Left Atrial Decompression on Venoarterial Extracorporeal Membrane Oxygenation: Getting to the Heart of the Matter*

Unauthorized reproduction of this article is prohibited Copyright © 2019 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies. Unauthorized reproduction of this article is prohibited 780 www.pccmjournal.org August 2019 • Volume 20 • Number 8 Many aspects of the care of patients on extracorporeal membrane oxygenation (ECMO) are driven by both institutional and provider “religions” rather than scientific evidence or at the very least—consensus. From sedation regimens to anticoagulation targets and ventilator strategies, there has yet to really be standardization in practice across institutions. In this issue of Pediatric Critical Care Medicine, Zampi et al (1) from the Pediatric Interventional Cardiology Early Career Society (PICES) tackle the important yet controversial topic of the timing (and approach) for left atrial (LA) decompression during venoarterial ECMO. Many of us have struggled to care for (and regretted not intervening earlier on) patients in whom venoarterial ECMO cannulation is associated with the rapid or insidious development of pulmonary edema secondary to LA hypertension in the setting of left ventricular (LV) failure without adequate ejection but with continued right ventricular ejection. This results in worsening pulmonary dynamics and the continued need for respiratory as well as cardiac extracorporeal support. Additionally, although difficult to quantify, the resultant left heart distention is likely to impact upon the ability to provide effective myocardial rest during ECMO. Through their well-done 11 site descriptive study of 137 patients cared for in PICES centers, they make a sound and well thought out argument for early consideration of LA decompression in patients on venoarterial ECMO as a way to reduce length of time on ECMO and decrease the duration of mechanical ventilation. In this patient cohort, a majority of patients had a cardiomyopathy (47.4%) or myocarditis (16.8%) with severely depressed LV function, of whom 41.6% required extracorporeal cardiopulmonary resuscitation and had mean pre-ECMO lactates of greater than 7. With these particular etiologies and events, providing relief to the myocardium via adequate unloading to allow for the best chance at recovery is imperative. Multiple studies have shown that persistent ventricular distention fails to appropriately decrease myocardial oxygen demand and can promote ongoing ventricular injury (2–4). However, data to date have conflicted on optimal timing and overall utility of LA decompression, with some authors such as Baruteau et al (5) arguing it is beneficial, whereas Mistry et al (6) showed no difference in outcomes between patients who did and did not undergo decompression. Clearly, the impact of selection bias, access to the use of ventricular assist device in selected patients, as well as details of adequacy of LV ejection, aortic valve opening, and even the presence of a native interatrial communication may affect these observations. As an institution where LA decompression is performed, we are believers in its importance and impact on outcomes. As Zampi et al (1) astutely point out, although this procedure does not in and of itself have an impact on mortality, using mortality as an outcome measure at this point in time is too blunt of an instrument by which to gauge the effectiveness of a procedure. Reduction in mortality, if used as the barometer by which studies were measured, would limit the number of studies that could be published to help us guide practice. Prolonged time on ECMO increases the risks of neurologic, infectious, bleeding, and other major complications while also predicting worse functional status at time of discharge (7–10). The study by Zampi et al (1) presents the largest-to-date published cohort, and shows a shorter time to decannulation from ECMO in those patients who underwent early LA decompression; therefore, it is not unreasonable to infer the complication *See also p. 728.