Original Contribution

Comparison of Percutaneous Closure Versus Surgical Femoral Cutdown for Decannulation of Large-Sized Arterial and Venous Access Sites in Adults After Successful Weaning of Veno-Arterial Extracorporeal Membrane Oxygenation

Nicolas Majunke, MD;  Norman Mangner, MD;  Axel Linke, MD;  Enno Boudriot, MD;  Sandra Erbs, MD;  Franziska Tietz, MD;  Sabrina Wolff, MD;  Stephan Sch√ºrer, MD;  Gerhard Schuler, MD;  Marcus Sandri, MD

Nicolas Majunke, MD;  Norman Mangner, MD;  Axel Linke, MD;  Enno Boudriot, MD;  Sandra Erbs, MD;  Franziska Tietz, MD;  Sabrina Wolff, MD;  Stephan Sch√ºrer, MD;  Gerhard Schuler, MD;  Marcus Sandri, MD

Abstract: Objective. Surgical femoral cutdown for decannulation after veno-arterial extracorporeal membrane oxygenation (VA-ECMO) is considered standard practice. However, access-site complications with this technique are not rare. The objective of this study is to evaluate feasibility, safety, and efficacy of a complete percutaneous decannulation procedure after VA-ECMO compared with the conventional surgical cutdown approach. Methods. In 35 patients who were successfully weaned from VA-ECMO support, femoral artery and vein access sites were closed using a completely percutaneous approach in 15 patients, whereas 20 patients had conventional surgical cutdown for access-site closure. Data concerning all 35 patients were collected retrospectively and analyzed regarding immediate vascular closure success, associated complications, and clinical outcomes. Results. Technical deployment success of the percutaneous vascular closure devices was achieved in all patients. Immediate success of closure was achieved more frequently in the surgical group (29% vs 100%; P<.05). Severe wound complications requiring surgery occurred only in the surgical group (0% vs 35%; P=NS). Surgical cutdown was associated with a significantly greater need for transfusion of packed red blood cells (1.6 ± 1.4 vs 2.2 ± 1.2; P<.05). Mean hospital stay was shorter in the percutaneous group (32 ± 18 days vs 36 ± 12 days; P=NS). One patient in the surgical group complained about sustained paresthesia after discharge. Conclusions. Complete percutaneous closure of the femoral access site after VA-ECMO is feasible, effective, and safe when compared with conventional surgical closure and performed by experienced operators.  

J INVASIVE CARDIOL 2016;28(10):415-419

Key words: cardiogenic shock, veno-arterial extracorporeal membrane oxygenation, VA-ECMO

The use of veno-arterial extracorporeal membrane oxygenation (VA-ECMO) as a treatment option for temporary circulatory support in refractory cardiogenic shock can improve short-term and long-term survival in adults and is becoming an increasingly accepted procedure.1

In VA-ECMO, blood is drained via venous cannulation from the vena cava and returned oxygenated to the cannulated arterial system while bypassing the heart and lungs, and allows for full cardiac and pulmonary support. 

To maintain high blood flow during ECMO, large-caliber sheaths are required. The standard approach for placement of the arterial and venous cannulae is via the femoral vessels using the Seldinger technique. This can be performed using a completely percutaneous approach or after surgical cutdown.

Available options for access-site management after successful weaning of ECMO include manual compression following the usage of a mechanical femoral compression system to induce hemostasis2 or a surgical cutdown technique, which is considered standard practice after weaning from VA-ECMO, especially in patients with a surgical cutdown during implantation but also after percutaneous cannulation.2,4

Nevertheless, the use of femoral vessel cutdown and surgical closure may be accompanied by an increased rate of bleeding, wound infection, and delayed mobility. 

We developed a complete percutaneous approach for closure of the femoral access site after VA-ECMO to protect the patient from the potential hazards associated with surgical cutdown. 

In the present study, we compare clinical results using a percutaneous suture-mediated closure technique in combination with a collagen-based closure device with surgical closure for access-site management following successful weaning from VA-ECMO. 


Patients. We retrospectively studied patients undergoing percutaneous closure of femoral arterial and venous access sites using conventional vascular closure systems or undergoing conventional surgical cutdown for decannulation after successful weaning from VA-ECMO between February 2010 and March 2015. Since May 2014, the complete percutaneous approach has replaced surgical cutdown for the management of large-sized arterial and venous access sites after weaning from VA-ECMO at our center.

Twenty patients underwent surgery, while 15 patients were treated by percutaneous intervention using vascular closure systems.

In all patients, initial cannulation was performed by experienced interventional cardiologists in the catheterization laboratory under angiographic guidance using a percutaneous approach. The arterial cannula was implanted into the common femoral artery and the venous cannula was implanted into the common femoral vein. The size of cannula used was based on the size of the patient, the size of the vessel to be cannulated on angiography, and the predicted blood flow. An antegrade peripheral leg cannula was inserted into the superficial femoral artery to perfuse the ipsilateral distal leg whenever possible. No arterial evaluation with duplex ultrasound of the vascular system at the access site was performed before cannulation.

Procedural description. In the percutaneous group, cannulas were explanted in a standard fashion at the patient’s bedside or the cath lab. The Perclose ProGlide system (Abbott Vascular) and the AngioSeal device (St. Jude Medical) were used in all procedures. The mechanisms and technical features of the Perclose ProGlide device and AngioSeal device have been reported in detail previously.5,9 Briefly, the Perclose ProGlide device applies a non-absorbable suture to achieve closure of the femoral access site. The AngioSeal consists of an intravascular anchor connected to an extravascular collagen sponge by a self-tightening absorbable suture. When inserted, hemostasis is achieved by sandwiching the arterial puncture site between the intravascular anchor and the extravascular collagen plug.

Once the decision is made to decannulate, the arterial and venous lines are clamped. First, the arterial cannula is cut through and a guidewire is placed via the arterial cannula. Thereafter, the arterial cannula is completely removed while an assistant is manually compressing the femoral artery to avoid bleeding. After the sheath is removed, the first Perclose ProGlide device is advanced over the wire to the femoral artery and the first sutures are deployed. Before tying and securing the first suture, a second Perclose ProGlide device is advanced over the reinserted guidewire at an angle of about 45° to the first suture. If there is still relevant bleeding at the access site, a third or fourth suture is applied. Finally, an AngioSeal device is deployed in a conventional manner to augment the suture closure, irrespective of whether complete hemostasis was already achieved after suture closure. After hemostasis is achieved at the arteriotomy site, the venous femoral cannula is removed using a figure-8 stitch or a Perclose ProGlide device (Figure 1). Due to safety reasons, a sandbag was routinely put on the puncture site even when immediate hemostasis was achieved. In cases of oozing from the puncture site, a compression bandage was applied for 4-6 hours. 

In the surgical group, decannulation was performed at the patient’s bedside or in the operating room. General anesthetic techniques were used. The common femoral arteries and femoral veins are exposed and a single purse-string suture is applied around the arterial and venous sheath. Thereafter, the sheath is removed and the purse-string is tied. Pulse pressure is checked distal to the tied purse-string after vessel closure. Limited endarterectomy at the access site is performed as indicated. Once hemostasis is ensured, the wound is closed in layers with absorbable running suture and finally staples to close the skin.

In both groups, the puncture sites and the peripheral arterial circulation of the involved extremity were assessed clinically for ischemia and wound complications every 2 hours during the first 24 hours. Additionally, adequate flow was assessed using Doppler ultrasound semi-daily for the first 24 hours. Thereafter, the involved groin and extremity were monitored clinically 3 times daily during the stay on the intensive care unit.

During follow-up, patients or their general practitioners were contacted by phone. 

Statistical analysis. Continuous variables are expressed as mean ± standard deviation or as median and interquartile range where appropriate. Categorical variables are presented as percentage. Nominal variables were compared with the chi-square or Fisher’s exact test where appropriate, and continuous variables were compared using the t-test. A P-value <.05 was considered statistically significant.


Fifteen patients had their arterial and venous access sites closed with conventional vascular closure systems, while 20 patients were treated with surgical cutdown. There were no significant differences with regard to patient characteristics (Table 1). 

Percutaneous group. Percutaneous closure was technically successful in all patients (100%). Immediate hemostasis was achieved in 5 patients (29%) after decannulation. In the remaining patients, a compression bandage was applied for a few hours due to light oozing from the puncture sites. Finally, hemostasis of the arterial and venous access sites was achieved in all patients treated percutaneously. 

During hospital stay, 2 patients (13%) developed vascular complications on the former arteriotomy site. In both patients, angiography showed a total arterial occlusion (left external iliac artery in 1 patient and common femoral artery in 1 patient) due to a dissection and were followed by successful recanalization and stenting of the lesion. In 1 patient, a second procedure was necessary due to an early in-stent stenosis. This was treated successfully by percutaneous recanalization (Table 2). Both patients were discharged without sequelae and can exercise without limitation.

Two additional patients (12%) developed a local wound infection at the access site that required antibiotic treatment. The wound healed completely. In-hospital mortality in this group was 33% (n = 5). The majority of the patients died due to multiple organ failure and no death was related to the explantation procedure.

Surgical group. In the surgical group, immediate hemostasis was achieved in all patients (100%). A total of 7 patients (35%) developed wound complications at the access site. Wound dehiscence occurred in 2 patients, requiring surgical revision. The remaining 5 patients were diagnosed with an early deep groin infection. All patients were treated successfully with debridement, antibiotics, and vacuum-assisted closure (VAC) therapy. In 1 patient, critical limb ischemia occurred due to a superficial femoral artery in-stent restenosis. The patient was treated successfully by percutaneous angioplasty.

Total in-hospital mortality in this group was 25% (n = 5). None of the deaths were related to the decannulation procedure (Table 3). Patients in the surgical group required significantly more packed red blood cell transfusions versus those with totally percutaneous decannulation (mean, 2.2 ± 1.2 versus 1.6 ± 1.4; P<.05). Mean hospital stay of the patients treated with surgical cutdown was longer (36 ± 12 days versus 32 ± 18 days; P=NS) (Table 2).

Follow-up. Mean patient follow-up was 8 ± 5 months (range, 1-19 months) in the percutaneous group and 32 ± 16 months (range, 2-52 months) in the surgical group. In the percutaneous group, 2 patients died during follow-up (7 days and 1 month after discharge) due to multi-organ failure. Follow-up data were available for 10/10 (100%), 8/8 (100%), and 3/8 (38%) eligible patients at 1, 6, and 12 months, respectively.

In the surgical group, 2 patients died during follow-up (2 months and 3 months after discharge). One patient died due to multi-organ failure and the other died due to sepsis. Follow-up data were available for 15/15 (100%), 13/13 (100%), 13/13 (100%), 12/13 (92%), and 7/13 (54%) eligible patients at 1, 6, 12, 24, and 36 months, respectively.

None of the discharged patients in the percutaneous group had late vascular complications, or complained about paresthesia or claudication if physical activity was possible. One patient in the surgical group complained about sustained paresthesia of the left thigh (ECMO cannulation side) 22 months after discharge. The remaining patients were free of symptoms.


In the present study, we describe our clinical experience of a complete percutaneous decannulation procedure after multiple days of VA-ECMO. As described above, we used two conventional vascular closure devices for closure of the femoral access site. Figure 1 shows our percutaneous closure technique. This cohort was compared with a group of patients who underwent conventional surgical cutdown. The attraction in percutaneous decannulation lies in avoidance of open femoral access to decrease local postoperative complications and to facilitate earlier discharge.

Severe cannulation site complications remain a major safety issue during VA-ECMO and after decannulation, respectively. Additionally, patients with cardiogenic shock are at an increased risk for developing vascular complications.4

In a retrospective analysis, late vascular complications occurred in 12 hospital survivors (12%) after VA-ECMO support. In all patients, the arterial and venous cannula was placed after surgical cutdown through the femoral vessels. Decannulation was also performed by using a surgical approach. The mean follow-up was 30 ± 10 months. All patients had suffered from a stenosis at the former arteriotomy site and were treated by surgery (n = 7) or percutaneous angioplasty (n = 5). Predictors of late vascular complications were technical problems during ECMO explantation and history of peripheral vascular disease.3 

Decannulation is usually performed under direct visualization using the traditional surgical cutdown approach. However, this technique has several potential disadvantages, such as the need for general anesthesia, blood loss, femoral nerve neuropathy, wound-healing complications, and delayed mobility. 

The use of conventional suture-based vascular closure devices for large access-site vessel closure in patients who have undergone structural interventional procedures requiring large-sheath technology (transcatheter aortic valve implantation, abdominal aortic aneurysm repair, or percutaneous mitral valve repair) is a well-accepted method.6-8 

 We observed procedural success in all patients. Although immediate hemostasis was achieved more frequently in the surgical group, complete hemostasis was also achieved in the percutaneous group after treatment with a compression bandage for some hours. 

Severe local wound complications requiring surgical treatment occurred only in the surgical group. Patients who were treated by surgical cutdown required significantly more red blood cell transfusions versus those with percutaneous closure. Hospital stay was longer (not significant) in the surgical group. During follow-up, none of the patients treated by percutaneous closure complained about potential late access-site complications. In the surgical group, 1 patient complained about sustained paresthesia of the left thigh as a potential access-site complication.

The two most important factors in the successful use of vascular closure devices are the quality of the vascular puncture and the experience of the operator. In this series, experienced interventional cardiologists performed all cannulation and decannulation procedures and a percutaneous approach was used to achieve initial vascular access in all patients. This might have contributed to the high procedural success rate; a transfer to this closure technique for patients in which initial cannulation was performed by traditional surgical cutdown is not possible.

Hwang et al10 recently described their experience with a totally percutaneous approach for access-site closure after multiple days on VA-ECMO. They compared procedural outcomes and complications between 56 patients receiving a totally percutaneous technique for access-site closure and 59 patients undergoing traditional surgical cutdown. Although the need for open repair at the insertion site, limb ischemia, and infection of the insertion site occurred more often in the surgical cutdown group, the overall complication rate was statistically not significant between the two groups. Also, the length of intensive care unit stay after ECMO weaning and length of hospital stay after weaning were not significantly different. In comparison to our study, a different percutaneous closure technique was used. Instead, they performed a direct puncture in the proximal arterial cannula to achieve access to the vessel and used two suture-mediated closure devices (Perclose ProGlide) in all patients. Data on follow-up are lacking in this series.

Study limitations. First, this study is a non-randomized, retrospective study of a single-center experience. Second, we had no systematic vascular imaging of the groin such as ultrasound or angiography after access-site closure. Therefore, potential vascular complications without presenting symptoms may have gone unrecognized. Third, interventional cardiologists skilled in deploying the vascular closure devices performed all procedures, which may have contributed to the favorable results.


We conclude that percutaneous closure of large access sites after VA-ECMO in adults using conventional closure devices is feasible, effective, and safe when performed by operators with experience in obtaining vascular access with very large sheaths and trained in the use of vascular closure devices. However, prospective studies are needed to prove the concept of percutaneous closure in a larger patient group.


1.    Gaffney AM, Wildhirt SM, Griffin MJ, Annich GM, Radomski MW. Extracorporeal life support. BMJ. 2010;341:c5317.

2.    Bisdas T, Beutel G, Warnecke G. Vascular complications in patients undergoing femoral cannulation for extracorporeal membrane oxygenation support. Ann Thorac Surg. 2011;92:626-631.

3.    Zimpfer D, Heinisch B, Czerny M, et al. Late vascular complications after extracorporeal membrane oxygenation support. Ann Thorac Surg. 2006;81:892-895.

4.    Aziz F, Brehm CE, El-Banyosy A, Han DC, Atnip RG, Reed AB. Arterial complications in patients undergoing extracorporal membrane oxygenation via femoral cannulation. Ann Vasc Surg. 2014;28:178-183.

5.    Byrne RA, Cassese S, Linhardt M, Kastrati A. Vascular access and closure in coronary angiography and percutaneous intervention. Nature Rev Cardiol. 2013;10:27-40.

6.    Jaffan AA, Prince EA, Hampson CO, Murphy TP. The Preclose technique in percutaneous endovascular aortic repair: a systematic literature review and meta-analysis. Cardiovasc Interv Radiol. 2013;36:567–577.

7.    Griese DP, Reents W, Diegeler A, Kerber S, Babin-Ebell J. Simple, effective and safe vascular access site closure with the double-ProGlide preclose technique in 162 patients receiving transfemoral transcatheter aortic valve implantation. Catheter Cardiovasc Interv. 2013;82:E734-E741.

8.     Rüter K, Puls M, von der Ehe K, et al. Preclosure of femoral vein access site with the suture-mediated ProGlide device during MitraClip implantation. J Invasive Cardiol. 2013;25:508-510.

9.    Upponi SS, Ganeshan AG, Warakaulle DR, Phillips-Hughes J, Boardman P, Uberoi R. AngioSeal versus manual compression for haemostasis following peripheral vascular diagnostic and interventional procedures-a randomized controlled trial. Eur J Radiol. 2007;61:332-334.

10.    Hwang JW, Yang JH, Sung K, et al. Percutaneous removal using Perclose ProGlide closure devices versus surgical removal for weaning after percutaneous cannulation for venoarterial extracorporeal membrane oxygenation. J Vasc Surg. 2016;63:998-1003.e1. Epub 2016 Jan 26.

From the University of Leipzig, Heart Center, Department of Internal Medicine/Cardiology, Leipzig, Germany. 

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted May 24, 2016, provisional acceptance given June 21, 2016, final version accepted July 28, 2016.

Address for correspondence: Nicolas Majunke, MD, University of Leipzig, Heart Center, Department of Internal Medicine/Cardiology, Strümpellstrasse 39, 04289 Leipzig, Germany. Email: nicolas.majunke@medizin.uni-leipzig.de