients with implantable LVAD appear to be good, outcomes for patients on ECMO are poor. In the current study, almost half of patients in the ECMO cohort requiring noncardiac surgical procedures died within 30 days. This is not unexpected, because patients requiring temporary ECMO are known to have high mortality risk.16,17 The most common surgical procedure in the ECMO group was placement of a tracheostomy (n = 5), extremity/vascular procedures (n = 4), and abdominal exploration/bowel resection (n = 4). Vascular complications are a known risk of arterial cannulation in ECMO and can often require surgical intervention.17 In the current study, all patients requiring vascular/extremity procedures received peripheral femoral artery cannulation. As experience with ECMO increases and outcomes improve,19 more and more patients on ECMO will require noncardiac surgical procedures. Further studies are needed to improve outcomes in this very high-risk group. Because of the risk of thromboembolic complications,20 patients with LVAD and ECMO are usually anticoagulated and on aspirin.10,21 This presents a conundrum to surgeons as they are faced with high risk of bleeding versus inducing a thromboembolic complication. Current guidelines for patients with LVAD recommend maintaining an INR between 1.5 and 2.5 to minimize pump thrombosis and thromboembolic complications.10,21 Because of the need for anticoagulation in this patient population, perioperative bleeding is a known complication during noncardiac surgical procedures11,23 and was found to be problematic in the current study. In the current study, 25.5%
Table 4. Postoperative Outcomes
Mean total length of stay (d) Mean intensive care unit stay (hr) Requirement of postoperative mechanical ventilation, n (%) Duration mechanical ventilation (hr) Requirement of postoperative vasopressor support, n (%) Time requiring vasopressor support (hr) Increase in creatinine (mg/dl) Requirement of blood transfusion within 24 hours of surgery, n (%) Average packed red blood cells transfused within 24 hours Perioperative mortality, n (%)
Left Ventricular Assist Device (N = 47)
Extracorporeal Membrane Oxygenation (N = 15)
p Value
37.0 ± 33.6 28.7 ± 33.2 32 (68.1) 216.7 ± 277.0 19 (36.2) 228.6 ± 653.1 0.07 ± 0.57 12 (25.5) 0.8 ± 1.7 3 (6.4)
30.1 ± 42.2 24.9 ± 38.8 15 (100.0) 347.1 ± 447.8 9 (19.1) 48.7 ± 78.8 0.44 ± 0.67 11 (73.3) 2.8 ± 3.2 7 (46.7)
0.52 0.71 0.01 0.19 0.24 0.29 0.047 0.002 0.003 0.001
673
PROCEDURES IN ECMO AND LVAD PATIENTS Table 5. Univariate Regression Analysis for Survival
Age Male White Black Hispanic Extracorporeal membrane oxygenation History of biventricular assist device Ischemic cardiomyopathy Preoperative coumadin Preoperative INR Preoperative aspirin History of hyperlipidemia History of diabetes History of hyperlipidemia History of chronic lung disease History of chronic renal insufficiency Postoperative blood transfusion Preoperative creatinine Postoperative creatinine
Hazard Ratio
95% Confidence Interval
p Value
0.99 0.74 0.80 0.81 1.07 2.90 2.02 0.75 0.59 0.85 0.48 0.65 0.97 0.65 1.41 0.85 3.14 1.00 0.95
0.97–1.01 0.41–1.35 0.42–1.54 0.44–1.49 0.33–3.51 1.46–5.78 1.01–4.04 0.40–1.41 0.21–1.66 0.53–1.38 0.26–0.89 0.35–1.20 0.53–1.76 0.35–1.20 0.70–2.87 0.42–1.69 1.66–5.93 0.76–1.34 0.75–1.22
0.74 0.33 0.51 0.50 0.91 0.002 0.049 0.78 0.32 0.52 0.02 0.17 0.91 0.17 0.34 0.63
