INTRODUCTION
The benefit of ECMO in septic shock is still under debate [1,2] due to the difficulties in meeting the required flow demands in septic shock and fluctuating outcomes, together with scanty information available in adults [3]. In addition, the indications for ECMO usage in septic shock are still unclear. On the other hand, we experience an increased application of ECMO, due to an expansion of indication and inclusion of higher risk patients [4].
Due to distributive circulatory failure with capillary leakage in septic shock [5], uncommonly high flows might be required. Thus, the technical limitations of the cannulae and the oxygenator represent a burden for appropriate flow delivery. The venous cannula(e) can be limiting due to the generation of highly negative pressures, depending on their diameter and the flow needed. These negative pressures might provoke hemolysis. In addition, high blood flows through the oxygenator reduce its gas exchange performance (GEP) [6]. GEP also decreases over time because of coagulation inside of the oxygenator [7].
Facing these problems, two parallel-running ECMO circuits were used in a complex patient, one in vv and one in va configuration. Reported cases of the use of two separate, simultaneous extracorporeal circuits are scarce [8,9,10]. This option might be a feasible measure to get around technical limitations of cannulation and oxygenator performance.
CASE HISTORY
A 29-year-old male patient (179 cm, 64 kg, BSA 1.81, BMI 20) with cystic fibrosis was admitted to the ICU because of exacerbated respiratory insufficiency secondary to hospital-acquired pneumonia. He required invasive respiratory support; Horowitz Index was 167. A vv ECMO implantation was decided, aiming at extubation. The patient was then waitlisted for lung transplantation (LTPL). A 27F double-lumen cannula was inserted into the right jugular vein under x-ray control and ECMO support was delivered with 2.8 l/min blood flow, 4 l/min gas flow and 80 % FiO2. The patient was extubated and mobilized to stand upright.
The patient status deteriorated over the following days, the main problem being CO2 elimination. ECMO settings were adapted to 3.3 l/min blood flow, >10/min gas flow and 100 % FiO2. On day 6, the ECMO system (Maquet Cardiohelp HLS 7.0, Getinge GmbH, Rastatt, Germany) was exchanged because of decreasing oxygenator performance, and the respiratory parameters could not be held within the physiologic range despite maximum gas flow with a 15 l gas blender. HLS 7.0 system was used for all the configurations in this case. Oxygenator exchange was effective and gas flow could be reduced to 8 l/min – but only short-time. An additional 23 F venous cannula was percutaneously inserted in the right femoral vein and advanced to the right atrium under TTE control on day seven to deliver higher flowrates. A few hours later the patient’s hemodynamics deteriorated, requiring vasopressor support, and
a 15 F arterial cannula was inserted in the right subclavian artery for circulatory support. CO2 removal and oxygenation were improved by the first upgrade and only oxygenation deteriorated again with the second upgrade while CO2 levels remained. The patient was reintubated. The ECMO setup now complied with a vv va configuration. The system was driven by a single console; thus, the flow distribution over the different limbs had to be measured with additional flow probes (Emtec Sono TT, emtec GmbH, Finning, Germany) and controlled by Hoffmann tubing clamps.
Septic shock manifested on days seven to nine with decreasing MAP (norepinephrine >50 mcg/min), tachycardia (>130 bpm) and fever >39 °C. Invasive respiratory support was applied pressure controlled, 100 % FiO2. Esmolol administration (75 mcg/kg/ min) was beneficial for SaO2.
Overall, ECMO blood flow was now 6 lpm, divided 50/50 over arterial and venous access for the next 37 days. Norepinephrine demand decreased to values around 8 mcg/min. System exchange was necessary 11 times because of decreasing GEP, every 3.36 days on average. During every exchange, cardiopulmonary parameters decreased to critical levels. We did not observe recirculation phenomenon by sampling pre-oxygenator blood gases. Internal discussions led to the implementation of two parallel simultaneous ECMO circuits for support aiming at stabilizing PaO2 levels and exchange procedures.
Fig. 1: Charting overview on stepwise ECMO upgrades and cannulation in prospective order. Venous lines are colored blue, arterial lines red. Arrows point in flow direction.
For the stepwise exchange procedure (fig. 1), the AV limbs were clamped and the ECMO ran on vv configuration. A new system was connected to the AV limbs, and while this was running, the vv ECMO was exchanged by a new system. Both systems were set to a blood flow of around 3 l/min. The SvO2 immediately increased to normal values and MAP stabilized at these flows. Gas flow was set to 5 l/min with 100 % FiO2 in the vv circuit and 1 l/min with 50 % FiO2 in the va circuit.
The patient remained on parallel ECMO for 30 days without major technical complications. Exchange of both systems at once was performed 4 times because of increased D-dimer levels, with an average of one ex- change per 7.5 days. During exchange procedures, vital parameters were stable (fig. 2).
Anticoagulation was maintained by continuous delivery of unfractionated heparin and anti-Xa monitoring. Target value was 0.3–0.4 IU/ml, which is higher than normal practice (0.2–0.3 IU/ml. For safety, all connection sites were secured by cable straps and every tubing was fixated by adhesive strips on head, trunk or extremities twice. ECMO consoles were placed on opposite sides of the bed to avoid confusion.
Average PaO2 was calculated daily (fig. 3). Until day 45, the patient was on one ECMO circuit, and on two circuits from there on. During parallel ECMO, PaO2 was always above 10 kPa. After 75 days on ECMO, the patient developed acute liver failure and was removed from the LTPL list. Supportive therapy was terminated, and the patient eventually died.
Fig. 2: Chronologic overview on ECMO modes with corresponding minimal and maximal settings, ventilation status and number of exchanges
Fig. 3: Graphical display of average PaO2 values for every day on ECMO calculated from every blood gas analysis taken. Upgrade to parallel circuits on day 45
DISCUSSION
The decision for ECMO implantation was to allow physical therapy and to spare mechanical ventilation, which might be beneficial for survival in bridge to LTPL patients [11]. Patient selection and vascular access were performed analogously to published recommendations from other experienced centers [11, 12]. Unfortunately, the double lumen cannula alone did not allow higher flow when the patient deteriorated. This might have had anatomical and physiological reasons like intravascular hypovolemia, while sizing of the cannula should have been sufficient for the calculated flow of 4.3 l/min (BSA x 2.4). Therefore, an additional venous limb to improve drainage is in conformance with ELSO guidelines [13]. Addition of an arterial inflow-limb for circulatory support could be favorable for outcome in sepsis-induced cardiogenic shock [14], although inclusion criteria and setup in our case differed from Bréchot et al. (no lactacidemia/impaired LVEF). Vav ECMO might be a rescue strategy for patients with combined respiratory and hemodynamic failure [15]. We did this, but with two ECMO consoles and a cannulation strategy, that was undoubtedly born of necessity.
Therapeutic application of two parallel ECMO circuits might be an alternative, especially in a setting described in this case-report: in a combined respiratory and circulatory failure together with high cardiac output state that is resistant to other rescue strategies like beta-blockade [16].
Two parallel simultaneous ECMO circuits with vv and va configuration may help in delivering unrestricted high flows to patients in septic shock and avoid reduced GEP of a single oxygenator. Lithmathe and Dapunt [8] were able to deliver 9 l/ min overall flow with a combination of two va ECMOs. Malik et al. [9] delivered approximately 8–9 l/min with two vv ECMOs to handle hypoxemia in a hyperdynamic state.
An additional advantage of va and vv ECMO is the individual controllability of two different circuits, allowing maximum flexibility in support and adaptability to patient demands.
On the other hand, doubling the foreign surface could have been detrimental [17], although there were no signs to suggest this here, except for elevated D-dimer.
Monitoring D-dimers might help to predict clot formation inside oxygenators and to avoid system-induced coagulation disorders [18], the latter especially in relation to fibrinogen and platelet counts. However, on its own absence of D-dimers can exclude consumptive coagulopathy, but elevated levels alone are not able to show evidence of DIC [19]. In our case, D-dimer values decreased after each exchange of both circuits (fig. 4), as described by Dornia et al. after exchange of one [18].
Higher anticoagulation for ECMO in septic shock is recommended in the literature [3]. Other safety measures do not differ substantially from normal practice.
Riera et al. [3] refer to central cannulation as the last option to extend the limits of peripheral cannulation, which could seem a too aggressive approach in a patient without prior sternotomy and can result in significant morbidity in critically ill patients [16].
Kredel et al. conclude from a setting of combined septic and cardiogenic shock that a double peripheral approach is a valuable alternative to central cannulation [20] – but with a single circuit and a formally vv aa configuration and artrial septostomy with one draining cannula for left ventricular unloading.
Four instead of two cannulae might be associated with higher mortality [16]. So our cannulation strategy with one double lumen cannula for vv ECMO together with subclavian arterial access for parallel va ECMO, without retrograde arterial flow, seems to combine some advantages in context with current evidence.
Two oxygenators in one circuit might have been beneficial for gas exchange but would not allow higher flows. All hybrid forms using one console have the problem that overall flow has to be split and a compromise has to be made between circulatory and oxygenating support. Contento et al. handle this problem with an additional pump in the venous line, behind one oxygenator and conclude better controllability together with lower propensity for clotting in a partially clamped limb [21].
Fig. 4: Graphical display of all D-dimer tests during ECMO support. System exchanges are spotted on the x-axis. Upgrade to parallel circuits on day 45
With our double ECMO setup, the same overall flow as with vv va configuration was delivered, but with measurably better PaO2 values. Higher flows would have been easily applicable if needed.
The exchange intervals were extended, and the indication changed from GEP resp. Hypoxemia to D-dimer elevation. The patient condition did not abruptly deteriorate during the exchange procedures. We affiliate this to two oxygenators both working far below their limits and one system always running while one is stopped. Complete stop of any support for exchange could have been potentially lethal. Acute drug- or infection-triggered liver failure led us to terminate the therapy. Had an organ been available, this could have been a successful lung-bridging period. Rosenbaum et al. [10] reported a successful bridging period of four months to heart, lung and kidney transplantation with a setup of rvad ECMO plus vv ECMO via double lumen cannula.
This case shows that it is feasible to run two parallel simultaneous ECMO circuits, one in vv and one in va configuration, over a prolonged period. This could be an option besides central cannulation and hybrid forms, when one system is pushed to the limits, as under septic shock conditions, especially when there is an oxygenation and a circulation problem.
DECLARATION OF CONFLICTING INTERESTS
The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
FUNDING
This research received no specific grant from any funding agency in the public, commercial, or non-profit sectors.
KOMMENTAR ZUM ARTIKEL:
„PARALLEL VV AND VA ECMO WITH THREE VESSEL ACCESS AS BRIDGE TO LUNG TRANSPLANTATION IN A PATIENT WITH SEPTIC SHOCK: THE PERFUSIONIST PERSPECTIVE.“
Der dargestellte Fall von Mertha et al. zeigt die Komplexizität und die wachsenden Herausforderungen während der ECMO-Therapie auf.
Nicht immer gelangt man mit einem Standard ECMO-Setup an das gewünschte Ziel und weiterführende Lösungen müssen gefunden werden. Insbesondere ist der hyperdyname Patient mit respiratorischem Partial- oder Globalversagen schwierig zu betreuen, da neben der schwer zu erreichenden systemischen Oxygenierung mitunter auch extreme Mengen an Kohlenstoffdioxid anfallen können, was einen Oxygenator schnell an seine Leistungsgrenze führt. Im vorliegenden Fall wurde nach einer Vielzahl von ungünstigen Ereignissen die Entscheidung getroffen, eine veno-arterio-venöse ECMO zunächst mit einer, später mit zwei Konsolen, zu betreiben. Es gibt Literaturberichte, in denen über den Einsatz zweier Systeme an einem Patienten berichtet wurde. Sie sind jedoch rar und es stellt einen neuen Ansatz zur Realisierung einer High-Flow ECMO > 7LPM dar. Allerdings ist es unabdingbar, exakt zu unterscheiden, welche Art der Unterstützung der jeweilige Patient benötigt. Hat der Patient eine erhaltene biventrikuläre Herzfunktion, so sollte man aufgrund der möglichen Komplikationen und ungleichmäßigen Blutverteilung von der va ECMO absehen und primär auf ein vv ECMO-System gehen. Ist die Oxygenierung und/oder Dekarboxylierung mit einem System/Oxygenator unzureichend, so muss man sich eine Strategie überlegen, um die gewünschte Performance zu erhalten. Im vorliegenden Fall gab es Dekarboxylierungsprobleme, denen mit häufigen Systemwechseln entgegnet wurde, um stets eine maximale Gastransferleistung des Oxygenators zu gewährleisten. Jeder Komponentenwechsel birgt Risiken und sollte so selten wie möglich durchgeführt werden.
Mittels zwei parallel geschalteten Oxygenatoren kann man eine höhere CO2-Auswaschung erreichen, sofern nicht der Volumenumsatz der ECMO das Problem ist und man ein ECMO-System zur Verfügung hat, welches getrennte Komponenten (Pumpe, Oxygenator) aufweist. Beschränkte Blutflussraten mit Doppellumenkanülen sind nicht selten. Hier kann man wie von Mertha et al. beschrieben, eine zweite Drainagekanüle implantieren, in der Hoffnung, einen höheren Blutstrom zu generieren.
Benötigt man im Verlauf einen Blutstrom jenseits der 5LPM, sollte man auf großlumigere Doppellumenkanülen wechseln oder eine große femorale Drainagekanüle implantieren und die Doppellumenkanüle nach Repositionierung als reine Rückgabekanüle nutzen oder auf bifemoral-venös umbauen.
Ein Umbau auf vav ECMO, um CO2 zu eliminieren, ist kritisch zu betrachten. Dies kann aber in Erwägung gezogen werden, wenn ein kardinales Versagen hinzukommen sollte. Eine spezifischere Herzunterstützung mittels Impella LV in Kombination mit vv ECMO oder ra-pa ECMO stellt ebenfalls eine elegante und gut steuerbare Maschinerie dar. Solange aber die Hämodynamik kein Problem darstellt, sollte man alles tun, um mit einer vv ECMO den pH- Wert und die Oxygenierung zu steuern und so eine hämodynamische Stabilität zu erzeugen.
Thomas Schultze, Halle
