Venoarterial extracorporeal membrane oxygenation (ECMO) is a temporizing therapy for patients with refractory cardiogenic shock, and it may also be a bridge to destination therapy or cardiac transplantation. Patients requiring venoarterial ECMO are, by definition, on the verge of cardiopulmonary collapse. Acute disruption of forward flow in ECMO-dependent patients will result in severe hypoperfusion and must be avoided at all costs. It is common practice to use contrast transthoracic echocardiography to improve the visualization of endocardial borders and allow more accurate assessments of ventricular function. This information is used clinically to determine whether to wean patients from ECMO support.

The ECMO system used at our institution (CARDIOHELP; MAQUET Medical Systems USA, Wayne, NJ) has an integrated sensor that uses ultrasound to detect changes in flow dynamics associated with the presence of air bubbles or thrombi >5 mm in dimension. Activation of the bubble detector results in an alarm, followed by pump shutdown. If the alarm is not overridden within 6 sec, “zero-flow mode” is engaged. This mode applies sufficient revolutions per minute to prevent backflow from the arterial cannula into the venous cannula and prevents forward flow as well, hence the name zero-flow mode. This safety feature is designed to prevent deleterious cardiovascular and neurologic sequelae associated with air embolism during ECMO. Nevertheless, although this is a safety feature, stopping the ECMO circuitry and therefore the blood flow may be followed by deleterious consequences such as hypotension or the occurrence of thrombi.

Although the administration of contrast agents has been deemed safe, we have had two instances at our institution in which the peripheral administration of a microbubble contrast agent, namely, Definity (perflutren lipid microsphere; Lantheus Medical Imaging, North Billerica, MA) during contrast transthoracic echocardiography resulted in activation of the ECMO bubble detector alarm and subsequent engagement of zero-flow mode, resulting in hemodynamic instability. Expedient restoration of flow is necessary to avoid grave hemodynamic consequences in patients who are already refractory to maximum inotropic and vasopressor therapy, as well as organ ischemia and circuit thrombosis. However, the need to restore forward flow must be balanced with the need to detect and prevent arterial embolization. We believe that the cause of the air bubble alarm in these cases was the administration of a contrast medium specifically designed to enhance ultrasonographic imaging, which magnified the signal perceived by the air bubble detector. It seems intuitive that the administration of echocardiographic contrast could create an ultrasonographic pattern consistent with turbulent flow that may be interpreted by the sensor as an embolus or a similar occurrence. In any case, it is detected as an event worthy of alarm that requires immediate clinical evaluation to prevent a shutdown for safety reasons. An unexpected alarm of this level of clinical significance should not be haphazardly silenced or overridden.

Further investigation into the circumstances surrounding these incidents revealed several learning points that may be helpful in anticipating and avoiding similar events in the future. Safely proceeding with contrast echocardiography in an ECMO patient requires that the provider anticipate and plan for potential sensor alarms and their downstream consequences, including engaging a zero-flow state. We recommend that the circuit be inspected visually before contrast administration, to minimize the chance of a true embolic event being disregarded as an artifact. An experienced team should be present during the study and should include, if possible, the attending cardiothoracic surgeon, the intensivist, and the perfusionist in addition to the echocardiography team. If possible, contrast transthoracic echocardiography in ECMO patients should be performed during regular working hours, when an abundance of resources and personnel are readily available to aptly handle such occurrences; other possible approaches are to completely avoid the use of contrast agents in ECMO patients and to temporarily disconnect the bubble alarm while a study is under way. In any event, further data are needed before a valid decision can be made on the best way to proceed in these circumstances.