If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address for reprints: Dominik Wiedemann, MD, Department of Cardiac Surgery, Medical University of Vienna, Waehringer Guertel 18-20, 1090 Vienna, Austria.
To assess the influence of primary arterial access in patients receiving peripheral postcardiotomy extracorporeal life support on associated complications and outcome.
Methods
Of 573 consecutive patients requiring PC-ECLS between 2000 and 2019 at a single center, 436 were included in a retrospective analysis and grouped according to primary arterial extracorporeal life support access site. Survival and rate of access-site–related complications with special emphasis on fatal/disabling stroke were compared.
Results
The axillary artery was cannulated in 250 patients (57.3%), whereas the femoral artery was used as primary arterial access in 186 patients (42.6%). There was no significant difference in 30-day (axillary: 62%; femoral: 64.7%; P = .561) and 1-year survival (axillary: 42.5%; femoral: 44.8%; P = .657). Cerebral computed tomography-confirmed stroke with a modified ranking scale ≥4 was significantly more frequent in the axillary group (axillary: n = 28, 11.2%; femoral: n = 4, 2.2%; P = .0003). Stroke localization was right hemispheric (n = 20; 62.5%); left hemispheric (n = 5; 15.6%), bilateral (n = 5; 15.6%), or infratentorial (n = 2; 6.25%). Although no difference in major cannulation site bleeding was observed, cannulation site change for bleeding was more frequent in the axillary group (axillary: n = 13; 5.2%; femoral: n = 2; 1.1%; P = .03). Clinically apparent limb ischemia was significantly more frequent in the femoral group (axillary: n = 12, 4.8%; femoral: n = 31, 16.7%; P < .0001).
Conclusions
Although survival did not differ, surgeons should be aware of access–site-specific complications when choosing peripheral PC-ECLS access. Although lower rates of limb ischemia and the advantage of antegrade flow seem beneficial for axillary cannulation, the high incidence of right hemispheric strokes in axillary artery cannulation should be considered.
Surgeons should be aware of specific access-site–related adverse events when choosing PC-ECLS access. The high incidence of right hemispheric strokes in axillary artery cannulation must be considered.
PC-ECLS is a lifesaving bailout therapy for patients with postcardiotomy shock, but is associated with relevant mortality and morbidity. Several complications of PC-ECLS are related to the vascular access site. Knowledge of possible cannulation–site-related complications enables surgeons to choose the most suitable peripheral access site for individual patients.
See Commentary on page 1559.
Complexity of cardiac surgical procedures has risen during the past decades. Along with increasing age and morbidity of the patient population, the number of patients requiring postcardiotomy temporary extracorporeal life support (PC-ECLS) steadily increases.
Indications for PC-ECLS are mainly inability to separate from cardiopulmonary bypass, or cardiopulmonary failure of various etiologies during the early postoperative period. Although the primary therapeutic goal is bridging to myocardial recovery and weaning from the device, bridging to durable mechanical circulatory support (MCS) or heart transplantation is an option for eligible patients when ECLS weaning is unlikely or unsuccessful.
Despite technical and medical advances, the mortality of PC-ECLS patients remains high with reported in-hospital survival rates ranging from 25% to 46%.
Early and late outcomes of 517 consecutive adult patients treated with extracorporeal membrane oxygenation for refractory postcardiotomy cardiogenic shock.
Furthermore, PC-ECLS is associated with a number of potentially devastating complications that may arise during therapy and limit outcome, some of which are directly related to the site of cannulation, including limb ischemia and cannulation site bleeding. Cerebrovascular events are another frequent and often fatal complication.
Main advantages over central cannulation are the facilitated sternal closure, the lower risk for infection and mediastinal bleeding, as well as the easier intensive care unit handling and ECLS explantation. A recent meta-analysis also reported higher mortality in patients with central cannulation.
Evidence on which peripheral access site is preferable is still scarce. Different peripheral arterial access sites have variable flow properties related to the location of the arterial cannula in the bloodstream. Cannulation of the axillary artery allows for predominantly antegrade body perfusion and proximal shift of the watershed, therefore preventing differential hypoxemia in patients with compromised pulmonary function.
Feasibility and safety of watershed detection by contrast-enhanced ultrasound in patients receiving peripheral venoarterial extracorporeal membrane oxygenation: a prospective observational study.
Risk of harlequin syndrome during bi-femoral peripheral VA-ECMO: should we pay more attention to the watershed or try to change the venous cannulation site?.
Another obvious advantage of axillary cannulation is the uncompromised leg perfusion, avoiding leg ischemia and related problems. For the stated reasons, indirect axillary artery cannulation via side graft has evolved to be the preferred access site for elective implantation at our center in recent years, whereas the femoral artery remains the access of choice in cases where rapid ECLS initiation is required.
Despite all advantages, cerebral blood flow properties are more likely to be influenced in axillary artery cannulation because of the anatomic proximity of the cannulation site, which might represent a potential source of cerebral embolism. A possible relation of vascular ECLS access site and cerebrovascular adverse events is conceivable, but yet unclear. Furthermore, indirect axillary cannulation via an anastomosed side graft poses a risk of bleeding from the cannulation site.
The aim of this study was to compare axillary and femoral artery cannulation in terms of outcome and the incidence of cannulation-site–related and cerebrovascular adverse events, as well as further assessment of the location of stroke in these patients.
Methods
Study Population
The data of all patients undergoing PC-ECLS at the Department of Cardiac Surgery, Medical University of Vienna, were collected within an institutional database approved by the local ethics committee of the Medical University of Vienna (institutional review board No. 1086/2019), in compliance with the Declaration of Helsinki. Informed consent was waived due to the retrospective study design.
All consecutive patients who received PC-ECLS from 2000 to 2019 (N = 573) at the Department of Cardiac Surgery at the Medical University of Vienna were screened for the inclusion and exclusion criteria stated below.
Inclusion Criteria
Patients aged 18 years or older who received ECLS after cardiopulmonary bypass (CPB) either intraoperatively or within 72 hours postoperatively for postcardiotomy low cardiac output syndrome (LCOS), cardiopulmonary resuscitation, or hemodynamic/respiratory instability of other etiologies necessitating initiation of ECLS.
There is no defined age cutoff for PC-ECLS at our institution, and whenever a patient is determined to be a candidate for cardiac surgery, advanced age is not regarded a contraindication for ECLS if required to treat postcardiotomy shock. However, older patients are less likely to be eligible candidates for durable MCS or heart transplantation (HTX) and therapy may be withdrawn in the case the patient cannot be weaned from ECLS. In this case, a weaning attempt is made and ECLS explanted. HTX age cutoff is 70 years in our center, and although there is no absolute age cutoff for durable MCS, we are more restrictive with durable MCS implantation in patients aged 75 years or older. In cases when we anticipate that the patient will not be weanable from ECLS because of an underlying structural cardiac defect, we are more restrictive with ECLS implementation when the patient is no candidate for ventricular assist device (VAD)/HTX.
Exclusion Criteria
Patients were excluded in the case that 1 or more of the following criteria was met (Figure 1):
•
Central ECLS cannulation,
•
ECLS already installed before surgery/before CPB,
•
Patients who received ECLS for temporary right ventricular support at the time of left VAD implantation,
ECLS implanted >72 hours after the end of surgery,
•
Patients younger than age 18 years,
•
Missing data, and/or
•
Patients transferred on ECLS from another hospital.
Figure 1Study flowchart. Five hundred seventy-three consecutive patients undergoing postcardiotomy extracorporeal life support (PC-ECLS) from 2000 to 2019 were screened for inclusion and exclusion criteria. Four hundred thirty-six patients were included in a retrospective analysis and grouped by primary arterial cannulation site (axillary artery n = 250 or femoral artery n = 186). LVAD, Left ventricular assist device.
Patients who received ECLS >72 hours after surgery were excluded to exclude patients with delayed complications of surgery that occurred after the patient has already been transferred to the regular ward, such as unexpected, brisk bleeding leading to tamponade and cardiopulmonary resuscitation and ECLS implantation under emergency conditions on a normal ward. These patients have unfavorable prognosis due to the unexpected and sudden deterioration in a less-controlled environment (ward vs intensive care unit), and we therefore believe these patients are not comparable to patients receiving ECLS during or shortly after cardiac surgery. However, because the majority of patients who required PC-ECLS at our center received ECLS intraoperatively or within 72 hours after surgery, using a cutoff of 72 hours, only a very low number of patients had to be excluded.
Clinical Definitions and End Points
Indication for ECLS implantation was:
•
Inability to separate from CPB because of LCOS, respiratory or metabolic instability with signs of anaerobic metabolism upon CPB weaning despite optimized supportive measures (ie, adequate vasopressor and inotropic support and fluid status), or
•
Postoperative LCOS or hemodynamic or respiratory instability of any cause not amenable to conservative measures and requiring the implantation of ECLS, or postoperative cardiac arrest.
All-cause 30-day mortality was set as the primary end point. Incidence of cerebrovascular and access-site–related complications (stroke with modified ranking scale [MRS] ≥4, cannulation site bleeding requiring surgical revision and/or change of cannulation site, extremity ischemia and wound healing disorders), as well as all-cause long-term mortality were secondary end points.
Peripheral ECLS Cannulation
For axillary cannulation, the axillary artery is exposed and proximal and distal control is obtained. In the case that the patient is still on CPB at the time of implantation, an activated clotting time of >400 seconds is maintained. In the case that ECLS implantation is performed in a patient who is not on CPB, 5000 IU unfractionated heparin is administered before a side-biting clamp is applied to the vessel. An 8-mm polyethylene terephthalate graft is then anastomosed end-to-side. A 19F or 21F arterial cannula is inserted via the graft with the cannula tip placed 0.5 to 1 cm proximal to the anastomosis and secured to the graft with silk ligatures. A silicone vessel-loop is placed distally around the axillary artery, and a biradial invasive blood pressure monitoring is installed to enable regulation of the perfusion of the right arm. A venous drainage cannula is percutaneously inserted via the femoral vein.
For femoral cannulation, the right or left common femoral artery is either percutaneously cannulated using the Seldinger technique or surgically exposed. A distal perfusion cannula (DPC) is used depending on vessel diameter and surgeon's preference. A representative video of axillary arterial cannulation is provided (Video 1).
Patient Management on ECLS
Anticoagulation on ECLS was performed with continuous intravenous administration of 7.5 to 20 IU/kg/h unfractionated heparin and monitored by activated partial thromboplastin time (aPTT) with a target therapeutic aPTT range of 1.5 to 2.5 × baseline. In case of confirmed or suspected heparin induced thrombocytopenia, argatroban was used instead of heparin with the same target aPTT. Anticoagulation was adapted or discontinued in case of severe mediastinal bleeding.
Patients who receive ECLS during surgery are transferred to an intensive care unit specialized on the postoperative care of patients after cardiac surgery. The decision to initiate ECLS, irrespective of intraoperatively or postoperatively, is made with a multidisciplinary team approach involving anesthesiologists, intensivists as well as the surgeon who performed the surgery.
Although a standardized weaning protocol has been developed at our institution within recent years, this was not available for the majority of the study period. In case of absent recovery, patients are evaluated for durable left ventricular assist device/HTX or ECLS is withdrawn after repeatedly unsuccessful ECLS weaning attempts if the patient is not eligible for left VAD/HTX, after thorough ethical consideration, interdisciplinary discussion of the case and involvement of the patient's family.
Follow-up
Survival data were retrieved from federal statistics (Statistics Austria, Vienna, Austria) and patient records. Hospital records were used to characterize intra- and postoperative course, as well as incidence of cerebral and access site related adverse events.
Cerebral computed tomography (CCT) images were acquired during clinical routine in case of clinical suspicion of stroke and were retrospectively assessed for stroke location. Stroke severity was graded by clinical presentation and neurological assessment using MRS at the time of hospital discharge. Stroke with MRS ≥2 at the time of hospital discharge was classified as a disabling stroke according to the proposed definition for standardized neurological end points for cardiovascular clinical trials published by The Neurologic Academic Research Consortium,
together with evidence of ischemic or hemorrhagic stroke on CCT. Additionally, the rate of patients with MRS ≥4 is reported.
Statistical Analysis
All statistical analyses were performed using Prism for Mac OS version 8.1.2 (GraphPad Software Inc, San Diego, Calif). Data are presented as median and interquartile range (IQR) for continuous variables, and absolute and relative frequencies for categorical variables. Survival was estimated using Kaplan-Meier curves and group comparisons of survival depending on primary arterial cannulation site performed by log rank test. For group comparisons of baseline characteristics, procedural variables, and adverse event rates, Fisher exact test was used for categorical variables, and Mann-Whitney U test for continuous variables. To identify risk factors of fatal/disabling stroke, a multivariable binary logistic regression model was used, including clinically relevant and statistically significant parameters from the bivariate analysis. Two-tailed P < .05 was considered statistically significant.
Results
Patients
Between February 2000 and December 2019 a total of 573 patients required PC-ECLS at our center. After application of inclusion and exclusion criteria listed in Figure 1, 436 patients with primary peripheral ECLS cannulation were included in the analysis and divided into groups according to the primary arterial access site (axillary artery, n = 250 or femoral artery, n = 186).
Baseline Characteristics
Patients with primary axillary arterial cannulation were older, had a higher European System for Cardiac Operative Risk Evaluation II score (EuroSCORE II), higher prevalence of coronary artery disease, lower preoperative bilirubin level, aspartate aminotransferase level, and hemoglobin level, and a higher prevalence of peripheral arterial disease. There was no significant difference in the other studied baseline variables. Baseline characteristics are presented in Table 1.
Table 1Baseline characteristics of the study cohort grouped by primary arterial extracorporeal life support (ECLS) access site
Values are presented as median (interquartile range) for continuous variables and absolute numbers (%) categorical variables. EuroSCORE, European System for Cardiac Operative Risk Evaluation score BMI, body mass index; VAD, ventricular assist device; LVEF, left ventricular ejection fraction; ASAT, aspartate aminotransferase; ALAT, alanine aminotransferase; gamma-GT, gamma-glutamyl transferase; TIA, transient ischemic attack; HTX, heart transplantation.
Of 436 patients with peripheral ECLS cannulation, the axillary artery was cannulated in 250 patients (57.3%), whereas the femoral artery was used as primary access site in 186 patients (42.6%). In case of axillary cannulation, the right axillary artery was used in the majority of patients (n = 242; 96.8%), whereas the left axillary artery was used infrequently (n = 8; 3.2%).
Operative Data and ECLS Indication
Procedure duration, CPB and aortic crossclamp (Xclamp) time were significantly longer in the axillary group, and the rate of surgery for type-A aortic dissection was higher in the axillary group. Periprocedural data are shown in Table 2. The rate of patients requiring cardiopulmonary resuscitation (CPR) before ECLS initiation, as well as the number of patients implanted under ongoing CPR was significantly higher in the group with primary femoral cannulation.
Table 2Procedure data of the study cohort grouped by primary arterial extracorporeal life support (ECLS) access site
Indication for ECLS was intraoperative failure to separate from CPB (73.6%), or postoperative LCOS or hemodynamic/metabolic instability/cardiopulmonary resuscitation within 72 hours from surgery (26.4%). The duration of ECLS was significantly longer in the axillary group (4.6 days [IQR, 2.9-7.1 days] vs 4 days [IQR, 2.5-6.1 days]; P = .044) (Table 3).
Table 3Extracorporeal life support (ECLS) indications, run details, and adverse event rates of the study cohort grouped by primary arterial ECLS access site
Variable
Total study cohort (N = 436)
Indirect axillary (n = 250)
Femoral (n = 186)
P value
ECMO indication
CPB weaning failure (implanted during initial surgery)
321 (73.6)
188 (75.2)
133 (71.5)
.4419
Hemodynamic decline/CPR/respiratory failure after cardiac surgery (implanted within 72 h after end of surgery)
There was no significant difference in survival after 30 days (axillary: 62%; femoral: 64.7%; P = .561) and 1 year (axillary: 42.5%; femoral: 44.8%; P = .657) after ECLS initiation between patients with primary arterial ECLS cannulation of the axillary artery and femoral artery. There also was no difference in overall survival (P = .766) (Figure 2).
Figure 2Kaplan-Meier estimates of survival. Survival in patients with axillary (ax) and femoral (fem) arterial extracorporeal life support (ECLS) cannulation is visualized by Kaplan-Meier curves, and survival compared between the groups by log rank test. There was no significant difference in survival between patients with ax and fem primary arterial ECLS access site (ax vs fem P = .766). The shaded areas represent the 95% CI.
Older age, lower preoperative glomerular filtration rate, longer duration of ECLS support, and performed procedure were significantly associated with mortality in a Cox proportional hazards model for mortality (Tables E1 and E2).
Cerebral and Access-Site–Related Adverse Events
Incidence of CCT-confirmed stroke with an MRS ≥4 was significantly higher in the axillary group (axillary: n = 28, 11.2% vs femoral: n = 4, 2.2%; P = .0003). Although there was no significant difference in major cannulation site bleeding requiring surgical revision, change of cannulation site because of bleeding was significantly more frequent in the axillary group (axillary: n = 13; 5.2% vs femoral: n = 2; 1.1%; P = .03).
Clinically apparent limb ischemia (axillary: n = 12, 4.8%; femoral: n = 31, 16.7%; P < .0001) was significantly more frequent in the femoral group. The rate of complications arising from limb ischemia in patients with femoral arterial cannulation were reduced with use of a DPC; however, this reduction lacked statistical significance (Table E3). Of 12 patients with limb ischemia in the axillary group, the right arm was affected in 8 patients and leg ischemia occurred in 4 patients. The likely origin/pathomechanism of limb ischemia in patients with axillary arterial cannulation, as well as resulting morbidity, is described in detail in Table E4. Moreover, wound healing disorders requiring surgical intervention were significantly more frequent in the femoral group (Table 3).
Of a total of 32 cases of stroke with MRS ≥4 in the study cohort, localization was right hemispheric in the majority of cases (n = 20; 62.5%), left hemispheric (n = 5; 15.6%), bilateral (n = 5; 15.6%), and infratentorial (n = 2; 6.25%) (see Table 4 and Figure 3).
Table 4Stroke localization
Primary arterial cannulation site
Total No. of patients with stroke MRS ≥4
Right hemispheric
Left hemispheric
Bilateral
Infratentorial
Total study population (N = 436)
32 (7.3)
20 (62.5)
5 (15.6)
5 (15.6)
2 (6.25)
Axillary artery (n = 250)
28 (11.2)
18 (64.3)
4 (14.3)
4 (14.3)
2 (7.1)
Femoral artery (n = 186)
4 (2.2)
2 (50)
1 (25)
1 (25)
0 (0)
Values are presented as absolute numbers (%). MRS, Modified ranking scale.
Figure 3The aim of the present study (section 1), the definition of the study cohort and end points (section 2), as well as the main results and conclusion of the study (sections 3and4). Although no difference in survival between patients with femoral (fem) and axillary (ax) arterial extracorporeal life support (ECLS) cannulation was observed, the rate of stroke and cannulation site change for bleeding was significantly higher in the axillary group, whereas the incidence of limb ischemia and cannulation site wound healing disorders was significantly higher in the femoral group. The distribution of stroke localization in 32 patients with stroke with a modified ranking scale (MRS) ≥4 is shown. PC-ECLS, Postcardiotomy extracorporeal life support.
Comparing baseline and periprocedural data of patients with (n = 32) and without (n = 404) fatal/disabling stroke, patients with severe stroke had significantly longer ECLS run duration (P = .034), longer aortic Xclamp time during main surgery (P = .004), higher rates of primary axillary arterial cannulation (P = .0003), aortic surgery (P = .043), and CPR before implantation (P = .174). All results of the bivariate analysis are depicted in Table E5. To identify risk factors for fatal/disabling stroke, the following variables were included in a multivariable binary logistic regression model: ECLS run duration, aortic Xclamp time, initial arterial cannulation site, and aortic surgery.
Primary axillary arterial ECLS cannulation was identified as the strongest risk factor for severe stroke with an adjusted odds ratio of 4.51. Additionally, ECLS duration and aortic Xclamp time were identified as risk factors (Table 5).
Table 5Risk factors for stroke modified ranking scale (MRS) ≥4 were identified using a binary logistic regression model, including clinically relevant and significant parameters from the bivariate analysis
PC-ECLS is often a rescue therapy for patients with cardiopulmonary failure after cardiac surgery not amenable to conservative measures, which is reflected by a tremendously high mortality. Previous studies report in-hospital survival rates after PC-ECLS ranging from 25% to 46%.
Early and late outcomes of 517 consecutive adult patients treated with extracorporeal membrane oxygenation for refractory postcardiotomy cardiogenic shock.
In our study cohort, survival after 1 year from PC-ECLS implantation was 42.5% in patients with primary axillary cannulation, and 44.8% in patients with primary femoral cannulation.
Although patients requiring PC-ECLS may have little or no chance to survive without temporary circulatory support, ECLS itself poses a considerable source of related complications that for themselves carry a high burden of morbidity and mortality.
These ECLS-related complications might be partly related to general and often inevitable circumstances of ECLS such as increased risk of surgical site bleeding under systemic anticoagulation, but others might be preventable and represent a vantage point for improving the outcome of patients on PC-ECLS, such as complications related to the choice of cannulation site.
In the present study we evaluated 2 different peripheral arterial cannulation strategies for PC-ECLS in a cohort of 436 patients treated at a single center in terms of survival and incidence of cerebral and access-site–related adverse events. Although femoral arterial cannulation is a commonly used access site at many centers, indirect axillary artery cannulation has evolved to be a preferred arterial access site for PC-ECLS at our center for the benefit of predominantly antegrade body perfusion, avoidance of lower limb ischemia, as well as reduction of differential hypoxemia (ie, Harlequin effect) in patients with impaired pulmonary function.
Feasibility and safety of watershed detection by contrast-enhanced ultrasound in patients receiving peripheral venoarterial extracorporeal membrane oxygenation: a prospective observational study.
Risk of harlequin syndrome during bi-femoral peripheral VA-ECMO: should we pay more attention to the watershed or try to change the venous cannulation site?.
can be addressed easily by application of a silicone vessel loop distal to the cannulation site and biradial invasive blood pressure monitoring to enable regulation of distal blood flow.
In the setting of CPB weaning failure where a direct switch from CPB to ECLS is conducted, there is usually enough time to allow the more time-consuming indirect axillary artery cannulation. In emergency situations or whenever rapid initiation of ECLS is required, the femoral artery is still the preferred access site.
Although survival did not differ between axillary and femoral arterial cannulation in our study (30-day survival: axillary: 62% vs femoral: 64.7%) we found a different spectrum and incidence of access-site–related adverse events. In particular, we observed a significantly higher rate of severe stroke in patients cannulated via the right axillary artery (axillary: 11.2% vs femoral: 2.2%), predominantly these strokes were located in the right hemisphere. Pisani and colleagues
described a low incidence of local complications in axillary arterial cannulation in a mixed cohort of PC and non-PC ECLS patients; however, cerebral complications were not assessed. Toivonen and colleagues
found that neurologic injury (ischemic or hemorrhagic stroke) is a common complication in PC-ECLS patients with a reported incidence as high as 19% and associated with adverse outcome in a multicenter meta-analysis of 781 PC-ECLS patients; however, the influence of the arterial access site was not addressed.
Hypotheses for the high incidence of right hemispheric strokes are the anatomical proximity of the cannulation site together with flow turbulences at the anastomosis site. Especially in patients with smaller subclavian arteries and higher flow requirements, this might play an underestimated role.
On the other hand, in some cases there is a chronological relation between ECLS weaning and onset of stroke (Table E6), although due to the retrospective design of the study, we could not prove this point. Nevertheless, it is recommended not to keep the patient on lower flow for a too-long time period.
Measures to avoid thromboembolism during explant include proximal clamping of the subclavian/axillary artery during the explant procedure; however, this does not protect against thromboembolism from the cannulation site during the ECLS run. Based on the results of the present analysis, we consider it important to implement a standard operating procedure for axillary ECLS explant because a relevant proportion of major strokes seem to be in relation to the explant procedure. However, we emphasize that we cannot yet tell whether or not these measures are able to mitigate the occurrence of stroke in any way, and this must be subject of future studies. We suggest achieving proximal and distal control of the axillary artery before stopping ECLS and excluding the vessel from circulation. The cannula should then be removed, and the graft inspected for any thrombus formation, which is removed carefully. If a thrombus is present in the graft, a thrombectomy should then be performed from distally and proximally, before the artery is unclamped, the graft flushed in a retrograde fashion and ligated. To avoid thrombus formation at the cannulation site, adequate anticoagulation should be maintained, especially during the weaning phase, and a structured weaning protocol followed to avoid prolonged low-flow promoting thrombus formation during this phase. Because ECLS duration was also identified as a risk factor for stroke in this study, another important conclusion is that explantation of ECLS should never be unnecessarily delayed (eg, for logistic reasons) because longer ECLS run duration may increase the risk for adverse events and also for mortality, which has already been suggested by the results of previous studies,
and was also confirmed in the present study. Explantation of indirect axillary cannulated ECLS following the abovementioned standard operating procedure is uncomplicated and can be performed in the intensive care unit in most cases, thus not requiring any operating room capacity. Potentially, stroke incidence can be reduced by strictly adhering to these measures; however, surgeons should be aware of the specific risk in axillary cannulation.
We observed no significant difference in cannulation site bleeding requiring surgical revision; however, change of cannulation site for severe bleeding was significantly more frequent in the axillary group. This may suggest that bleeding at the axillary cannulation site can be more difficult to manage and treat with success, thus requiring change to another site more frequently. This finding is in line with the results of previous studies investigating axillary ECLS cannulation.
Of note, bleeding at the axillary cannulation site and hematoma formation can lead to malperfusion of the right arm and a vicious circle causing further complications, and this was the cause of upper extremity ischemia in 5 of 12 patients with extremity ischemia in the axillary cannulation group in our study (Table 5).
As expected, we saw a significantly higher rate of clinically apparent limb ischemia in the group with femoral cannulation, and the rate of limb ischemia necessitating cannulation site change was more frequent in the femoral group. This was to be expected; however, rates of severe associated complication such as fasciotomy and amputation were similar in both groups.
There was a visible reduction of limb ischemia and associated complications with the use of a DPC in our study; however, these reductions missed statistical significance and the study was probably underpowered to prove a benefit of DPC utilization. Additionally, although it became standard to implant a DPC at the time of ECLS implantation within recent years, this was not the case in earlier years where DPCs were often only implanted with already apparent clinical limb ischemia (Table E3).
We believe that DPC placement right at the time of femoral ECLS implantation, as well as close clinical monitoring of the extremity to detect limb ischemia before complications arise, is of paramount importance and may be able to reduce the problem of severe limb complications with femoral ECLS cannulation. Furthermore, it is important to understand that axillary cannulation likewise carries a risk of limb complications, often related to cannulation site bleeding and subsequent impaired perfusion of the arm (Table E4).
Despite the higher incidence of stroke, survival in the axillary group was the same as in the femoral group. The reasons for this might be that the femoral cannulation strategy has other disadvantages like the higher incidence of leg ischemia and the higher left ventricular afterload together with poorer oxygenation of the upper body. Therefore, according to our data, we cannot clearly recommend one cannulation site over the other but obviously the occurrence of major stroke is a relevant obstacle of the right axillary artery as cannulation site. Limb complications occurring with femoral access can be reduced by DPC utilization and close clinical monitoring. Furthermore, it should be emphasized that limb complications are not limited to femoral access but also occur in axillary cannulation, often as a consequence of cannulation site bleeding and hematoma formation. Pros and cons of each cannulation site need to be taken into consideration for each individual patient. Additionally, the quality of cannulation in each technique is highly dependent on the ECLS implanting surgeon. Attention must be paid to every detail: appropriate cannula selection, meticulous performance of the anastomoses of the graft to the axillary artery, as well as protection of leg perfusion in case of femoral arterial cannulation are essential. Anastomosis site bleeding as well as thrombus formation within the system need to be avoided, because a technically not optimally implanted ECLS might be the source of severe complications regardless of the site of cannulation.
Nevertheless, our study shows that the primary access site can significantly influence the outcome of an individual patient. In our eyes, a prospective randomized trial comparing arterial ECLS access would be justified and necessary. Moreover, other options for primary access site need to be addressed such as cannulation of the left instead of the right subclavian artery, indirect femoral cannulation with a site graft, and even direct cannulation of the axillary artery in comparison to the indirect option.
Limitations
The retrospective design and potential for a historical bias are inherent to the present study because indirect axillary cannulation evolved to be the preferred access site at out center within the recent years, whereas femoral cannulation was more common in earlier years (Figure E1). Axillary cannulation is mostly used for ECLS access in controlled situations, such as weaning failure from CPB, allowing for cannulation while the patient is still on CPB, whereas femoral access naturally is preferred in emergency settings. Also, the management of ECLS patients in an intensive care unit is likely to have changed over time. A randomized trial evaluating different peripheral ECLS access sites for postcardiotomy ECLS is justified. Whereas preceding ECLS related events leading to stroke were detectable in some of the patients, it is not possible in every case to differentiate between perioperative and ECLS related origin of strokes in a retrospective study, especially in patients undergoing procedures with a high risk of perioperative stroke, such as patients with aortic dissections. Furthermore, strokes without clinical neurological symptoms in sedated patients were not apprehended because no routine CCTs were performed in absence of clinical neurological deficits.
Conclusions
Surgeons should be aware of different complication profiles when choosing peripheral arterial access site for postcardiotomy ECLS. Although lower rates of limb ischemia and the advantage of antegrade flow seem beneficial for axillary cannulation, especially the high incidence of right hemispheric strokes in axillary artery cannulation should be considered.
Dr Wiedemann is a proctor for Abbott and Medtronic. All other authors reported no conflicts of interest.
The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.
Representative video of axillary arterial cannulation for postcardiotomy extracorporeal life support (PC-ECLS) using an 8-mm polyethylene terephthalate side graft. Video available at: https://www.jtcvs.org/article/S0022-5223(21)01659-7/fulltext.
Representative video of axillary arterial cannulation for postcardiotomy extracorporeal life support (PC-ECLS) using an 8-mm polyethylene terephthalate side graft. Video available at: https://www.jtcvs.org/article/S0022-5223(21)01659-7/fulltext.
Appendix E1
Figure E1Histogram showing the utilization of axillary and femoral arterial postcardiotomy extracorporeal life support (PC-ECLS) cannulation over the study period.
Table E3Distal perfusion cannula (DPC) utilization and complications related to extremity ischemia in patients with femoral arterial extracorporeal life support (ECLS) cannulation
Complication
All patients (N = 186)
DPC at implant (n = 121)
No or delayed DPC (n = 65)
P value
Leg ischemia
31 (16.7)
17 (14)
14 (21.5)
.2181
Compartment syndrome
14 (7.5)
7 (5.8)
7 (10.8)
.2500
Fasciotomy
13 (7)
6 (5)
7 (11)
.2256
Amputation
3 (1.6)
2 (1.7)
1 (1.5)
>.9999
Cannulation site change due to leg ischemia
7 (3.8)
2 (1.7)
5 (7.7)
.0518
Values are presented as absolute numbers (%). DPC, Distal perfusion cannula.
Table E6Characterization of all strokes (N = 48) with timing and clinical indication for cerebral computed tomography (CCT), preceding extracorporeal life support (ECLS)–related events, and possible causes
Cannulation site group
Patient
Timing of CCT
Indication for CCT
Hemorrhagic/ischemic
MRS at discharge
Likely cause of stroke
Axillary artery
1
Respiratory weaning/after explant
Reduced vigilance
Hemorrhagic
1
Unknown
2
Respiratory weaning/after explant
Left-sided hemiplegia
Ischemic
4
ECLS
3
Respiratory weaning/after explant
Left-sided hemiparesis
Ischemic
4
ECLS
4
Respiratory weaning/after explant
Coma
Hemorrhagic
6
ECLS
5
Directly after explant
Sudden onset fixed dilated pupils
Hemorrhagic
6
ECLS explant
6
2 d after explant
Left-sided hemiplegia, decrease of right-sided near-infrared spectroscopy to 29% directly after ECLS explant
Ischemic
4
ECLS explant
7
Respiratory weaning/after explant
Reduced vigilance
Ischemic
2
ECLS
8
Respiratory weaning/after explant
Coma
Ischemic
4
ECLS
9
After extubation
Paresis of right arm
Ischemic
2
ECLS or surgery related (type a aortic dissection)
10
2 wk after ECLS explant, on regular ward before discharge
Visual field loss
Ischemic
1
Unknown/patient also had type a dissection
11
During ECLS run (day 6)
Sudden onset fixed dilated pupils on day 6 during ECLS run
Hemorrhagic
6
ECLS
12
Respiratory weaning/after explant
Reduced vigilance
Ischemic
1
Unknown, patient also underwent perioperative CPR
13
Respiratory weaning/after explant
Left-sided hemiplegia
Ischemic
4
Patient also had mechanical mitral valve thrombosis
14
During ECLS run (day 3)
Sudden onset anisocoria at day 3 of ECLS
Ischemic
6
ECLS
15
Respiratory weaning/after explant
Coma, leftsided hemiplegia, embolectomy of right brachial artery after ECLS explant
ischemic
6
ECLS related: CCT at beginning of ECLS normal, embolectomy right brachial artery after explant
16
Respiratory weaning/after explant
Tetraparesis
Ischemic
5
ECLS
17
Respiratory weaning/after explant
Tetraplegia and coma
Ischemic
6
Known left atrial thrombus and systemic embolism, also to left leg
Sudden onset fixed dilated pupils at day 3 of ECLS
Hemorrhagic
6
On ECLS
21
Respiratory weaning/after explant
Hemiplegia
Ischemic
4
ECLS
22
During ECLS run (day 4)
Sudden onset anisocoria on day 4 of ECLS
Hemorrhagic
6
On ECLS
23
Respiratory weaning/after explant
NCSE
Ischemic
5
Unknown, also underwent CPR
24
Respiratory weaning/after explant
Aphasia, leftsided Hemiparesis
Ischemic
3
ECLS explanted because of device thrombosis despite adequate anticoagulation
25
Respiratory weaning/after explant
Hemiplegia
Ischemic
4
ECLS, arterial cannula was changed due to thrombus formation
26
During ECLS run
Seizure
Ischemic
2
ECLS
27
Respiratory weaning/after explant
Reduced vigilance
Ischemic
1
Unknown
28
Respiratory weaning/after explant
Left-sided hemiplegia and dysphagia
Ischemic
4
Patient had normal CCT 2 d before ECLS explant, large right sided ischemic stroke in CCT 3 d after ECLS explant
29
After arterial cannula change (same day)
Sudden onset anisocoria after cannula exchange
Hemorrhagic
6
Normal CCT 2 d prior to explant, onset of anisocoria after arterial cannula change due to thrombus formation
30
Respiratory weaning/after explant
Myoclonia
Ischemic
5
Unknown, potentially ECLS related
31
Respiratory weaning/after explant
Reduced vigilance, positive Babinski right side
Ischemic
4
ECLS
32
Directly after explant
fixed dilated pupils after explant, thrombus in arterial cannula
Ischemic
6
2 d after explant, occlusion of the right internal carotid artery was diagnosed and patient underwent thrombectomy; however, patient developed a fatal stroke. Thrombotic material in the arterial cannula was noted at the time of explant
33
On ECLS (day 3)
Sudden onset fixed dilated pupils
Hemorrhagic
6
ECLS
34
Respiratory weaning/after explant
Coma
Ischemic
6
Unknown, 30 min CPR
35
During ECLS (day 5)
Seizure
Ischemic
1
ECLS
36
On ECLS (day 3), after revision for mediastinal bleeding
Sudden onset fixed dilated pupils
Hemorrhagic
6
Hypertensive phase during revision for bleeding; fixed unresponsive pupils after revision→CCT
37
Respiratory weaning/after explant
Reduced vigilance and seizure
Ischemic
3
Unknown
38
During ECLS run (day 24)
Unknown
Ischemic
6
ECLS
39
Respiratory weaning/after explant
Hemiparesis and aphasia after extubation
Ischemic
4
ECLS
40
During ECLS (day 17)
Seizures, dilated pupils
Hemorrhagic
6
ECLS
41
Respiratory weaning/after explant
Unknown
Ischemic
1
Unknown, also had aortic dissection
Femoral artery
1
Respiratory weaning/after explant
Reduced vigilance
Ischemic
1
ECLS
2
Respiratory weaning/after explant
Reduced vigilance
Ischemic
4
Embolic? Patient also underwent CPR before implant
3
Respiratory weaning/after explant
Unknown
Ischemic
1
Watershed infarct, CPR before implant
4
During ECLS run (day 6)
Anisocoria
Ischemic
6
On ECLS, cannulation site was changed from femoral to axillary artery 4 d before the event
5
Respiratory weaning/after explant
Hemiparesis
Ischemic
3
Unknown, perioperative CPR
6
On ECLS (day 8)
Anisocoria during ECLS run
Ischemic
6
ECLS, Cannulation site was changed from femoral to axillary artery on day 4 after implant
7
On ECLS (day 7)
Evaluation for durable left ventricular assist device implantation
Ischemic
6
Mechanical mitral valve thrombosis
CCT, Cerebral computed tomography; MRS, modified ranking scale; ECLS, extracorporeal life support; CPR, cardiopulmonary resuscitation; NCSE, nonconvulsive status epilepticus.
Early and late outcomes of 517 consecutive adult patients treated with extracorporeal membrane oxygenation for refractory postcardiotomy cardiogenic shock.
Feasibility and safety of watershed detection by contrast-enhanced ultrasound in patients receiving peripheral venoarterial extracorporeal membrane oxygenation: a prospective observational study.
Risk of harlequin syndrome during bi-femoral peripheral VA-ECMO: should we pay more attention to the watershed or try to change the venous cannulation site?.
With great interest we read the study by Schaefer and colleauges,1 who conducted a detailed analysis of the outcomes of postcardiotomy venoarterial extracorporeal membrane oxygenation (VA-ECMO) support focusing on stroke and cannulation-related complications. The stroke rate of right axillary (RAX) VA-ECMO was greater than that of femoral cannulation. In both axillary and femoral VA-ECMO, the right hemisphere was the most common stroke location (64.5% in RAX and 50% in femoral). This stroke laterality trend in RAX cannulation was similar in our experiences.
Postcardiotomy extracorporeal life support (PC-ECLS) is an essential component of contemporary cardiac surgical care. Temporary PC-ECLS can permit recovery of native myocardial or pulmonary function, correction of underlying derangements precluding the same, or allow time for bridging to more durable forms of support, as needed. The recent seminal European Association for Cardio-Thoracic Surgery, the Extracorporeal Life Support Organization, the Society of Thoracic Surgeons, and the American Association for Thoracic Surgery expert consensus statement provided recommendations on several aspects for PC-ECLS; however, detailed discussion of cannulation strategies was less granular in detail.
01659-7&id=fx1.jpg){kind=link}
01659-7&id=fx2.jpg){kind=link}
01659-7&id=gr1.jpg){kind=link}
01659-7&id=gr2.jpg){kind=link}
01659-7&id=gr3.jpg){kind=link}
01659-7&id=fx4.jpg){kind=link}
01659-7&id=fx5.jpg){kind=link}