The use of venovenous gas exchange as a supportive therapy in severe respiratory failure has increasingly been used during the past decade. Favorable outcome data, most probably because of improved management, and technical advancement have broadened the indications for respiratory extracorporeal membrane oxygenation (ECMO) from its classical role as an ultimate rescue therapy in the most severe cases of lung failure to a valid supportive therapy to maintain adequate gas exchange in severe acute respiratory distress syndrome (ARDS).1 The commonly used configurations of venovenous ECMO are single-site cannulation of the right internal jugular vein by double-lumen cannulae or two-site cannulations typically involving the right internal jugular vein as well as the femoral veins in different configurations. Critically ill patients may, however, bear indwelling venous catheters for continuous drug infusion including vasopressors, analgosedation, or anticoagulation as well as large-bore catheters for renal replacement therapy. Vascular malformation or presence of thrombosis may additionally limit the possibilities of central venous access.
Yoffa2 described the supraclavicular subclavian venipuncture and catheterization for insertion of infusion lines as early as 1965. The supraclavicular approach never gained broad acceptance, although a low complication rate and a rather simple insertion technique was emphasized by the authors. Several reports have since affirmed the advantages of the supraclavicular approach.3–6 In our institution, the slightly modified originally described technique from both sides of the neck is widely used as an alternative to the puncture of the internal jugular vein. If a patient needs ECMO but the right internal jugular vein or the femoral veins are not easily accessible, the supraclavicular approach to the subclavian vein may serve as an alternative. We retrospectively analyzed the data of patients undergoing venovenous ECMO with either a two-site or a single-site cannula configuration comparing the “classical” jugular cannulation and the supraclavicular approach with respect to successful cannulation as well as performance of the extracorporeal circuit.
The study protocol was approved by the institutional review board of the Medical University of Vienna. Over a period of 36 months, all patients requiring venovenous ECMO for acute lung failure or terminal chronic lung failure during bridging to lung transplantation were identified and their charts and the separate ECMO documentation were analyzed. Patients were eligible for analysis if undergoing full venovenous ECMO in a femoro-jugular or femoro-subclavian setting or if undergoing mid-flow ECMO with the main goal of decarboxylation caused by severe respiratory acidosis using double-lumen cannulae via the right internal jugular or a subclavian vein, respectively. Patients were either equipped with drainage and reperfusion cannulae by Maquet (Rastatt, Germany) with the cannula size chosen according to venous diameter measured by ultrasound before insertion as well as expected need of blood flow or with 22 or 24 French double-lumen cannulae (Novaport Twin, Novalung, Heilbronn, Germany). Charts were analyzed for demographics, indication for ECMO, cannula-associated complications, performance of the circuit, as well as duration of cannulae in place. All consecutive patients fulfilling these criteria were included in the analysis.
Location and diameter of the respective vein was determined by ultrasound before venipuncture. In all cases, puncture and insertion of cannulae was done after local anesthesia using an introducer set enabling stepwise dilatation over a guidewire (Percutaneous Insertion Kit, Maquet, Rastatt, Germany). The cannulae were then inserted over their inherent introducers to the desired depth. A supraclavicular approach was chosen if the right internal jugular vein was inaccessible and ultrasound revealed an accessible and open subclavian vein. Reasons for choosing a supraclavicular approach were a right jugular catheter in place in eight patients, partial thrombosis of the right jugular vein in two patients, and local infection involving the right lateral neck in one patient. The technique used for the supraclavicular approach was slightly modified compared with the original description by Yoffa,2 with the puncture point at the lateral margin of the clavicular head of the sternocleidomastoid muscle at its insertion point to the clavicle (Figure 1). The needle was directed to bisect the angle of the muscle and the clavicle and was directed 10° to 20° anterior to the coronal plane to avoid arterial and pleural puncture.4 Patients in need for full ECMO support received a two-cannula femoro-jugular or femoro-subclavian (supraclavicular) configuration. Cannulae were then connected to an ECMO system (Cardiohelp, Maquet, Rastatt, Germany), and blood as well as sweep gas flow was titrated to desired levels. In patients primarily in need for decarboxylation, 22 or 24 French double-lumen cannulae (Novaport Twin, Novalung, Heilbronn, Germany) were inserted either via the right internal jugular vein or via the right or the left subclavian vein by the supraclavicular approach (Figure 2). The cannulae were then connected to a previously described pump-driven gas exchange system (ILA Activve, Novalung, Heilbronn, Germany) equipped with a membrane (ILA membrane ventilator, Novalung, Heilbronn, Germany) optimized for low to mid blood flow.7,8 Blood flow in all cases was chosen to keep suction pressure well below −100 mm Hg. Removal of all cannulae in patients weaned from extracorporeal life support (ECLS) was performed percutaneously after stopping heparin for at least 1 hour. Jugular and supraclavicular cannulae were then removed, and the insertion site was compressed until bleeding ceased. An inflatable transparent compression dressing (Safeguard©, Maquet, Rastatt, Germany) was then placed over the insertion site for at least 6 hours being shortly deflated every 2 hours to allow capillary refill (Figure 3).
The primary end-point of the study was the extracorporeal blood flow 12 hours after ECLS start. For the primary end-point, as well as for the secondary end-points “P/F improvement” and “pump speed (RPM),” two-sided 95% confidence intervals (CIs) for the mean group difference were calculated based on a Welch test. Because of the small sample, interpretation of group differences was regarded as descriptive with respect to clinical relevancy. Therefore, no p values are displayed in Table 1. Quantitative data are reported as mean ± standard deviation and qualitative data as absolute frequency. Statistical analyses were carried out with the statistics program SAS 9.4 (SAS Institute Inc.).
Altogether, 35 patients were analyzed. Fifteen patients underwent a two-cannula setting, 10 of them with the reperfusion cannula placed into the right internal jugular vein and in five of them placed into the left (n = 4) or right (n = 1) subclavian vein via the supraclavicular approach. Twenty patients underwent insertion of double-lumen cannulae, 14 of them via the right internal jugular vein and 6 of them via the right (n = 3) or the left (n = 3) subclavian vein via the supraclavicular approach. Four patients were cannulated while awake and breathing spontaneously, all of them receiving double-lumen cannulae, three via a jugular and one via a supraclavicular access. Overall, no insertion-associated complications occurred. In one patient undergoing right-sided supraclavicular cannula insertion, minor continuing bleeding at the insertion side occurred a few hours after cannulation, making an additional purse-string suture necessary. No bleeding events were observed after removal of all cannulas. Demographics, performance of the extracorporeal gas exchange circuits, and cannula-associated parameters are shown in Table 1. Analysis of the primary end-point revealed no clinically relevant difference between the supraclavicular versus jugular setting with respect to extracorporeal blood flow 12 hours after ECMO start (mean group difference [95% CI]: 148 [−621 to 917] ml/min). For the secondary end-point pump speed, no clinically relevant group difference was observed either (mean group difference [95% CI]: 115 [−1176 to 1406] RPM). Whereas, for P/F improvement, a markedly greater increase of P/F ratio was found in the supraclavicular setting compared with the jugular setting (mean group difference [95% CI]: 55 [−15 to 124]. Duration of a cannula in place ranged between 4 and 43 days in the jugular group and between 4 and 32 days in the supraclavicular group. No cannula-associated event influenced the duration of their use, and no cannula had to be removed prematurely.
The classical insertion sites for venous ECMO cannulae—left and right femoral vein as well as the right internal jugular vein—sum up to a choice of three. Critically ill patients are usually equipped with venous catheters at these very places or these access sites may not be available because of anatomic deviations, local infections, surgical procedures, hematomas, or presence of intravascular thrombosis, for instance. In cases where ECMO is needed, this may lead to troubles identifying an appropriate vascular access site. Alternative cannulation sites have been proposed and successfully used like the left internal jugular vein9 or the left subclavian vein via the infraclavicular access.10 The left jugular vein might bear a higher risk for malposition, arterial puncture,11 and vessel perforation when inserting a large-bore cannula because of more angulations through the venous system on its way to the superior vena cava or the right atrium.9 A left-sided infraclavicular approach to the subclavian vein might be associated with a higher risk for pneumothorax4 as well as difficulties in overcoming the resistance of the narrow between clavicle and rib. Malposition, pneumothorax as well as inadvertent arterial puncture rate, was lower in a large number of patients undergoing central venous catheter insertion via the supraclavicular approach compared with the complication rate associated with internal jugular and infraclavicular subclavian access as summarized in a review by Kusminsky.11 Air embolism especially in spontaneously breathing patients may be a possible severe, if rare, complication12 making specific precautions like Trendelenburg’s position and a structured insertion and removal procedure advisable. Literature, however, does not link the supraclavicular approach to a higher risk for air embolism.3–6
In this retrospective analysis, ECMO cannulae were placed via a right or left supraclavicular approach in 11 patients. These sites were chosen because of the lack of accessibility of the right internal jugular vein. No difference in complication rate and ECMO performance was observed between the jugular and supraclavicular groups, respectively. Blood flow was similar between ECMO patients in whom reperfusion cannulae were placed via the jugular or the supraclavicular access. The markedly larger increase in oxygenation in patients undergoing supraclavicular cannulation can be explained by two of the three patients equipped with larger double-lumen cannulae. Twenty-four French double-lumen cannulae allow for higher extracorporeal blood flow, thus enhancing oxygenation. Accordingly, blood flow in the subgroup of patients equipped with supraclavicular double-lumen cannulae was higher (1700 ± 645 vs. 1379 ± 351 ml/min). The larger cannulae were chosen to achieve an enhanced oxygenation effect additionally to the primary goal of CO2 elimination, as well as because of their length of 27 cm compared with 17 cm in the 22 F model. When placing a left supraclavicular cannula, 17 cm may be too short, especially in larger patients. Adequate cannula length is necessary to avoid the catheter resting in the proximal superior vena cava and also to avoid the tip of the catheter pointing toward the lateral vessel wall and causing perforation (Figure 2).13 In patients with two-cannula setting, drainage cannulae diameters and reperfusion cannulae diameters were comparable between the groups, as was the average blood flow (Table 1). In contrast, improvement in PaCO2 seemed to be larger in the jugular group. This fact can be explained by the lower baseline CO2 levels in the supraclavicular group. As the goal was to normalize pH, not PaCO2, a lower elimination rate resulted within the supraclavicular group.
Another major difference between jugular and supraclavicular groups was observed with respect to duration of the respective cannula in place. This time approximated duration of extracorporeal gas exchange therapy as no cannula was removed prematurely because of mechanical problems.
The supraclavicular approach has been shown to be a safe and simple access site for central venous catheters. Large bore catheters up to a diameter of 16 French used for renal replacement therapy were placed successfully in a series of 208 catheterization episodes with a remarkably low rate of complications.3 A review of available studies on the supraclavicular access to insert central venous catheters demonstrated a markedly lower complication rate compared with a comprehensive review describing complication of central venous catheter placement via the internal jugular or infraclavicular subclavian approach.4,11 Rates of puncture failure, malposition, pneumothorax, and arterial puncture seemed all lower when using the supraclavicular technique. There are no reports on complications associated with percutaneous removal of the catheters. In our series of patients, the insertion site seemed to be well compressible and the strategy followed for removal as previously mentioned was successful to avoid bleeding.
To our knowledge, supraclavicular placement of large-size intravascular cannulae as used for ECMO has not yet been described in the literature. Percutaneous venous cannulation for ECMO seems to be feasible and safe, if precautions like preinsertion ultrasound are used, yet can lead to fatal consequences if complications occur.14–16 The observations made in our small series of patients point toward the supraclavicular approach to be a safe and effective alternative to internal jugular access. They have to be viewed, however, in the light of several limitations: First of all, the inherent limitations of a retrospective analysis apply, although in our institution, data on complications associated with ECMO and its performance are documented prospectively in a local registry. All consecutive patients fulfilling the selection criteria within the defined time period were included in the analysis. Furthermore, the number of patients undergoing supraclavicular cannulation is low because of the fact that jugular approach is still the first choice for cannulation and the supraclavicular access is used only if jugular cannulation seems impossible or inadvisable. Thus, a certain selection bias cannot be ruled out. It has to be taken into account, moreover, that supraclavicular insertion of central venous catheters has been a standard procedure in our institution for many years, leading to extensive experience with respect to the procedure. Using the technique as a completely novel one for insertion of large-size ECMO cannulae might therefore not always reflect our own positive experience. We suggest approaching the technique by placing central venous infusion catheters at first to accustom to its benefits and pitfalls.
The supraclavicular access technique to the left and right subclavian veins could be a useful alternative for placing ECMO cannulae with a low complication rate, good tolerance, and performance comparable to standard right jugular cannulation if the right internal jugular vein is inaccessible. Further prospective studies are needed to prove the validity of this approach.
1. Fan E, Gattinoni L, Combes A, et al. Venovenous extracorporeal membrane oxygenation for acute respiratory failure. Intensive Care Med 201642: 712–724–XXX, .
2. Yoffa D. Supraclavicular
subclavian venepuncture and catheterisation. Lancet 1965.2: 614–617.
3. Muhm M, Sunder-Plassmann G, Apsner R, et al. Supraclavicular
approach to the subclavian/innominate vein for large-bore central venous catheters. Am J Kidney Dis 1997.30: 802–808.
4. Cunningham SC, Gallmeier E. Supraclavicular
approach for central venous catheterization: “Safer, simpler, speedier.” J Am Coll Surg 2007205: 514–516–XXX, .
5. Czarnik T, Gawda R, Perkowski T, et al. Supraclavicular
approach is an easy and safe method of subclavian vein catheterization even in mechanically ventilated patients. Anesthesiology 2009.111: 334–339.
6. Kocum A, Sener M, Caliskan E, et al. An alternative central venous route for cardiac surgery: Supraclavicular
subclavian vein catheterization. J Cardiothorac Vasc Anesth 2011.25: 1018–1023.
7. Hermann A, Staudinger T, Bojic A, et al. First experience with a new miniaturized pump-driven venovenous extracorporeal CO2
removal system (iLA Activve): A retrospective data analysis. ASAIO J 2014.60: 342–347.
8. Hermann A, Riss K, Schellongowski P, et al. A novel pump-driven veno-venous gas exchange system during extracorporeal CO2
-removal. Intensive Care Med 2015.41: 1773–1780.
9. Abrams D, Brodie D, Javidfar J, et al. Insertion of bicaval dual-lumen cannula via the left internal jugular vein for extracorporeal membrane oxygenation. ASAIO J 2012.58: 636–637.
10. Shafii AE, McCurry KR. Subclavian insertion of the bicaval dual lumen cannula for venovenous extracorporeal membrane oxygenation. Ann Thorac Surg 2012.94: 663–665.
11. Kusminsky RE. Complications of central venous catheterization. J Am Coll Surg 2007.204: 681–696.
12. Heckmann JG, Lang CJ, Kindler K, et al. Neurologic manifestations of cerebral air embolism as a complication of central venous catheterization. Crit Care Med 2000.28: 1621–1625.
13. Gravenstein N, Blackshear RH. In vitro
evaluation of relative perforating potential of central venous catheters: comparison of materials, selected models, number of lumens, and angles of incidence to simulated membrane. J Clin Monit 1991.7: 1–6.
14. Conrad SA, Grier LR, Scott LK, et al. Percutaneous cannulation for extracorporeal membrane oxygenation by intensivists: A retrospective single-institution case series. Crit Care Med 201543:1010–1015.
15. Rubino A, Vuylsteke A, Jenkins DP, et al. Direct complications of the Avalon bicaval dual-lumen cannula in respiratory extracorporeal membrane oxygenation (ECMO
): Single-center experience. Int J Artif Organs 2014.37: 741–747.
16. Hirose H, Yamane K, Marhefka G, et al. Right ventricular rupture and tamponade caused by malposition of the Avalon cannula for venovenous extracorporeal membrane oxygenation. J Cardiothorac Surg 2012.7: 36.