Pulse Pressure During VA-ECMO Support and Neurologic Outcomes: Worth Getting Pumped Up About?

Veno-arterial extracorporeal membrane oxygenation (VA-ECMO) is increasingly used for the management of refractory cardiogenic shock. Although VA-ECMO can stabilize hemodynamics and improve systemic flow, morbidity, and mortality on support remain high.1 Neurologic injuries, including ischemic brain injuries and central nervous system bleeding, are common in patients supported by VA-ECMO and are important contributors to morbidity and mortality.2 Previous studies have identified longer duration of support, higher pump flow, hyperoxemia, and other factors as predictors of acute brain injury (ABI) while on VA-ECMO.3 Due to the physiology of continuous flow cerebral perfusion on VA-ECMO, research has also focused on low pulse pressure during support as a potential mechanism for ABI. One small, single-center cohort previously demonstrated that low pulse pressure was associated with ABI.4 To further investigate the relationship between pulse pressure and ABI during VA-ECMO support, Kalra et al.5 analyzed data from the global Extracorporeal Life Support Organization (ELSO) Registry, with findings published in this issue of the ASAIO Journal. Kalra et al. examined the ELSO dataset for adults aged 18 and older who were supported with peripheral VA-ECMO for cardiogenic shock between 2018 and 2023. After exclusion criteria, the authors identified 8,894 patients, of whom 1,096 (11.1%) developed ABI during VA-ECMO support. The median pulse pressure was 27 mm Hg (interquartile range 14–4 2 mm Hg) in patients with ABI and 30 mm Hg (interquartile range 17–44 mm Hg) in patients without ABI, which was statistically significant. The authors also found that low pulse pressure (categorically defined as ≤10 mm Hg) was associated with ABI in both univariate and multivariable analyses. Other predictors of ABI in multivariable analyses included age, Charleson Comorbidity Index, pre-ECMO cardiac arrest, and on-ECMO ΔPaCO2, PaO2, hemolysis, arrhythmia, and renal replacement therapy. The authors concluded that low pulse pressure at 24 hours of VA-ECMO support was independently associated with ABI, with the findings led by the association between low pulse pressure and central nervous system ischemic events. The study by Kalra et al. is an important contribution to the literature on ABI in VA-ECMO support and highlights areas for further consideration and discussion. First, several important limitations should be noted. Almost half of the patients with VA-ECMO in the ELSO registry were excluded from the main analysis, primarily due to missing data or due to concomitant utilization of mechanical circulatory support for ventricular unloading. Patients with mechanical support for ventricular unloading were excluded from the main analysis because of potential difficulties in interpretation of pulse pressure, for example in the case of concomitant intra-aortic balloon pump use. Additionally, the timing of the ABI relative to the pulse pressure measurements (performed at 24 hours of support based on the data available in ELSO) is not reported. Although the authors adjust for relevant baseline differences between the lower and higher pulse pressure groups, residual confounding based on differences in comorbidity severity and both pre- and post-ECMO clinical course remains possible. The main takeaway question from the findings by Kalra et al. remains: to what extent is low pulse pressure merely a marker of adverse clinical course during support versus a modifiable target in trying to reduce the incidence of ABI? The authors suggest several clinical implications of their findings, including the importance of surveillance for low pulse pressure and potential modification of ECMO settings or vasopressor support to increase pulse pressure. Whether a certain pulse pressure threshold is required to minimize the detrimental effects of continuous flow on neurologic outcomes is unclear. Also, the best options to achieve a pulse pressure closer to physiologic norms without sacrificing systemic flow or peripheral perfusion remain to be determined. In the future, these questions may be best addressed by examining cohorts who undergo care with protocols to optimize pulse pressure until randomized trials of pulse pressure targets can be performed. Additionally, the current study highlights the need to explore pulsatile VA-ECMO systems, which are currently undergoing evaluation.6 Finally, future studies will be needed to determine whether the pulse pressure findings in the study by Kalra et al. will remain applicable in the setting of VA-ECMO plus mechanical unloading particularly utilizing Impella (ABIOMED, Danvers, MA), which is becoming increasingly common.7 This is important because the sensitivity analysis performed by Kalra et al. limited to the smaller number of patients with temporary mechanical support for ventricular venting demonstrated no significant relationship between narrow pulse pressure and ABI. In conclusion, this study by Kalra et al. provides important insights into the association between low pulse pressure and ABI in patients with cardiogenic shock supported by peripheral VA-ECMO. Although awaiting additional data, multidisciplinary treatment teams that currently provide care to VA-ECMO patients should remain vigilant to optimize neurologic health and should appreciate the role of low pulse pressure with its prognostic and potential treatment modification implications.