Background: Pulse pressure (or pulsatility) is a dynamic marker of cardiovascular function and is often decreased in patients on veno-arterial extracorporeal membrane oxygenation (VA-ECMO), which provides non-pulsatile flow. We explored the impact of pulse pressure on development of acute brain injury (ABI) in VA-ECMO patients. Methods: We retrospectively analyzed all adult (≥18 years old) patients receiving VA-ECMO at a single tertiary care center between 7/2016 and 1/2021. All patients received routine neurocritical care consultations with standardized neuromonitoring. ABI included intracranial hemorrhage, ischemic stroke, hypoxic ischemic brain injury, cerebral edema, seizure, and brain death. Blood pressure (BP) was recorded every 15 minutes during the first 24 hours after ECMO initiation. Pulse pressure (systolic minus diastolic pressure) was categorized as: very low (<10mmHg), low (1029mmHg), moderate (30-44mmHg), and high (≥45mmHg). Multivariable logistic regression, controlling for baseline illness severity, operative characteristics, and left ventricle (LV) venting strategy, was performed for the association of pulse pressure with ABI. LV venting strategies included intra-aortic balloon pump (IABP), percutaneous ventricular assist device, right superior pulmonary vein, and left ventricle apex vents. Results: We collected 9,179 pressure measurements from 123 VA-ECMO patients (median age=63, 63% male), of which 41 (33%) experienced ABI. Demographics, baseline clinical characteristics, and pre-ECMO arterial blood gases were not significantly different between those with and without ABI. Those with ABI had a lower median pulse pressure than those without (20 mmHg vs. 32, p=0.038). In a multivariable model controlling for cannulation site, preECMO lactate, and left ventricle venting strategy, low (adjusted odds ratio/aOR=7.04, 95%CI: 1.72-28.82, p=0.007) and very low pulse pressure (aOR=5.14, 95%CI: 1.18-22.34, p=0.029) were associated with ABI (Figure 1). In a model with the same covariates, every 10-mmHg decrease in pulse pressure was associated with a 1.29 increased adjusted odds of ABI (95%CI: 1.00-1.65, p=0.047) (Figure 2). In a sensitivity analysis, adjusting for systolic BP yielded similar associations between pulse pressure and ABI, suggesting pulse pressure acts independently of BP in mediating ABI. Conclusions: Decreased pulsatility in the first 24 hours of VA-ECMO was associated with ABI. Special attention should be given to management practices which sustain sufficient pulse pressure early during ECMO. Figure 1. Forest plot for multivariable model of pulse pressure and acute brain injury (ABI), with the covariates of central cannulation, lactate on ECMO day 1, and left ventricle venting strategy. Covariates are not shown in the figure. Dots represent adjusted odds ratios and brackets represent 95% confidence intervals. Low and very low pulse pressure were associated with increased odds of ABI. This model had a C-index of 0.76. Figure 2. Multivariable logistic regression model with pulse pressure treated as a continuous variable, with the covariates of central cannulation, lactate on ECMO day 1, and left ventricle venting strategy. The adjusted relative odds of acute brain injury decrease as pulse pressure increases.
Read full abstract