A Rare Cause of Severe Biventricular Dyssynchrony During Venoarterial Extracorporeal Membrane Oxygenation for COVID-19 Respiratory Failure

A midesophageal 4-chamber TEE view revealed diastolic flattening of the right and left ventricular (RV and LV, respectively) apices associated with dyssynchronous biventricular contraction (Figs 1 and 2; Video 1). A small pericardial effusion was also present. The midesophageal RV inflow-outflow view revealed pronounced RV outflow tract narrowing during diastole (Figs 3 and 4; Video 2). Similar findings were observed in the midesophageal aortic valve long axis view (Figs 5 and 6; Video 3). The transgastric midpapillary short axis view showed severely dyssynchronous LV contraction with apparent tethering of the LV anterior wall (Figs 7 and 8; Video 4). Collectively, the TEE findings strongly suggested that an extracardiac force was acting on the heart, as if cardiac compressions were being performed. The patient continued to experience hypoxemia (arterial oxygen saturations in the mid-80s), hypotension despite escalating infusion rates of vasoactive drugs, and frequent ECMO suction events that only could be managed with reduced flow rates. A chest radiograph was obtained to confirm position of a nasojejunal feeding tube that had been placed when the TEE examination was completed. A left pneumothorax was observed with notable flattening of the left heart border by the compressed lung parenchyma (Fig 9). A thoracostomy tube was inserted, which resulted in immediate improvement in the patient's arterial oxygen saturation and hemodynamics, a decrease in vasoactive drug requirements, and the ability to safely increase ECMO flow rates. A repeat TEE examination showed complete resolution of the abnormal cardiac motion.

Pneumothorax occurs in 0.16% to 2% of patients with COVID-19, but its incidence may be much higher (approaching 25%) in those who require mechanical ventilation. Tension pneumothorax during ECMO has been previously described in neonates, but to the authors’ knowledge, the current report is the first description of this complication in an adult with COVID-19 treated with VA-ECMO. A triad of paradoxically increased arterial oxygen tension, evidence of reduced peripheral perfusion (as indicated by narrowed pulse pressure and decreases in mixed venous oxygen saturation), and declining flow rates accompanied by hemodynamic instability were the characteristic features of tension pneumothorax in neonates treated with ECMO. These findings were replicated in a study of experimentally-induced tension pneumothorax in dogs receiving VA ECMO. The current patient treated with VA-ECMO displayed 2 of 3 of these previously reported signs of tension pneumothorax: compromised peripheral perfusion and declining ECMO flows rates associated with suction events and hemodynamic instability requiring high-dose vasoactive drug support. However, the current patient also developed hypoxemia and not an increase in arterial oxygen tension. A difference in ECMO cannulation site and the resulting location of the mixing cloud were probably responsible for this discrepancy. The common carotid artery and internal jugular vein were cannulated in neonates and dogs, whereas femoral-femoral arterial and venous cannulation was used here. Thus, arterial blood gas sampling from a right radial arterial catheter and the decrease in arterial oxygen tension observed in the current patient probably reflected a more distal mixing cloud.

The echocardiographic findings in patients with tension pneumothorax are rarely reported, most likely because of the emergent nature of the pathology. The major TEE findings in the current patient are similar to those that have been previously reported in the literature. Castro et al recently described their TEE findings in 2 ECMO patients with tension pneumothorax in whom dyssynchronous biventricular contraction and flattening of both ventricular apices were observed. Pneumopericardium was suspected initially in the first patient (who was receiving VA-ECMO after a pulmonary thromboembolectomy), but a chest radiograph revealed a large left pneumothorax. The RV of this patient also appeared to be tethered, similar to the current patient's LV in the transgastric midpapillary view (Figs 7 and 8). The other patient in the Castro et al series was undergoing venovenous ECMO cannulation under TEE guidance when acute changes in cardiac motion consisting of dyssynchronous contraction with diastolic flattening were observed. An iatrogenic pneumothorax after central venous catheter placement was suspected and subsequently confirmed with lung ultrasound imaging. A second case report described a patient undergoing minimally invasive coronary artery surgery who developed hypotension and elevated airway pressures during single-lung ventilation. TEE revealed a mass effect on the right atrium resulting from a right tension pneumothorax. This abnormality resolved after the right pleura was opened.