Abstract
Background:
Vascular complications remain a prevalent and devastating complication of extracorporeal membrane oxygenation (ECMO). Risks for vascular complications have risen as the volume and medical complexity of ECMO cases has rapidly increased in recent years1–3. This study assesses occurrence and clinical impact of vascular complications across all ECMO modes and cannulation methods in a high-volume academic center.
Methods:
A retrospective single-institution review was performed of all ECMO cannulations for central or peripheral veno-venous (VV) or veno-arterial (VA) ECMO in 2019–2020. Patients that expired during cannulation were excluded. Primary outcomes of vascular consultation rates at or following ECMO placement, limb loss, and mortality were assessed during index hospitalization.
Results:
229 patients were identified during the two-year study period. VA ECMO was used in the majority of patients (n=137, 60%), with 83% (N=114) undergoing peripheral cannulation. Vascular surgery was consulted in 54 (23.6%) patients. Complication rates ranged from 33.9% in peripheral VA cannulation to 7% in jugular VV cannulation. Overall, 65% of vascular consults required operative intervention; interventions were highest in peripheral VA ECMO (n=30/114, 26.3%). Across all ECMO types, acute limb ischemia (ALI) was the most common complication (n=38, 16.5%) with rates ranging from 26.1% in Central VA ECMO to 4.8% in jugular VV ECMO. Distal antegrade perfusion catheters (DPC) were employed in n=68/114 (59.6%) of all peripheral VA ECMO cases. Prophylactic DPC were found to be incorrectly placed in 10.2% (n=7/68) of cases, which obscured ability to full evaluate effect of prophylactic DPC on reducing rates of ALI. Major amputation (BKA/AKA) occurred in 6 peripheral VA patients (5.3%), 2 central VA patients (8.65), and 2 femoral-femoral VV patients (4%). Patients with ALI experienced significantly lower rates of in-hospital survival in Kaplan-Meier analysis (32.5% vs. 54%, Log Rank=0.023).
Conclusion:
This study highlights the prevalence of vascular complications, and their associated mortality impact, across all modes of ECMO and additionally identifies several areas for institutional performance improvement. ALI was the most common complication across all cohorts and was associated with decreased in-hospital survival. Impact of DPC on ALI was difficult to elucidate due to non-standardized placement patterns and selective use. In the care of these medically complex patients on multidisciplinary teams, review of outcomes and group discussions targeting areas for improvement are critical to success; in this study, findings resulted in development of a lower extremity perfusion management protocol.
Keywords: ECMO, Acute Limb Ischemia, Perfusion Catheter
Table of Contents:
Vascular complications are prevalent across all modes of ECMO and appear to influence in-hospital survival. Selective use of distal perfusion catheters was not associated with significantly reduced rates of ALI, but non-selective prophylactic usage is increasing and the impact on ALI is yet to be determined. Multidisciplinary team management remains key to success.
Introduction:
Vascular complications remain a prevalent and devastating complication of extracorporeal membrane oxygenation (ECMO). Risks for vascular complications have risen as the volume and medical complexity of ECMO cases has rapidly increased in recent years1–3. Vascular complications increase patient morbidity and healthcare utilization; as such, understanding and preventing these complications remains paramount4,5. The majority of research on vascular complications in ECMO focuses on peripheral VA ECMO6–8, however, clinically vascular surgeons are increasingly being called upon to participate in the care of all ECMO patients regardless of mode. The range of vascular complications, and their clinical impact, has yet to be well-defined for all modes of ECMO.
Within peripheral VA ECMO, the range of vascular complications appears to vary widely based on institutional volume, cannulation techniques, and patient population. Potential risk factors for vascular complications in peripheral ECMO include cannula size9,10 usage of distal perfusion catheters11, and percutaneous cannulation12,13; however, these results are not consistently reproduced across all institutions7,14. The impact of vascular complications on overall morbidity and mortality in peripheral VA ECMO has been persistent; with hemorrhage and acute limb ischemia (ALI) being reliably identified as risk factors for poor in-hospital survival15,16. Research on predictors of and consequences of vascular complications has yet to be well explored across all modes of ECMO within the vascular literature.
Use of protocols and standardized practices for the prevention and management of complications have been shown to improve clinical outcomes17–19. The Extracoporeal Life Support Organization (ELSO) publishes guidelines pertaining to ECMO indications and management; however, they do not yet have mandatory protocols for prevention or management of vascular complications20. Development and implementation of protocols therefore falls to the responsibility of individual institutions.
This study aims to evaluate occurrence and outcomes of vascular complications and their surgical management across all ECMO modes and cannulation configurations in a high-volume academic center. Results of this study were used to inform an institutional protocol for the prevention and management of lower extremity vascular complications following ECMO cannulation.
Methods:
Cohort Selection
This is a retrospective single center review of all inpatients maintained on ECMO between January 1st 2019-December 31th 2020. Cohort includes patients transferred to our institution for ECMO, including patients cannulated at an outside institution prior to transfer. Cohort included Veno-Arterial (VA) and Veno-Venous (VV) modes and all cannulation configurations including central, lower extremity peripheral, and upper extremity peripheral. Patient’s cannulated as part of extracorporeal cardio-pulmonary resuscitation (E-CPR) were included. Cannulations and ECMO management were performed by either Cardiac Surgery or Interventional Cardiology, with variation in techniques and cannulation locations across teams and providers. Patients that expired during initial cannulation procedure were excluded. Study protocol was approved by the Institutional Review Board at the University of Alabama at Birmingham.
Endpoints and Definitions
The primary endpoint was rate of ECMO-associated vascular complications (intra-operative or post-operative) requiring consultation to Vascular Surgery during index hospitalization. This endpoint did not capture vascular injuries and complications managed by the primary ECMO team, excessively futile injuries that did not warrant a surgical consultation, or complications not directly to ECMO (example: referral for prophylactic IVC filter placement). Focus was placed exclusively on vascular complications requiring consultation to Vascular Surgery for two reasons: in order to adequately reflect the degree of Vascular Surgery involvement in management of ECMO patients, and to avoid falsely elevating the complication rates with intra-operative vascular issues that were often managed successfully by the primary team with little clinical consequence. Complications were not mutually exclusive, and patients may experience more than one complication in more than one limb; however the overall rate of vascular complications is reported per person rather than per limb. Complications were classified as either directly related to cannulation procedure or those that were indirectly related to cannulation. Secondary outcomes included rates of prophylactic distal perfusion catheter usage (pDPC), vascular surgery intervention rates and techniques, and in-hospital mortality. pDPC catheters were defined as those placed during initial cannulation procedure. Amputations were categorized as major (below knee, above knee, and upper extremity transhumeral) or minor (digital, foot, and hand). Interventions were broadly classified and were not mutually exclusive.
Statistical Analysis:
Collected data was analyzed in SPSS statistical software, version 23.0 (SPSS Inc, Chicago, IL). Continuous variables were presented as mean ± SD and were compared using an independent Student t test. Categorical variables were presented as number and percentage, and were compared using a Chi-square test. Kaplan-Meier and Log-Rank analysis were used to plot outcomes related to survival. A p value <0.05 defined statistical significance.
Results:
Demographics and Procedural Details
A total of 229 patients were successfully cannulated for ECMO between January 1st 2019 and December 31th 2020. The majority of patients were cannulated for VA ECMO (n=137, 60%) compared to VV ECMO (n=92, 40%). Exact distribution of modes and cannulation configurations are shown in Figure 1. The median age was 53.2 years (IQR 27.8 years) and the majority of patients were Caucasian (48.9%) males (63.3%). The most prevalent comorbidities were hypertension (71.5%) and hyperlipidemia (53.7%). Table 1 highlights key demographics and comorbidities for the whole cohort.
Figure 1:
Distribution of patients across ECMO modes and cannulation configurations
Table 1:
Baseline Patient Demographics and Comorbidities, along with ECMO Procedural Details
| All (n=229) | VA (n=137, 60%) | VV ECMO (n=92, 40%) | p | |||
|---|---|---|---|---|---|---|
| Peripheral VA (n=114, 49.8%) | Central VA (n=23, 10%) | Femoral VV (n=50, 21.8%) | IJ Dual Lumen VV (n=42, 18.3%) | |||
|
| ||||||
| Age (years) (median, IQR) | 53.2 (27.8) | 55 (29.2) | 58.1 (19.7) | 46.9 (26.6) | 50.2 (30.4) | 0.125 |
| Male Gender | 145 (63%) | 67 (58.8%) | 12 (52.2%) | 35 (70%) | 31 (73.8%) | 0.158 |
| Caucasian Race | 112 (49%) | 53 (46.5%) | 12 (52.2%) | 25 (50%) | 22 (52.4%) | 0.293 |
| BMI (median, IQR) | 30 (11) | 29.4 (10.6) | 29.7 (13.3) | 31.4 (9.3) | 31.2 (9.9) | 0.783 |
| Comorbidities: | ||||||
| Hypertension | 164 (71.5%) | 79 (69.3%) | 22 (95.7%) | 34 (68%) | 29 (69%) | 0.063 |
| Hyperlipidemia | 123 (53.7%) | 65 (57%) | 18 (78.3%) | 24 (48%) | 16 (38.1%) | 0.013 |
| Diabetes Mellitus | 60 (26.2%) | 37 (32.5%) | 7 (30.4%) | 8 (16%) | 8 (19%) | 0.097 |
| COPD | 33 (14.5% | 20 (17.5%) | 3 (13%) | 6 (12.2%) | 4 (9.5%) | 0.571 |
| Coronary Artery Disease | 101 (44.1%) | 61 (53.5%) | 18 (78.3%) | 8 (16%) | 14 (33.3%) | <0.001 |
| Congestive Heart Failure | 74 (32.3%) | 52 (45.6%) | 8 (34.8%) | 5 (10%) | 9 (21.4%) | <0.001 |
| Chronic Kidney Disease | 34 (14.8%) | 16 (14%) | 4 (17.4%) | 4 (8%) | 10 (23.8%) | 0.195 |
| ESRD | 11 (4.8%) | 5 (4.4%) | 2 (8.7%) | 2 (4%) | 2 (4.8%) | 0.834 |
| Current Smoker | 39 (17.3%) | 16 (14.4%) | 10 (43.5%) | 5 (10.2%) | 8 (19%) | 0.075 |
| Transferred for ECMO | 104 (45.4%) | 47 (41.2%) | 13 (56.5%) | 25 (50%) | 19 (45.2%) | 0.498 |
| Cannulated at OSH | 14 (6.1%) | 7 (6.1%) | 5 (21.7%) | 2 (4%) | 0 | 0.005 |
| E-CPR Cannulation | 57 (24.9%) | 45 (39.5%) | 5 (21.7%) | 5 (10%) | 2 (4.8%) | <0.001 |
| Procedure Location | ||||||
| CV Operating Room | 116 (50.7%) | 50 (43.9%) | 17 (73.9%) | 29 (58%) | 20 (47.6%) | |
| Cardiac ICU Bedside | 40 (17.5%) | 27 (23.7%) | 1 (4.3%) | 4 (8%) | 8 (19%) | |
| Interventional Suite | 28 (12.2%) | 21 (18.5%) | 0 | 1 (2%) | 6 (14.3%) | |
| Medical ICU Bedside | 17 (7.4%) | 3 (2.6%) | 0 | 9 (18%) | 5 (11.9%) | |
| Outside Hospital | 14 (6.1%) | 7 (6.1%) | 5 (21.7%) | 2 (4%) | 0 | |
| Other | 14 (6.1%) | 6 (5.2%) | 0 | 5 (10%) | 3 (7.1%) | |
| Cannula Size (Fr) (median, IQR) | 21 (4) | 17 (2) | 21 (3.5) | 25 (6) | 32 (4) | 0.001 |
| Run Time (days) (median, IQR) | 9 (18.5) | 6 (13) | 6 (11) | 17.5 (31) | 13 9 (26.5) | 0.799 |
| Multiple ECMO Runs | 25 (10.9%) | 14 (12.2%) | 5 (21.7%) | 2 (4%) | 4 (11%) | 0.24 |
COPD, Chronic Obstructive Pulmonary Disease; ESRD, End Stage Renal Disease; OSH, Outside Hospital; E-CPR, extracorporeal cardiopulmonary resuscitation; IQR, interquartile range
Table 1 summarizes details of ECMO cannulation procedures. A total of 104 patients (45.4%) were transferred to our institution for ECMO management, of which, 14 (6.1%) were cannulated prior to transfer. Central VA ECMO had the highest prevalence of transfers (n=13, 56.5%) as well as the highest prevalence of cannulation at outside hospitals (n=5, 21.7%). A quarter of the whole cohort was cannulated as part of E-CPR (n=57, 24.9%), with VA ECMO patients having a significantly higher prevalence of E-CPR cannulation compared to VV ECMO. Cannulation within the CV operating room was the most prevalent (50.7%), followed by bedside in the cardiac ICU (17.5%), and in the interventional suite (12.2%). Open cannulation under direct visualization was performed in 20.5% (n=47) of the cohort, with rates ranging from 100% in Central VA ECMO to 6% (n=3) in Femoral VV ECMO. Cannulation method is unknown in the 14 cases (6.1%) where cannulation occurred at an outside hospital. Percutaneous cannulation was utilized in 73.3% of the cohort (n=168), and was most prevalent within IJ Dual Lumen VV ECMO. Ultrasound guidance was documented in the procedural note for 90% of percutaneous cannulation cases; failure to document ultrasound utilization or failure to use ultrasound was most commonly seen in procedures done as part of E-CPR. Only 34% of the cohort had any type of non-invasive diagnostic vasculature evaluation prior to cannulation for ECMO; none of the documented diagnostic of imaging evaluation were obtained as a direct relation to ECMO. CT scans were performed pre-cannulation in 18.7% of the cohort, and were most commonly obtained for evaluation of an ongoing medical condition (n=25), part of routine pre-operative evaluation (n=6, 2.6%) (example: planning for routine Transfemoral Aortic Valve Replacement), or as part of trauma evaluation (n=12, 5.2%). Ankle-Brachial Indexes were collected on 17 patients (7.4%) as part of routine pre-operative planning in cardiac and lung transplant patients. The median cannula size varied significantly across cohorts; within the peripheral VA cohort, the median arterial cannula was 17 Fr (range 10Fr-25Fr). Median run time was 9 days (IQR 4–22.5), with no significant variation across the cohorts. Multiple ECMO runs were seen in 10.9% of the whole cohort.
Vascular Complication and Intervention Details
54 patients (23.6%) required consultation to vascular surgery for iatrogenic vascular injuries and complications (Table 2): Within these 54 patients, 41 patients (76%) had complications directly related to cannulation site, while 13 patients (25%) had general vascular complications. Across the modes of ECMO consults to vascular surgery were seen in 33.3% of patients cannulated on peripheral VA ECMO (n=38/114), 26.1% of central VA patients (n=6/23), 14% of femoral-femoral VV (n=7/50), and 7.1% of internal jugular VV patients (n=¾2). Average time from cannulation to consultation was 10.5±14.5 days, with 14 patients having complications identified within 24 hours of original cannulation procedure. The peripheral VA cohort was the only group to experience multiple separate complications. A single patient experienced three separate complications; ALI followed by hemorrhage and finally a pseudoaneurysm. Across all cohorts, ALI was the most common complication (16.5%, n=38), ranging from 26.1% (n=6/23) in Central VA patients to 4.8% (n=2/42) in internal jugular VV patients. In all central VA and VV ECMO patients, ALI was a general complication and not directly related to cannula site injury, and affected arteries had concurrent or recent direct instrumentation with either a support device (Ex: intra-aortic balloon pump) or as a part of surgical intervention (ex: transfemoral aortic valve replacement). Additionally, one patient on IJ VV ECMO experienced spontaneous lower extremity ALI as a consequence of heparin-induced thrombocytopenia (HIT), which was considered a complication not directly related to the cannulation procedure. Bleeding complications were seen in 9.6% (n=22/229) of all patients, and were isolated to peripheral VA and Femoral-Femoral VV ECMO cohorts. Frank active hemorrhage was only seen within the peripheral VA cohort (n=7/114, 6.1%). Vascular was consulted on three patients (n=3/229, 1.3%) that developed significant Inferior Vena Cava thrombus leading to symptomatic pulmonary embolism.
Table 2:
Vascular Consultations and Complications (2019–2020)
| All (n=229) | VA (n=137, 60%) | VV ECMO (n=92, 40%) | |||
|---|---|---|---|---|---|
| Peripheral VA (n=114, 49.8%) | Central VA (n=23, 10%) | Femoral VV (n=50, 21.8%) | IJ Dual Lumen VV (n=42, 18.3%) | ||
|
| |||||
| Any Consult to Vascular Surgery | 54 (23.6%) | 38 (33.3%) | 6 (26.1%) | 7 (14%) | 3 (7.1%) |
| Direct Cannula Complications | 41 (17.9%) | 38 (33.3%) | 0 | 3 (6%) | 0 |
| General Complications | 13 (5.6%) | 0 | 6 (26.1%) | 4 (8%) | 3 (7.1%) |
| Total Vascular Complications | 63 | 47 | 6 | 7 | 3 |
| Number of Complications | |||||
| 1 | 46 (20.1%) | 30 (26.3%) | 6 (26.1%) | 7 (14%) | 3 (7.1%) |
| 2 | 7 (3%) | 7 (6.1%) | 0 | 0 | 0 |
| 3 | 1 (0.4%) | 1 (0.8%) | 0 | 0 | 0 |
| Complication Type* (# patients) | |||||
| Acute Limb Ischemia | 38 (16.5%) | 26 (22.8%) | 6 (26.1%) | 4 (8%) | 2 (4.8%) |
| Bleeding | 22 (9.6%) | 19 (16.6%) | 0 | 3 (6%) | 0 |
| Hemorrhage | 7 (3.1%) | 7 (6.1%) | 0 | 0 | 0 |
| Hematoma | 8 (3.5)% | 7 (6.1%) | 0 | 1 (2%) | 0 |
| Pseudoaneurysm | 7 (3.1%) | 5 (4.4%) | 0 | 2 (4%) | 0 |
| IVC Thrombus with VTE | 3 (1.3%) | 2 (1.8%) | 0 | 0 | 1 (2.3%) |
Not mutually exclusive
Complications resulted in operative intervention in 55% of cases (n=35/63) (Table 3); ranging from 63.8% in the peripheral VA cohort to 0% in the internal jugular VV cohort. Decision to not pursue operative intervention was most frequently based on futility of effort (e.g. brain death) or presence of significant physiologic instability. Median time until first reintervention was 4 days (IQR 1–19.5 days). Trips to the operating room ranged from 1 to 4, with more than 2 trips only seen in the peripheral VA cohort. Reported repair techniques shown in Table 3 are not mutually exclusive. The most common interventions were thromboembolectomy (n=25, 46.2%) and patch angioplasty arterial repair (n=24, 44.4%). Therapeutic distal perfusion catheters were placed in 14 patients (25.9%) on peripheral VA ECMO. Within patients receiving therapeutic DPC, two patients required subsequent major amputations (n=2/14, 14.2%). Fasciotomies were performed in 11 patients (20.3%), and were only seen in VA ECMO patients. Major amputation rate was 18.5% of patients with complications, corresponding to 4.3% of the entire cohort. One patient had bilateral below-knee amputations and one patient had an upper extremity transhumeral amputation. Diagnostic Angiograms were utilized in 5 patients, but no therapeutic endovascular techniques were employed in the treatment of ECMO patients.
Table 3:
Categorization and Distribution of Vascular Interventions within patients experiencing vascular complications (2019–2020)*
| All (n=54) | VA (n=44, 81.5%) | VV ECMO (n=10, 18.5%) | |||
|---|---|---|---|---|---|
| Peripheral VA (n=38, 70.3%) | Central VA (n=6, 11.1%) | Femoral VV (n=7, 12.9%) | IJ Dual Lumen VV (n=3, 5.5%) | ||
|
| |||||
| Patients Receiving Operations | 35 (15.3%) | 30 (26.3%) | 2 (8.7%) | 3 (6%) | 0 |
| Number of OR Trips | |||||
| 1 | 23 (10%) | 19 (16.7%) | 1 (4.3%) | 3 (6%) | 0 |
| 2 | 5 (9.2%) | 4 (3.5%) | 1 (4.3%) | 0 | 0 |
| 3 | 4 (1.7%) | 4 (3.5%) | 0 | 0 | 0 |
| 4 | 3 (1.3%) | 3 (2.6%) | 0 | 0 | 0 |
| Types of Intervention | |||||
| Thromboembolectomy | 25 (46.2%) | 24 (55.2%) | 1 (16.6%) | 0 | 0 |
| Patch Angioplasty Repair | 24 (44.4%) | 22 (57.8%) | 1 (16.6%) | 1 (14.2%) | 0 |
| DPC Placement | 14 (25.9%) | 14 (36.8%) | 0 | 0 | 0 |
| Fasciotomy | 11 (20.3%) | 10 (26.3%) | 1 (16.6%) | 0 | 0 |
| Amputation | 11 (20.3%) | 7 (18.4%) | 2 (33.3%) | 2 (28.5%) | 0 |
| Major Amputation | 10 (18.5%) | 6 (15.7%) | 2 (33.3%) | 2 (28.5%) | 0 |
| Minor Amputation | 1 (1.8%) | 1 (2.6%) | 0 | 0 | 0 |
| Hematoma Evacuation | 6 (11.1%) | 5 (13.1%) | 0 | 1 (14.2%) | 0 |
| Bypass | 5 (9.2%) | 4 (3.5%) | 0 | 1 (14.2%) | 0 |
| Angiogram | 3 (5.5%) | 3 (7.8%) | 0 | 0 | 0 |
| Muscle Flap Coverage | 4 (7.4%) | 4 (10.5%) | 0 | 0 | 0 |
| IVC Thrombectomy & Filter | 1 (1.8%) | 1 (2.6%) | 0 | 0 | 0 |
Not Mutually Exclusive.
OR, Operating Room; DPC, Distal Perfusion Catheter; IVC, Inferior Vena Cava
Factors Associated with Acute Limb Ischemia in VA ECMO
Across the whole cohort, univariate and Linear Cox Regression Analysis did not identify any pre-operative patient or procedural factors that appear to influence rates of ALI (Supplement Table 1). Patients with and without ALI were statistically equal in age and smoking status compared to patients without ALI. Similarly, peripheral VA patients did not differ in rates of transfers (36% vs 42.7%, 0.548) or rates of cannulation at OSH (4% vs 6.7%, p=0.614) compared to patients without ALI. Cannulation as part of E-CPR did not appear to have a relation to ALI rates. Arterial size was slightly larger in peripheral VA patients with ALI (18.1±2.5 Fr vs. 17.3±2.6 Fr, p=0.194) compared to those without ALI however this difference did not reach statistical significance. pDPC were used at a lower rate in patients with ALI (54%, n=14/26) compared to those without ALI (61.3%, n=54/88) however this difference was not statistically significant (p=0.674, Supplement Table 1). Peripheral VA patients with ALI did not experience more ECMO runs or have longer duration of cannulation time compared to patients without ALI.
Impact of Prophylactic Distal Perfusion Catheters
A total of 68 peripheral VA ECMO patients had DPC placed during their initial ECMO cannulation (n=68/114, 59.6%). Utilization increased from 56.6% in 2019 to 62.9% in 2020. All of the pDPC were antegrade and placed in the groin. The majority of pDPC were placed in the cardiac operating room (n=28/68, 41.2%), followed by bedside in the cardiac ICU (n=17/68, 25%) and the interventional suite (n=16/68, 23.5%). Despite having the highest gross number of pDPC placed in the cardiac OR, the utilization rate of pDPC compared to all cases performed in the cardiac operating room was 56%, compared to 77% in the interventional suite, 60% bedside in the cardiac OR. Of the 68 pDPC, 7 (10.2%) cases were found at the time of vascular consult to be incorrectly placed in either a vein (n=5) or subcutaneous tissue (n=2). It is unknown if any were incorrectly placed in patients that did not receive vascular surgery evaluation. Univariate statistical analysis did not identify any patient or procedural factors that appear to positively influence usage rates of pDPC (Supplement Table 2). Arterial cannula size did not significantly differ between peripheral VA ECMO patients with and without pDPC (17.6±1.8 Fr vs. 17.5±2.9 Fr, p=0.842). BMI was non-significantly lower in patients that did receive pDPC (29.5±7.6 vs. 31.5±7.5, p=0.173).
Impact of pDPC on ALI was difficult to elucidate due to non-standardized placement patterns and high prevalence of initial misplacement of the pDPC. Presence of pDPC did not appear to be significantly statistically associated with reduced rates of ALI in Peripheral VA ECMO patients (Table 4). ALI occurred at a rate of 20.6% in patients with pDPC compared to a rate of 26% in patients without pDPC; however, this reduction did not reach statistical significance. In Binary Logistic Regression, pDPC did demonstrate decreased odds for ALI in peripheral VA ECMO patients, however this did not reach statistical significance (OR 0.825 (0.336–2.024), p=0.674, Table 4).
Table 4:
Presence of pDPC does not appear to be significantly related to reduced rates of ALI in Peripheral VA
| All (n=114) | pDPC Present (n=68, 59.6%) | pDPC Absent (n=46, 40.4%) | Binary Logistic Regression | |
|---|---|---|---|---|
|
| ||||
| Acute Limb Ischemia | 26 (22.8%) | 14 (20.6%) | 12 (26.0%) | OR 0.825 (0.336–2.023), p=0.674 |
pDPC, prophylactic Distal Perfusion Catheter; ALI, Acute Limb Ischemia
ALI did have a significant relationship with initially incorrectly placed pDPC. Univariate statistical analysis did not identify any patient anatomic or procedural factors statistically associated with increased rates of incorrect pDPC placement; however, the majority of the incorrectly placed pDPC were placed bedside in the cardiac ICU (n=4, 57.1%) and as a part of E-CPR (n=4, 57.1%). BMI did not differ significantly between those with correct and incorrect pDPC (30.6±7.6 vs 25.2±6.0, p=0.070). Within these incorrectly placed pDPC, there was a 100% rate of ipsilateral limb ALI and a 100% rate of surgical intervention. Time between incorrect pDPC placement and recognition of ALI along with surgical intervention was <24 hours in all cases. In all cases, patients underwent open exploration with thromboembolectomy and patch angioplasty. No patients required major amputations during index hospitalization following incorrect pDPC placement; however 2 of these 7 patients had ischemia that would have required major amputation had they not expired prior to intervention, and 1 of these patients returned several months after index hospitalization for an amputation. In-hospital mortality rate was 57.1% in patients with incorrectly placed pDPC.
Impact of Vascular Complications on In-Hospital Mortality
Within the whole cohort, the median length of stay was 42 days (IQR 14–49.5 days) and the in-hospital mortality rate was 49.3%. Within the whole cohort, including all modes and cannulation configurations, patients with ALI experienced significantly lower rates of in-hospital survival (32.5% vs. 54%, Log Rank=0.023, Figure 2). In survival Cox Regression, presence of ALI had a Hazard Ratio of 1.6 (95% CI 1.1–2.6, p=0.026) for in-hospital mortality. Lack of operative intervention was associated with significantly decreased in-hospital survival (11% vs 57%, HR 3.9 (95% CI 1.8–7.5), p<0.001) compared to patients who underwent operative intervention for their injuries.
Figure2:
Impact of Acute Limb Ischemia on In-Hospital Survival Across all Modes of ECMO
Discussion:
This study highlights the prevalence of vascular complications, and their associated mortality impact, across all modes of ECMO and additionally identifies several areas for institutional performance improvement. In the care of these medically complex patients on multidisciplinary teams, review of outcomes and group discussions targeting areas for improvement are critical to success; in this study, findings resulted in development of a lower extremity perfusion management protocol.
This study is the one of the largest single institution studies dedicated to ECMO vascular complications, and one of very few studies to evaluate complications across all ECMO modes. Vascular complications were seen in 23.6% of the entire cohort; ranging from 7.1% in internal jugular dual lumen VV to 33.3% in peripheral VA. In the literature, rates of peripheral VA complications range from 10%8 in routine cannulation to 67%21 in those cannulated as part of E-CPR7. Historically, hemorrhage was the most common vascular complication following ECMO cannulation22, however in this study, ALI was the most common across all modes, reflecting findings of several previous modern studies14,21. Central VA had the highest ischemia rate (26.1%), but peripheral VA had the highest raw number of patients with ALI (n=26). Interestingly, frank infection rates were much lower than expected; the majority of patients with groin infections (most commonly presenting as PSA) were treated several months after index hospitalization and were therefore not included in this analysis. Prevalence of vascular complications across all ECMO modes highlights the fact that these complications are frequently multifactorial and related to injury, physiologic compromise, and underlying vascular disease.
Operative intervention was seen in 65% of patients with vascular complications. Multiple trips to the operating room were only seen in VA ECMO cohorts. Endovascular techniques were utilized as part of a diagnostic approach, but no therapeutic endovascular interventions were used in any patient. Major amputation rate of 4.3% is in line with national reports23. This study highlights that major amputations are not isolated to peripheral VA ECMO. Minor amputations at index hospitalization were significantly less common; minor amputations were intentionally delayed and planned for after index hospitalization when patient was no longer dealing with the immediate physiologic effects of ECMO. Muscle flap coverage was lower than expected; which we hypothesize is due to two factors. First, increasing recruitment of plastic surgery to perform complex coverage and second because the majority of groin infections presented after index hospitalization and therefore these repairs were not captured in this study timeframe. Future studies may want to investigate long-term complications and operative repairs performed outside of the index hospitalization, as we suspect variations in types of complications and modes of treatment would be present outside of the acute treatment window.
This study did not identify any anatomic or procedural factors associated with increased rates of ALI (Supplement Table 1). Arterial cannula size has been identified as a possible predictor for ALI9,10. However not all studies14, including this current study, are able to reproduce the relationship between arterial cannula size and ALI. Similarly, previous findings with lower rates of ALI in percutaneous cannulation13,24 were not replicated in this study. Other factors examined but that were not found to impact rates of ALI were cannulation as part of E-CPR, multiple ECMO runs, and duration of ECMO run time (Supplement Table 1).
Utilization of pDPC is increasing; however, prevalence of incorrectly placed pDPC obscured any obvious (and expected) benefits for the prevention of ALI. Studies have demonstrated reproducible trends of increasing pDPC utilization in peripheral VA ECMO11,23, and subsequently notable reductions in ALI rates25–27. In this study, utilization of pDPC increased between 2019 and 2020; however, the average utilization was only 59.6% within peripheral VA ECMO patients. The majority of pDPCs were placed in the operating room; however, relative to total cases performed, the operating room had a lower utilization rate compared to the interventional suite, highlighting differences in individual provider practice patterns. When non-standardized, the choice to place a pDPC can be impacted by a variety of factors that are difficult to accurately capture in a retrospective chart review. Although we were unable to demonstrate a benefit of DPC in the prevention of ALI, this variability in practice patterns certainly supports implementation of institutional guidelines. The lack of apparent statistical relationship between pDPC utilization and rates of ALI was perhaps due to selective utilization and non-standardized practice patterns, introducing bias into this retrospective analysis and likely represents a Type 2 statistical error. Despite the results of our study, our clinical experience along with results from other retrospective reviews lead us to believe that placement of pDPC in peripheral VA ECMO cases is a crucial element of proactive efforts to prevent ALI. The true impact of pDPC on ALI rates will need to be re-evaluated at our institution following implementation of mandatory pDPC protocol.
ALI occurred in 20.6% of patients with pDPC, compared to 26.0% of patients without pDPC, however this reduction was not statistically significant and likely influence by prevalence of incorrectly placed pDPC. ALI occurred in 100% of patients with incorrectly placed pDPC, and resulted in a 100% rate of operative intervention, likely skewing logistic regression results. We suspect that the rate of incorrectly placed pDPC is actually higher, as this study may not have well capture pDPC placed into the profunda rather than common femoral artery and may not have captured placement issues in patients that did not receive consults to vascular surgery. Unfortunately, this study did not identify any statistically significant patterns within incorrectly placed pDPC, although we suspect this is due to the small sample size (n=7). Some trends identified within incorrectly placed pDPC include placement as part of E-CPR (n=4, 57.1%) and placement at bedside in cardiac ICU (n=4, 57.1%). We hypothesized that patients with higher BMI may be at increased risk for incorrectly placed pDPC, however this was not found to be the case in our data. We believe that larger studies would be able to identify anatomic or procedural factors associated with incorrect pDPC placement, and would be able to help guide the need for open vs. percutaneous DPC placement.
ALI across all modes of ECMO was associated with significantly reduced in-hospital survival (Figure 2). In previous work, both hemorrhage15 and ALI4,16 have demonstrated a relationship with decreased survival, although this is not a finding universally found in all studies27. Patients with vascular complications that did not undergo operative intervention, either due to inappropriateness of consult, physiologic compromise, or futility of efforts, experienced a 11% rate of in-hospital survival. This markedly poor survival in patients not selected for intervention highlights appropriate patient selection and insight into appropriateness of surgical management verse comfort care. Relationship between ALI and mortality reinforces importance of complication prevention rather than reactive intervention.
The results of this study were used to inform discussions with the UAB Mechanical Circulatory Support Device Quality Improvement Multidisciplinary Team, resulting in the development of a peripheral VA ECMO perfusion management protocol. Deficiencies and areas for improvement were identified at every level of care; including operative technique, vendor availability, physician communication, bedside monitoring, and inadequate resource provisions from the hospital. The largest barriers to optimizing care were time (examples: retraining nursing staff, increased requirements for EHR documentation) and resources (limited number of ultrasound technicians, difficult to modify pre-existing vendor supply chains). Protocol now involves mandatory placement of pDPC in all peripheral VA patients, placement confirmation with duplex ultrasound, monitoring perfusion with near-infrared reflectance spectroscopy (NIRS)28,29, and bedside nursing adherence to hourly monitoring check-list (Supplement Figure 1). Lower extremity bedside monitoring protocol now requires hourly physical and pulse exam with Doppler ultrasound and subsequent documentation in the chart, hourly evaluation of tubing patency with documentation in the chart, a bedside sign with documentation of available access and cannulation locations, and bedside placement of a call tree for concerns related to peripheral ischemia. All patients on ECMO will now have comprehensive evaluation by a member of the ECMO Management Team every 4 hours in addition to hourly evaluation by the Intensive Care Unit team member. Certain protocol details such as mandatory ultrasound utilization during percutaneous cannulation, preferences for smaller arterial cannulas, and inclusion of bedside flow monitoring methods, are still being evaluated and finalized. Following a year of implementation, we plan to report our institutional rates of vascular complications, acute limb ischemia, and amputations along with an extensive report of protocol detail.
Limitations:
The retrospective nature of this study introduces inherent bias and limitations. This review captures only vascular complications for which vascular surgery was consulted and does not include injuries treated directly by cardiac surgery; leaving the possibility that rates are underestimated. Non-standardized practice patterns with regard to DPC utilization reduced our ability to draw well-supported conclusions regarding impact of DPC on rates of ALI. We suspect that the relationship between ALI and pDPC hinges on issues of cannula size relative to native vessel and extent of collateral perfusion pathways in a manner that our study is not primarily designed or powered for. Despite these limitations, we feel that the information captured within this study is relevant to practicing vascular surgeons and reflects true clinical practice patterns.
Conclusion:
This study highlights the prevalence of vascular complications, and their associated mortality impact, across all modes of ECMO and additionally identifies several areas for institutional performance improvement. ALI was the most common complication across all cohorts and was associated with decreased in-hospital survival. Impact of pDPC on ALI was difficult to elucidate due to non-standardized placement patterns. In the care of these medically complex patients on multidisciplinary teams, review of outcomes and group discussions targeting areas for improvement are critical to success; findings of this study resulted in development of a standardized lower extremity perfusion management protocol.
Supplementary Material
Article Highlights.
Type of Research:
Single-center retrospective review of all ECMO patients between January 1st 2019 and December 31st 2020.
Key Findings:
Vascular complication rate was 23.6% within ECMO patients, and was highest within peripheral VA ECMO (33.3%). Acute Limb Ischemia (ALI) was the most common complication (16.5%), and was associated with decreased in-hospital survival (32.5% vs. 54%, Log-Rank=0.023). Prophylactic distal perfusion catheters were used in 60% of peripheral VA ECMO Cases.
Take Home Message:
Vascular complications were prevalent in all ECMO modes. ALI was the most common complication irrespective of ECMO mode, and was associated with decreased in-hospital survival. Non-selective prophylactic distal perfusion catheter usage is increasing in response to ALI events, and the impact of limb loss will be evaluated in the future.
Acknowledgments
Disclosures:
Blakeslee-Carter, J: Supported by the National Institute of Health Agency for Healthcare Research and Quality grant 5T32HS013852
Shao, C: Supported by the National Institute of Health Agency for Healthcare Research and Quality grant 5T32HS013852
Spangler, E L: Supported by National Institute of Health grant KL2 TR 003097
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Makdisi G, Makdisi T, Wang IW. Use of distal perfusion in peripheral extracorporeal membrane oxygenation. Ann Transl Med. 2017;5(5): 103–105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kim H, Cho YH. Role of extracorporeal cardiopulmonary resuscitation in adults. Acute Crit Care. 2020;35(1):1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Extracorporeal Life Support Organization Registry Report; 2021. https://www.elso.org/Registry/Statistics/InternationalSummary.aspx [DOI] [PubMed]
- 4.Tanaka D, Hirose H, Cavarocchi N, Entwistle JWC. The Impact of Vascular Complications on Survival of Patients on Venoarterial Extracorporeal Membrane Oxygenation. Ann Thorac Surg. 2016;101(5):1729–1734. [DOI] [PubMed] [Google Scholar]
- 5.Bonicolini E, Martucci G, Simons J, Raffa GM, Spina C, Coco VL, Arcadipane A, et al. Limb ischemia in peripheral veno-arterial extracorporeal membrane oxygenation: A narrative review of incidence, prevention, monitoring, and treatment. Crit Care. 2019;23(1):266–267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Roussel A, Al-Attar N, Khaliel F, Alkhoder S, Raffoul R, Alfayyadh F, Rigolet M, et al. Arterial vascular complications in peripheral extracorporeal membrane oxygenation support: A review of techniques and outcomes. Future Cardiol. 2013;9(4):489–495. [DOI] [PubMed] [Google Scholar]
- 7.Yau P, Xia Y, Shariff S, Jakobleff WA, Forest S, Lipstiz EC, Scher LA, et al. Factors Associated with Ipsilateral Limb Ischemia in Patients Undergoing Femoral Cannulation Extracorporeal Membrane Oxygenation. Ann Vasc Surg. 2019;54:60–65. [DOI] [PubMed] [Google Scholar]
- 8.Bisdas T, Beutel G, Warnecke G, Hoeper MM, Kuehn C, Haverich A, Teebken OE. Vascular complications in patients undergoing femoral cannulation for extracorporeal membrane oxygenation support. Ann Thorac Surg. 2011;92(2):626–631. [DOI] [PubMed] [Google Scholar]
- 9.Lamb KM, DiMuzio PJ, Johnson A, Batista P, Moudgill N, McCullough M, Eisenberg JA, et al. Arterial protocol including prophylactic distal perfusion catheter decreases limb ischemia complications in patients undergoing extracorporeal membrane oxygenation. J Vasc Surg. 2017;65(4):1074–1079. [DOI] [PubMed] [Google Scholar]
- 10.Kim J, Cho YH, Sung K, Park TK, Lee GY, Lee JM, Song YB, et al. Impact of Cannula Size on Clinical Outcomes in Peripheral Venoarterial Extracorporeal Membrane Oxygenation. ASAIO J. 2019;65(6):573–579. [DOI] [PubMed] [Google Scholar]
- 11.Ohira S, Kawamura M, Ahern K, Cavarocchi N, Hirose H. Aggressive placement of distal limb perfusion catheter in venoarterial extracorporeal membrane oxygenation. Int J Artif Organs. 2020;43(12):796–802. [DOI] [PubMed] [Google Scholar]
- 12.Son AY, Khanh LN, Joung HS, Guerra A, Karim AS, McGregor R, Pawale A, et al. Limb ischemia and bleeding in patients requiring venoarterial extracorporeal membrane oxygenation. J Vasc Surg. 2021:73(2): 593–600. [DOI] [PubMed] [Google Scholar]
- 13.Danial P, Hajage D, Nguyen LS, Mastroianni C, Demondion P, Schmidt M, Bougle A, et al. Percutaneous versus surgical femoro-femoral veno-arterial ECMO: a propensity score matched study. Intensive Care Med. 2018;44(12):2153–2161. [DOI] [PubMed] [Google Scholar]
- 14.Foley PJ, Morris RJ, Woo EY, Acker MA, Wang GJ, Fairman RM, Jackson BM. Limb ischemia during femoral cannulation for cardiopulmonary support. J Vasc Surg. 2010;52(4):850–853. [DOI] [PubMed] [Google Scholar]
- 15.Aubron C, Cheng AC, Pilcher D, Leong T, Magrin G, Cooper DJ, Scheinkestel C, et al. Factors associated with outcomes of patients on extracorporeal membrane oxygenation support: A 5-year cohort study. Crit Care. 2013;17(2):73–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kaushal M, Schwartz J, Gupta N, Im J, Leff J, Jakobleff WA, Leyvi G. Patient Demographics and Extracorporeal Membranous Oxygenation (ECMO)-Related Complications Associated With Survival to Discharge or 30-Day Survival in Adult Patients Receiving Venoarterial (VA) and Venovenous (VV) ECMO in a Quaternary Care Urban Center. J Cardiothorac Vasc Anesth. 2019;33(4):910–917. [DOI] [PubMed] [Google Scholar]
- 17.Ban KA, Berian JR, Ko CY. Does Implementation of Enhanced Recovery after Surgery (ERAS) Protocols in Colorectal Surgery Improve Patient Outcomes? Clin Colon Rectal Surg. 2019;32(2):109–113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kollef MH, Micek ST. Using protocols to improve patient outcomes in the intensive care unit: Focus on mechanical ventilation and sepsis. Semin Respir Crit Care Med. 2010:31(1):19–30. [DOI] [PubMed] [Google Scholar]
- 19.Thuzar M, Malabu UH, Tisdell B, Sangla KS. Use of a standardised diabetic ketoacidosis management protocol improved clinical outcomes. Diabetes Res Clin Pract. 2014;104(1):8–11. [DOI] [PubMed] [Google Scholar]
- 20.Concensus E. ELSO Guidelines for Cardiopulmonary Extracorporeal Life Support; 2017. https://www.elso.org/Resources/Guidelines.aspx
- 21.Cheng R, Hachamovitch R, Kittleson M, Patel J, Arabia F, Moriguchi J, Esmailian F, et al. Complications of extracorporeal membrane oxygenation for treatment of cardiogenic shock and cardiac arrest: A meta-analysis of 1,866 adult patients. Ann Thorac Surg. 2014;97(2):610–616. [DOI] [PubMed] [Google Scholar]
- 22.Makdisi G, Wang IW. Extra Corporeal Membrane Oxygenation (ECMO) review of a lifesaving technology. J Thorac Dis. 2015:7(7):166–176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wagner AM, Phair J, Scher L, Shariff S, Jakobleff W, Lipsitz E. Vascular Complications Associated With Extracorporeal Membrane Oxygenation. J Vasc Surg. 2015;62(3):791–792. [Google Scholar]
- 24.Pellenc Q, Girault A, Roussel A, Aguir S, Cerceau P, Longrois D, Mal H, et al. Preclosing of the femoral artery allows total percutaneous venoarterial extracorporeal membrane oxygenation and prevents groin wound infection after lung transplantation. Eur J Cardio-Thoracic Surg. 2020;58(2):371–378. [DOI] [PubMed] [Google Scholar]
- 25.Augusto R, Passos S, J C, Coelho A, Coelho N, Semiao AC, Pinto E, et al. Arterial Vascular Complications in Peripheral Venoarterial Extracorporeal Membrane Oxygenation Support. Rev Port Cir Cardiothorac Vasc. 2017;24(4):104. [PubMed] [Google Scholar]
- 26.Yeo HJ, Yoon SH, Jeon D, Kim YS, Cho Wh, Kim D, Lee SE. The Utility of Preemptive Distal Perfusion Cannulation During Peripheral Venoarterial Extracorporeal Membrane Oxygenation Support. J Interv Cardiol. 2016;29(4):431–436. [DOI] [PubMed] [Google Scholar]
- 27.Kaufeld T, Beckmann E, Ius F, Koigeldiev N, Sommer W, Mashaqi B, Fleissner F, et al. Risk factors for critical limb ischemia in patients undergoing femoral cannulation for venoarterial extracorporeal membrane oxygenation: Is distal limb perfusion a mandatory approach? Perfus (United Kingdom). 2019;34(6):453–459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Patton-Rivera K, Beck J, Fung K, Chan C, Beck M, Takayama H, Takeda K. Using near-infrared reflectance spectroscopy (NIRS) to assess distal-limb perfusion on venoarterial (V-A) extracorporeal membrane oxygenation (ECMO) patients with femoral cannulation. Perfus (United Kingdom). 2018;33(8):618–623. [DOI] [PubMed] [Google Scholar]
- 29.Steffen Robert J, Sale Shiva, Anandamurthy Balaram, Cruz Vincent B, Grady Patrick M, Edward G Soltesz NM. Using Near-Infrared Spectroscopy to Minitor lower extremities in patients on venoarterial extracorporeal membrane oxygenation. Ann Thorac Surg. 2014;98(5):1853–1854. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.