Usefulness of Self-Expanding Drainage Cannula in Venovenous Extracorporeal Membrane Oxygenation: Tips, Tricks, and Results of an Early Experience

Abstract

Inadequate venous drainage decreases the efficiency of extracorporeal membrane oxygenation (ECMO). Pump augmentation may even make it worse due to collapse of the venous system under negative pressures. Furthermore, recirculation is a phenomenon that occurs when oxygenated blood supplied through the infusion cannula is withdrawn directly through the drainage cannula without contributing to the oxygenation of the patient and also compromises the efficacy of the therapy. Large drainage cannulas allow for similar flow rates at lower pump speed. But percutaneous insertion of these larger cannulas could be challenging. When using a self-expandable cannula, the diameter of the cannula for the insertion can be reduced, and once inserted, its intravascular diameter maximized, resulting in a large venous cannula due to in situ expansion after mandrel removal (up to 36F). We present a retrospective series of selfexpanding venous cannula 430 or 530 mm in length in six consecutive patients undergoing venovenous (VV) ECMO. No vascular or cardiac iatrogenic injury was caused during implantation. Target flows were reached, and no clinically significant recirculation was described in any case. The use of selfexpanding drainage cannulas was safe, and efficient drainage was achieved with easy and definitive unique positioning during cannulation.

The main objective of ECMO is to provide systemic perfusion or gas exchange during an injury of the heart and/or the lungs as bridge to recovery, transplant, destination, or decision.¹

Patients with isolated lung dysfunction can be supported with VV ECMO, usually provided by using two separate cannulation sites: a drainage cannula to the circuit, and a return cannula to the patient.

Inadequate drainage into the circuit may occur, leading to end-organ malperfusion or blood trauma. This occurs more often in the case of cannula malposition or small diameter, and methods to improve venous drainage are of high importance.

In addition, increased blood flow to overcome suboptimal drainage in VV ECMO may increase recirculation, making the therapy less efficient. It is more likely to occur with femoro-jugular configuration, when both cannulas are closer. It can be reduced by adding a second drainage cannula,² or by replacing the drainage cannula with one with a larger diameter and longer drain surface.

self-expanding venous cannulas minimize the diameter of the cannula for insertion while maximizing the intravascular diameter once the cannula is inserted and allow drainage all along the intravascular path. The uncovered part acts as a spacer that prevents the venous system from collapsing, while the drainage point at the end of the covered part is located considerably away from the atrium (Figure 1). No reports were found on the use of these cannulas in ECMO therapy.

Figure 1.: A: Smartcanula design: Tubing length (T) 100 mm, covered length (C) 80 mm, and uncovered length (U) 160, 250, or 350 mm; once the length of the uncovered segment is variable depending on the total size of the cannula. B: The open-wall design allows drainage not only through the tip but also all along the intravascular path, making recirculation highly unlikely.

Von Segesser and coworkers redesigned the interface between the venous system and the extracorporeal circulation by introducing the concept of “collapsed insertion and expansion in situ” to obtain more efficient venous drainage.³ This provides two important advantages in cardiopulmonary bypass. First, this cannula expands to 36F once it reaches the target vessel (expanding within the cava vein to 36F), while the diameter of the collapsed cannula during insertion through the access vessel is 18F (Figure 2), which is relevant for percutaneous remote access.^4,5 The other important advantage is that the open-wall design allows to act as spacers, avoiding the collapse of the cava.⁶ We have previously described this approach as single cannulation for tricuspid valve reoperations.⁷

Figure 2.: A: Collapsed cannula (18F) when stretched the mandrel, to facilitate insertion (B) expanded cannula (36F) after removal of the mandrel, once fully inserted into the venous system. The wire must be removed before the mandrel to prevent dislodgment.

METHODS

From June 2015 to March 2017, we used the 340, 430, or 530 mm length self-expanding venous cannula for drainage into the circuit, in six VV ECMO (Smartcanula LLC, Lausanne, Switzerland). The only inclusion criterion was operator preference. The indications for therapy were acute respiratory distress syndrome during the H1N1 epidemic in three cases, chickenpox pneumonia, bacterial pneumonia, and chest trauma.

Biometric characteristics, cannula sizes, and flow achievements were recorded. We review any possible malfunctions (loss of effectiveness, flow limitation, visceral complications and malperfusion, or edema of the extremities). Biochemical patterns of hemolysis were also verified.

SURGICAL TECHNIQUE

The VV ECMO configuration places the artificial lung in series with the patient’s lungs. In all reported cases, a double venous cannula system has been used, with a femoral-jugular configuration: the self-expanding drainage cannula placed in the femoral vein and the return cannula to the patient placed through the right internal jugular vein, and advancing toward the right atrium in front of the tricuspid valve.

Femoral drainage cannulation was performed with the collapsed 18 F cannula reaching 36F intravascularly when the mandrel was removed for venovenous femoral-jugular configuration. For the three first cases, an open cut-down semi-Seldinger technique was used. Recent cases were performed percutaneously with long-kink resistant wires (>2 m). A short return cannula (<20 cm) was introduced via the right jugular vein.⁸ All implants were performed using bicaval transesophageal echo view to ensure that the wire is directed toward the superior vena cava. Heparin (5,000 IU) is administrated. Serial dilatation of the vein is performed taking care that the wire moves freely within the dilator to prevent it from bending. Dilation up to 20 F is easy and hemostatic, once the cannula expansion after mandrel removal is above that size. After initiation of therapy, a heparin infusion was used to achieve an activated clotting time of 160–180 seconds.

RESULTS

There were no major vascular or cardiac complications with this technique. Neither was recirculation detected during echo follow-up in any of the cases (Table 1). The open-wall design allows drainage at all levels of affluence and makes this complication unlikely. Flow achievements met clinical demands in all cases. In one patient only, the venous cannula was exchanged for a traditional polyurethane cannula after nine days of treatment due to reduced flow in the context of an increase in the intra-abdominal pressure (22 mm Hg). The improvement after the exchange was parallel to normalization of the intra-abdominal pressure. Another case presented venous congestion and elevation of the limb was sufficient to treat this condition. No significant hemolysis was detected. When the therapy ended, the cannulas were carefully evaluated and only small clots were found attached to the open-wall cannula (Figure 3).

Table 1. - Therapy Specifications: Cannulas Sizes (Expanded Cannula Diameter/Full Cannula Length), Calculated Cardiac Output (2.4 L/min/m²), Average and Maximal Flows Achieved, and Duration of the Therapy.

Drainage Cannula Size (Fr/mm)	Return Cannula Size (Fr/mm)	Calculated Cardiac Output (L/min)	Average Flow (L/min)	Maximal Flow (L/min)	Mean P1 (mm Hg)	Minimum P1 (mm Hg)	Duration (days)
36/530	21/180	4.3	3.6	3.8	-32	-50	13
36/530	19/150	4.5	4.0	5.2	-53	-68	11
36/530	17/150	4.5	4.9	5.0	-60	-65	25
36/430	19/150	5.3	3.5	3.8	-30	-58	8
36/430	17/150	4.7	4.0	4.9	-70	-92	52
36/530	21/180	4.7	4.0	4.6	-30	-50	17

Figure 3.: Removed selfexpanding venous cannula after ECMO therapy. Small clots added to the mesh.

LIMITATIONS AND ETHICS

This series shows our initial experience and ethical decision to use the self-expanding cannula were based on the principles beneficence and no maleficence, supported by our surgical background and cannula availability.

The small sample size and limited duration of the therapy are the main flaws of this series.

DISCUSSION

Despite significant improvements, drainage of systemic venous return remains the main limiting factor for extracorporeal life support. The amount of systemic venous return is directly correlated with the amount of pump flow,⁹ and according to Poiseuille’s law, halving the radius of a tube reduces a laminar flow by a factor of 16. Although less predictable because of the turbulence, maximizing intravascular diameter of the drainage cannula should improve drainage into the circuit, but diameter is generally limited by the access vessel. The type of venous cannula is a determining factor in venous return, and cannulas can be optimized using a thin-wall, maximizing internal diameter and adding lateral drainage holes to prevent collapse of venous wall with negative pressures.

In traditional thin-walled cannulas, as the size of the venous cannula approaches the size of the vein, there is a decrease in the maximum flow rate,¹⁰ due to impaired collateral drainage, particularly of the hepatic veins when the tip of the cannula is positioned beyond them.

The Smartcanula (Smartcanula LLC, Lausanne, Switzerland) is a wall-less cannula consisting of a mesh that acts like a stent and behaves like an intravascular spacer. Collapsed insertion (18F) and expansion in situ (36F) allow to overcome the diameter limitations by the access vein, and together with a reduced wall thickness of the mesh (0.2 mm) favors an internal/external diameter ratio and provides significant superior drainage. Furthermore, the open-wall design allows venous drainage in all venous openings throughout the intravascular route, except for the 8 cm proximal covered part that remains at the femoral level (Figure 1A).¹¹ At the same time, this device also behaves as an stent and prevents the vein from collapsing and allows all collateral flow to drain, further improving drainage.¹² These advantages lead to superior performance of these cannulas compared to traditional ones,⁴ which is why we have widely used self-expanding cannulas in open heart surgery as single venous cannulation, particularly in redo cases.⁷

In addition to venous drainage limitation, recirculation is the other pitfall related to venous cannulation, and exclusively affects to VV ECMO. This phenomenon occurs when re-infused oxygenated blood is withdrawn through the drainage cannula without passing through the systemic circulation, thereby decreasing the efficiency with which ECMO provides oxygenation. Several factors favor recirculation: proximity of infusion and drainage cannulas, femoral-jugular configuration or high pump speed. The larger the drainage cannulas, the lower the pump speed required for a comparable blood flow,¹³ potentially mitigating blood trauma and hemolysis.¹⁴ Furthermore, the functional tip of the self-expanding cannulas is at the end of the covered part of the intravascular segment (at 8 cm). Pump flow allows open-mesh venous drainage from the stented part to flow down to this point. Since this drainage point is very far from the return cannula, recirculation is very unlikely with the only possible position of the cannula (Figure 1B). As a consequence, mobilization of the cannula, although not possible, it is not necessary. The uncovered part, as mentioned above, acts as a stent, preventing collapse of the venous system.

The growth of endothelium on the mesh could be a concern for long-term therapies. Electron microscopy performed on venous stents shows a limited neointima covering the surface of the stent wires, which increases with time, especially from the fourth week after implantation, but intimal hyperplasia between the wires was not described, and blood backflow is not blocked.¹⁵ We did not notice any underperformance of the cannula over time during the therapy in our series related to this issue.

The removal of the cannula is easy and safe, as shown in the steps represented in Figure 4; and some small clots were attached to the mesh, as shown in Figure 3.

Figure 4.: Decannulation maneuver: (A) removing the cannula with one hand while compressing the vascular access with the other, exerting traction and allowing the diameter to be reduced, to extract it in a collapsed state. The difference between the expanded (B) and collapsed states (C) is made by pulling and stretching the cannula. In percutaneous technique, a deep figure-of-eight suture is sufficient to control bleeding after decannulation.

CONCLUSIONS

Femoral venous cannulation with self-expanding venous cannula is safe, easy, and hemostatic. Seldinger’s peripheral percutaneous cannulation is also feasible. No repositioning was necessary since no recirculation was detected, and target flows were achieved using this approach. Sustained negative pressure or intimal hyperplasia does not appear to affect the performance of the cannula during the therapy.

The cannula under discussion is ISO 9001 certified, and all available types of smartcanula are CE 0344 labeled, but have not yet been approved by FDA for ECMO use.