A 63-year-old man presented with confirmed coronavirus disease 2019 (COVID-19) and comorbid conditions (hypertension and diabetes mellitus) but no history of prior vaccination. Initially, isolation at a residential treatment center was planned. However, due to worsening dyspnea and poor mask-driven oxygen saturation, the patient was transferred to our emergency room on the day after confirmation. Upon arrival, his blood pressure was recorded at 135/77 mm Hg, and his pulse rate was 101 beats per minute. The patient’s respiratory rate was 33 breaths per minute, and his oxygen saturation was 87%, despite a flow rate of 15 L/min through a non-rebreather mask. Even when a high-flow nasal cannula (60 L/min) was used with a fraction of inspired oxygen (FiO₂) of 100%, the measured partial oxygen pressure (PaO₂) remained low, at 68 mm Hg.

Endotracheal intubation was performed to initiate mechanical ventilation with a positive end-expiratory pressure (PEEP) of 12 cmH₂O at 100% FiO₂. The patient’s chest X-ray revealed diffuse bilateral pulmonary infiltrates, consistent with pneumonia. Unfortunately, our ventilatory approach of low tidal volume and high PEEP proved ineffective. Hypoxemia was aggravated (PaO₂/FiO₂ ratio, 60.9 mm Hg) and was accompanied by ventilator dyssynchrony and reduced lung compliance (tidal volume, 390 mL; peak inspiratory pressure, 31 mm Hg). As a bridge-to-recovery measure, we implemented venovenous (VV) extracorporeal membrane oxygenation (ECMO) after day 2 of intubation (Fig. 1).

Figure 1. Chest X-ray depicting bilateral haziness of the lung fields, with an endotracheal tube and venovenous extracorporeal membrane oxygenation cannulation. The distance between the cannulae was approximately 10.5 cm.

The patient weighed 62 kg, with a calculated body surface area of 1.74 m². Following an intravenous bolus injection of heparin (3,000 units), we inserted a 19F drainage cannula into the right common femoral vein and a 15F return cannula into the right internal jugular vein. The initial ECMO flow was set at 3.35 L/min, with gas flow maintained at 2.5 L/min (100% FiO₂). Intravenous heparin infusion was initiated and controlled by targeting an activated coagulation time (ACT) of 160–220 seconds. Despite these measures, the patient’s severe and seemingly intractable hypoxia remained unaffected by maximal VV ECMO support (blood flow, 4 L/min; gas flow, 7 L/min; FiO₂, 100%). Moreover, a systolic blood pressure of 90 mm Hg could not be maintained, even with the administration of norepinephrine (>0.3 μg/kg/min). Six days after the initiation of VV ECMO, the patient exhibited progressive oliguria (less than 25 mL/hr) and worsening lactic acidosis (increasing from 1.38 to 4.01 mmol/L in arterial blood). We were also faced with the challenge of managing aggravated hypercapnic acidosis (PaCO₂, 68 mm Hg; pH=7.1) and refractory hypoxemia (PaO₂, 44 mm Hg) with secondary cardiac dysfunction.

One option was conversion to veno-arterial-venous (V-AV) ECMO, a process that necessitates a brief interruption of the ECMO circuit for the insertion of a third cannula. However, even a minimal pause in ECMO risked jeopardizing the patient’s condition, which was being precariously maintained by the ECMO support. Additionally, while the patient required increased ECMO blood flow for enhanced oxygenation, the V-AV method merely divides and distributes the ECMO blood flow without augmenting the total flow. Consequently, we elected to implement a separate venoarterial (VA) ECMO circuit that would operate in conjunction with the existing VV circuit. This approach not only increases the total blood flow of the parallel ECMO but also allows for independent regulation of the VV and VA flows based on the patient’s oxygen and circulatory support needs.

After 6 days of VV ECMO, VA ECMO was initiated by inserting a 19F drainage cannula in the left common femoral vein and a 15F return cannula in the left common femoral artery (Fig. 2). Additional heparin administration was not necessary due to the ongoing infusion. The total flow of both circuits was 7–8 L/min (VV ECMO, 4.0 L/min; VA ECMO, 3.8 L/min), which gradually improved the patient’s oxygenation and hemodynamic status. However, after 7 days of parallel ECMO, complications arose, including hematochezia (120 g), bleeding at the VA ECMO cannulation site, and subconjunctival hemorrhage. We lowered the ACT target to 120–150 seconds, but due to persistent bleeding, the heparin infusion was eventually discontinued. A significant improvement in cardiac function was observed on transthoracic echocardiography compared to before the application of additional VA ECMO. Additionally, the systemic circulation was acceptably maintained without inotropic agents. Consequently, we could successfully wean the patient from VA ECMO 8 days after the initiation of parallel ECMO.

Figure 2. Parallel venovenous (VV) and venoarterial (VA) extracorporeal membrane oxygenation (ECMO) support during hospitalization for coronavirus disease 2019. (A) The circles indicate 2 parallel ECMO circuits applied in the intensive care unit. (B) Illustration of the parallel flow directions of VV ECMO (from the right femoral vein to the right jugular vein) and VA ECMO (from the left femoral vein to the left femoral artery). Written informed consent for the publication of this image was obtained from the patient.

We then attempted weaning of VV ECMO. Bedside tracheostomy was performed 18 days after intubation to facilitate long-term ventilatory support. We reduced the flow rates for VV ECMO (from 3.5 to 2.0 L/min) and gas delivery (from 4 to 1 L/min) and later discontinued gas flow for 1 day. On day 20 following initial VV ECMO insertion, the results of blood gas analysis were normal, allowing us to complete the decannulation. By the 34th day of hospitalization, the patient’s recovery had progressed enough to warrant a transfer to a less intensive facility. His oxygen saturation remained stable in room air, eliminating the need for a mechanical ventilator. However, his consciousness had not fully returned. To effectively improve his respiratory hygiene, the patient was transferred to a conservative care facility while still in a state of tracheostomy (Fig. 3).

Figure 3. Chest X-ray depicting resolving lung infiltrates (diffuse and bilateral) at discharge.

This study was reviewed and approved by the institutional review board of Eunpyeong St. Mary’s Hospital, and the requirement for informed consent was waived (IRB no., PC22ZASI0248).

ECMO has been widely used to treat COVID-19 acute respiratory distress syndrome when the condition is refractory to conventional management. According to one meta-analysis, approximately 7% of patients have undergone VV ECMO in these circumstances, with considerable in-hospital mortality (39%) [1]. If maximal VV ECMO support fails to adequately improve oxygenation, the insertion of an additional cannula may be warranted to augment venous drainage or facilitate cardiac support.

Two fundamental strategies exist for supplemental ECMO. The first is a hybrid approach, which involves adding peripheral arterial or venous cannulae to existing ECMO circuits via Y-connectors. The second is a parallel method, in which additional cannulae are connected to a separate ECMO circuit [2]. Lee et al. [3] reported a case of hybrid ECMO use involving conversion from VA to V-VA ECMO through the addition of a venous return cannula. Malik et al. [4] also described the use of parallel VV ECMO circuits to manage oxygenation in a patient with refractory hypoxemia and high cardiac output. While using standard VV ECMO with cannulation of the right common femoral vein and the right internal jugular vein, they established a second circuit involving cannulation of the left common femoral vein and left subclavian vein. Similarly, Patel et al. [5] implemented parallel ECMO circuits for patients with COVID-19 who maintained or developed severe hypoxemia despite maximal ECMO support with a single circuit. The researchers utilized bilateral femoral veins and bilateral internal jugular veins for parallel circuits. Cannulation with the left subclavian vein or left internal jugular vein carries risks of pneumothorax and vessel perforation due to the relatively angular venous pathway to the superior vena cava and adjacency to the pleura. We mitigated these risks by preserving the veins on the left side. Because no patient showed signs of cardiac failure, no transition was made to parallel VV and VA ECMO circuits [5]. To our knowledge, the present report is the first to describe provision of parallel VV and VA ECMO circuits for a patient with COVID-19 combined with cardiac dysfunction. In a similar report, Navas-Blanco et al. [6] employed VV and VA ECMO to address refractory hypoxemia and coexistent cardiovascular failure. They enhanced the cardiovascular support by transitioning from VV ECMO to hybrid VV-V ECMO, and ultimately to parallel VV and VA ECMO. Their patient was still undergoing ECMO treatment at the time of the report. Unlike their case, we directly implemented parallel ECMO without transitioning through an intermediate stage such as hybrid ECMO. Our patient is also the first to be successfully weaned from parallel VV and VA ECMO.

Compared to a hybrid alternative, parallel ECMO offers distinct advantages, including the ability to independently control blood flow through separate circuits. The hybrid approach can lead to unbalanced flow diversion due to differences in cannula size and length, as explained by the Hagen-Poiseuille equation. The addition of a cannula can also increase the total ECMO blood flow beyond the maximum level achievable with a single drainage cannula. In the present case, the total circuit flow increased from 3.35 to 8.0 L/min after conversion to parallel ECMO. Furthermore, the distribution of flow in a dual-circuit ECMO system can help alleviate shear stress and prevent hemolysis. This has been evidenced by a significant decrease in lactate dehydrogenase levels upon switching to parallel VV and VA ECMO [6]. Under these parallel conditions, an additional drainage cannula can increase the maximum achievable blood flow, while limiting negative drainage pressure and reducing the recirculation fraction [5]. Notably, manufacturer specifications suggest that pressure gradients across the proximal and distal ends of 19F cannulae may exceed 100 mm Hg, which could lead to hemolysis [7]. Ultimately, parallel ECMO is a relatively safe procedure that can enhance ongoing ECMO circulation without disturbance. In contrast, transitioning to a hybrid mode requires brief clamping of the ECMO circuit. During this process, critically ill patients are at risk of severe hypoxia and cardiac arrest.

While implementing parallel ECMO, certain factors require careful consideration. The strategy used with our patient involved positioning 2 cannulas within the inferior vena cava, which raised concerns about potential thrombotic complications [4]. After running parallel ECMO for 7 days, we encountered bleeding complications, leading us to halt the heparin infusion. Fortunately, no thrombotic events occurred before the ECMO weaning process was completed. The optimal anticoagulation protocol for parallel ECMO is yet to be established and warrants further research. Additionally, if both femoral veins have ECMO cannulae installed, transfemoral percutaneous venting of the left ventricle (LV) could pose a problem, particularly in cases where LV hypocontractility or pulmonary edema worsens. Moreover, the cost of a second ECMO circuit could be prohibitive or not covered by insurance at all. The eligibility criteria for parallel ECMO or applicable payment restrictions (1 circuit versus 2 circuits) will likely be determined by regional coverage policies. Patel et al. [5] opted to treat patients with parallel ECMO based on evidence of physiological need. A parallel circuit was reserved for patients who maintained or developed severe hypoxemia despite maximal ECMO support with a single circuit. This required a combination of lung-protective ventilator settings and adjunctive therapies, including deep sedation, neuromuscular blockers, and inhaled pulmonary vasodilators [5].

In conclusion, parallel ECMO may be an effective approach for patients with concomitant refractory cardiopulmonary failure despite best efforts applied through conventional ECMO. Broader discussions of the indications for and management of parallel ECMO techniques are urgently needed.

Author contributions

Conceptualization: ESK, YHK. Data curation: ESK, SA. Formal analysis: ESK, YHK, JL. Methodology: ESK, YHK, JL, Project administration: SA, YHK. Visualization: ESK, YHK, JL. Writing–original draft: ESK, JL. Writing–review & editing: ESK, YHK. Final approval of the manuscript: all authors.

Conflict of interest

No potential conflict of interest relevant to this article was reported.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.