Extracorporeal cardiopulmonary resuscitation (eCPR) and cerebral perfusion: A narrative review

Abstract

Extracorporeal cardiopulmonary resuscitation (eCPR) is emerging as an effective, lifesaving resuscitation strategy for select patients with prolonged or refractory cardiac arrest. Currently, a paucity of evidence-based recommendations is available to guide clinical management of eCPR patients. Despite promising results from initial clinical trials, neurological injury remains a significant cause of morbidity and mortality. Neuropathology associated with utilization of an extracorporeal circuit may interact significantly with the consequences of a prolonged low-flow state that typically precedes eCPR. In this narrative review, we explore current gaps in knowledge about cerebral perfusion over the course of cardiac arrest and resuscitation with a focus on patients treated with eCPR. We found no studies which investigated regional cerebral blood flow or cerebral autoregulation in human cohorts specific to eCPR. Studies which assessed cerebral perfusion in clinical eCPR were small and limited to near-infrared spectroscopy. Furthermore, no studies prospectively or retrospectively evaluated the relationship between epinephrine and neurological outcomes in eCPR patients. In summary, the field currently lacks a comprehensive understanding of how regional cerebral perfusion and cerebral autoregulation are temporally modified by factors such as pre-eCPR low-flow duration, vasopressors, and circuit flow rate. Elucidating these critical relationships may inform future strategies aimed at improving neurological outcomes in patients treated with lifesaving eCPR.

INTRODUCTION

Although technically challenging and resource intensive, veno-arterial extracorporeal membrane oxygenation (VA-ECMO) can be a potent therapy for select patients with refractory cardiac arrest due to an initial shockable rhythm.¹ Commonly known as extracorporeal cardiopulmonary resuscitation (eCPR), this lifesaving therapy restores adequate perfusion to vital organs and allows time for providers to treat the underlying cause of the arrest. eCPR has expanded significantly since it was first demonstrated in 1957.² In most cases of out-of-hospital cardiac arrest (OHCA), the ECMO cannulas are placed percutaneously in the femoral vessels. Blood is drained from the cavoatrial junction, passed through the gas exchanger, and then returned to the abdominal aorta in a retrograde manner where the heart and brain are subsequently reperfused.

Evidence is accumulating in support of eCPR over standard advanced cardiovascular life support (ACLS) for select patients.^3,4 Randomized data are now available from two clinical trials.^1,5 eCPR was superior to standard ACLS in the ARREST trial which included 30 patients with refractory ventricular fibrillation,¹ although a clear benefit was not observed in the Prague OHCA study which included both shockable and non-shockable rhythms.⁵ Results are pending for two additional randomized trials: the INCEPTION trial (NCT03101787) and the APACAR2 trial (NCT02527031).

eCPR patients are no exception to the general observation that acute brain dysfunction remains pervasive in critical illness including extracranial critical illness. Indeed, neurological injury remains a significant cause of morbidity and mortality in eCPR patients with OHCA. Treatment recommendations are largely based on extrapolation from studies which were not specific to eCPR patients. Surprisingly little attention has been paid to the role of cerebral perfusion. OHCA eCPR patients represent a unique population in which cerebral perfusion is rapidly restored after a prolonged ischemic insult. Cerebrovascular pathophysiology and optimal management strategies are poorly characterized in this context. A better understanding of current shortcomings in our knowledge of these factors may inform future investigations centered on ultimately improving neurological outcomes for patients treated with eCPR. In this narrative review, we explore current gaps in knowledge about cerebral perfusion over the course of cardiac arrest and resuscitation with a focus on patients treated with eCPR. Articles published in English from major databases (PubMed/MEDLINE, Embase, and Google Scholar) were identified by search with combinations of the keywords “eCPR,” “cerebral perfusion,” “epinephrine,” “cardiac arrest,” “cerebral autoregulation,” and “cerebral blood flow.” Additional articles were identified by in-text citations. Case reports and conference abstracts were not included.

NEUROLOGICAL INJURY IN ECPR PATIENTS

Despite promising evidence in support of eCPR, neurological injury remains a significant burden for patients, although data on long-term neurological function in survivors are limited. It is difficult to attribute the occurrence of these pathologies to a single culprit such as the ECMO circuit, clinical management, or ischemia/reperfusion injury.⁶ One retrospective single-center analysis which included 274 eCPR cases reported that anoxic brain injury accounted for 34% (69/203) of deaths.⁷ A recent meta-analysis which included 6,261 eCPR cases found that 27% of patients experienced at least one neurological complication after eCPR.⁶ Hypoxic-ischemic brain injury was the most common type of injury reported with an overall rate of 23%. Other neurological complications included ischemic stroke, seizures, and intracerebral hemorrhage. Seventeen percent of patients included in the analysis developed brain death.

CEREBRAL BLOOD FLOW

Cerebral Blood Flow in Cardiac Arrest and Conventional CPR

Cessation of forward blood flow to the brain after cardiac arrest leads to global primary hypoxicischemic brain injury characterized by cytotoxic edema, intracellular acidosis, calcium influx, and excitatory neurotransmitter release.⁸ Reestablishment of cerebral blood flow (CBF) by return of spontaneous circulation (ROSC) initiates reperfusion injury involving aberrant production of reactive oxygen species coupled with impaired nitric oxide production, endothelial dysfunction, intracellular calcium accumulation, and further release of excitatory neurotransmitters.⁸

In non-ECMO patients, patterns of global CBF after ROSC are generally categorized into four distinct phases which are discussed in detail elsewhere.⁹ Briefly, these phases are described as follows: 1) Immediate hyperemia (0 to 20 min.); 2) Early hypoperfusion (20 min. to 12 h); 3) Intermediate (12 h to 72 h); and 4) Recovery (>72 h).⁹ These generalizations come with the caveat that cerebral perfusion is heterogeneous, varying by brain region, ischemic duration, age, and cardiac arrest etiology.^9–11 In pediatric models, resuscitation after a more prolonged insult tends to produce less frequent hyperemia with more pronounced hypoperfusion;¹² however, the relationship between arrest duration and regional CBF has not been studied in adult models of cardiac arrest.⁹ At the level of the microcirculation, areas of no-reflow are observed, which likely contribute to pathology.⁸ Recent work demonstrates a striking correlation between early recovery of regional CBF in the thalamus with neurological outcome in rats subjected to asphyxia cardiac arrest, although further translational validation is needed to determine whether these findings are clinically relevant.¹³

Cerebral Blood Flow in eCPR

It remains to be determined whether general patterns of post-ROSC cerebral perfusion are relevant for the eCPR patient in which flow is rapidly restored by VA-ECMO after a prolonged period of CPR. For example, eCPR is often preceded by up to 60 minutes of chest compressions which confer only 15–30% of normal cardiac output.^14,15 Lactic acidosis and arterial pressure of carbon dioxide (PaCO₂) progressively increase during prolonged CPR times despite utilization of an automatic chest compression device.¹⁴ Reperfusion injury may be exacerbated in the case of eCPR where oxygenated blood flow is rapidly restored after prolonged hypoxia.

Studies which reported measurements of cerebral perfusion during eCPR are limited (Table 1). It is difficult to directly compare studies due to heterogeneity in several areas including experimental design, ECMO circuit flow rate, use of vasopressors, and the specific parameter of near-infrared spectroscopy (NIRS) utilized.^16–20

Table 1.

Summary of studies reporting on cerebral perfusion during eCPR in humans or swine. Case reports and conference abstracts were not included.

Reference	Subjects	Method	Parameters	Timepoints	Highlights
Bělohlávek et al.91	11 pigs	NIRS, Doppler guide wire	rSO2, carotid flow velocity	Arrest	In the absence of ROSC, rSO2 and carotid flow velocity increased after initiation of VA-ECMO. Addition of IABP counterpulsation did not affect these parameters.
Spinelli et al.27	13 pigs	brain O2 tension probe, ICP catheter	PBro2, ICP, CePP	Baseline, arrest, 0–6 h eCPR	eCPR quickly restored PBro2 and CePP to pre-arrest values. ICP values remained unchanged throughout VF and early eCPR, but significantly increased after 5 h of eCPR.
Luo et al.83	18 pigs	Transit time flow probe	Carotid blood flow	Baseline, arrest, 0–6 h eCPR	Carotid blood flow was significantly reduced during low flow VA-ECMO (30–35 ml/kg/min) compared to standard flow VA-ECMO (65–70 ml/kg/min) over 6 h of eCPR.
Wollborn et al.92	29 pigs	NIRS, transcranial Doppler US	rSO2, MCA blood flow velocity	Baseline, arrest, 0–6 h post-ROSC	rSO2 recovered by 30 min post-ROSC. Mean MCA blood flow velocity was impaired at 3, and 6 h post-ROSC in animals treated with conventional CPR or eCPR.
Ölander et al.93	10 pigs	Transit time flow probe, ICP catheter	Carotid blood flow, ICP, CePP	Baseline, 15–45 min CPR, 30–180 min eCPR	Initiation of eCPR restored carotid blood flow and partially restored CePP to pre-arrest values. Minimal changes in ICP were observed throughout the study.
Levy et al.43	12 pigs	NIRS, transit time flow probe, ICP catheter, jugular venous catheter	rSO2, carotid blood flow, ICP, CePP, SjvO2, PRx	Baseline, 0–30 min eCPR, 0–120 min post-ROSC	Targeting a higher MAP (80–90 mmHg vs. 65–75 mmHg) via epinephrine transiently improved PRx during eCPR but worsened cerebral hemodynamics after ROSC.
Ölander et al.94	12 pigs	Transit time flow probe, ICP catheter	Carotid blood flow, ICP, CePP	Baseline, CPR, 0–180 min eCPR	Initiation of eCPR partially restored carotid blood flow and CePP to pre-arrest values. Minimal changes in ICP were observed throughout the study.
Yagi et al.16	15 patients with OHCA	NIRS	TOI	Arrival, post-vasopressin, 0–2 h eCPR	In 14 patients with poor neurological outcome, TOI increased after administration of vasopressin and further increased after initiation of VA-ECMO + IABP. TOI decreased during eCPR in one patient with good neurological outcome.
Ehara et al.17	16 patients with OHCA	NIRS	rSO2	CPR, 2–10 min eCPR	Initiation of eCPR increased rSO2 in patients with a poor neurological outcome. rSO2 did not change with eCPR in patients with good neurological outcome.
Bartos et al.18	83 patients with OHCA	NIRS	Not specified	0, 24, 48 h eCPR	“NIRS values” increased from baseline over the first 48 h in survivors. Values remained stable in patients that died, and values declined in patients with brain death.
Yagi et al.20	18 patients with OHCA	NIRS	TOI	CPR, 0–20 min eCPR	TOI increased immediately after initiating eCPR. Neurological outcomes were not reported.
Roellke et al.19	6 patients with IHCA	NIRS	rSO2	CPR, 2.5–5 min eCPR	rSO2 increased immediately after initiating eCPR. No patients survived to hospital discharge.

Abbreviations: CePP, cerebral perfusion pressure; CPR, cardiopulmonary resuscitation; eCPR, extracorporeal cardiopulmonary resuscitation; IABP, intra-aortic balloon pump; ICP, intracranial pressure; IHCA, in-hospital cardiac arrest; MCA, middle cerebral artery; NIRS, near-infrared spectroscopy; OHCA, out-of-hospital cardiac arrest; PRx, pressure reactivity index; rSO₂, regional oxygen saturation; SjvO2, jugular venous oxygen saturation; StO₂, tissue hemoglobin saturation; TOI, tissue oxygenation index; VA-ECMO, veno-arterial extracorporeal membrane oxygenation; VF, ventricular fibrillation.

In two small observational studies (n=15, 16) which conducted NIRS for up to two hours after initiating eCPR, oximetry values tended to increase in patients with poor neurological outcome but decrease or remain unchanged in patients with good neurological outcome.^16,17 It is hypothesized that this phenomenon could be explained by a quick recovery of cerebral metabolic rate of oxygen in patients that ultimately experienced good neurological outcome, and vice versa.^16,17 Interestingly, different trends were observed when NIRS was conducted at later timepoints post-arrest. In an observational study by Bartos et al.¹⁸ which included 83 eCPR patients, NIRS was conducted at hospital arrival and over the first 48 hours after initiating eCPR. Oximetry values compared to baseline were increased at 24 hours and 48 hours in survivors. Values remained stable in non-survivors over the same timeframe but decreased in those with brain death.

CEREBRAL AUTOREGULATION

Cerebral Autoregulation and Monitoring Methods

The healthy brain maintains constant CBF over a wide range of cerebral perfusion pressures (CePP) (classically reported as 50–150 mmHg) by regulating arteriolar tone.²¹ This process, termed cerebral autoregulation, is governed by myogenic, neurogenic, metabolic, and endothelial mechanisms which are not completely understood.²² Although cerebral autoregulation is a relationship between CePP and CBF (i.e. the Lassen curve²¹), assessment in humans involves less invasive means in which CePP is replaced by mean arterial pressure (MAP) and CBF is replaced by either regional cerebral oxygen saturation (measured by NIRS) or middle cerebral artery blood flow velocity (measured by transcranial doppler, TCD).^23,24

MAP may serve as an adequate surrogate for CePP in cases where intracranial pressure (ICP) is constant (i.e., CePP = MAP - ICP). However, it should be noted that ICP can fluctuate after cardiac arrest.^25–27 Unfortunately, use of an invasive ICP monitor is unavailable for eCPR patients due to the need for cessation of anticoagulation which is currently not feasible,²⁸ although non-invasive methods to estimate ICP are currently under evaluation.²⁹ Estimation of CBF by TCD assumes that the diameter of the artery of interest remains constant throughout the study period.²³ Caution should be exercised when using NIRS to assess cerebral perfusion during or shortly after cardiac arrest, as measurements can be confounded by vasopressor-induced cutaneous vasoconstriction^26,30–33—a phenomenon which may depend on the specific oximeter used in the study.³⁴ Additionally, oximetry values could be influenced by cerebral metabolic rate of oxygen, which may vary according to prognosis.^{16,17 15}O-water positron emission tomography is considered the gold standard for in vivo measurement of CBF, although routine clinical use is limited by cost, complexity, and need for arterial cannulation.³⁵

In cases where autoregulation is absent, cerebral perfusion becomes “pressure passive” wherein CBF increases linearly with CePP.⁸ In this manner, the brain becomes vulnerable to hypoperfusion when CePP is low, yet vulnerable to relative hypertension when CePP is high. Autoregulation may also be right-shifted which increases CePP requirements for achieving adequate CBF, contributing to secondary ischemic brain injury when CePP falls below a pathologically elevated lower limit of autoregulation.²³

Cerebral Autoregulation in Cardiac Arrest

Dysfunctional cerebral autoregulation is observed in many patients after resuscitation and this phenomenon is associated with poor outcomes.^36,37 Nishizawa et al.³⁸ described autoregulation to be absent three days post-arrest in eight comatose patients. In a notable study by Sundgreen et al.,²³ cerebral autoregulation was assessed in 18 patients within the first 24 hours after resuscitation. Autoregulation was absent in eight patients and right-shifted in five patients (median 114 mmHg, range 80–120 mmHg). Considering this heterogeneity, it has been proposed that the optimal management for patients should be an individualized approach in which MAP is maintained over the lower limit of autoregulation identified by non-invasive methods such as NIRS or TCD.^8,9,36,37 Two small feasibility studies (n=23, 51) have demonstrated that an optimal MAP target may be individualized based on NIRS assessment of cerebral autoregulation post-arrest;^36,37 however, this strategy requires validation by large prospective clinical trials.³⁹ It should be noted that accuracy of NIRS-based assessment of optimal MAP has been called into question by a small study (n=10) which found limited agreement when directly compared to invasive measurements.⁴⁰

In support of an individualized approach to MAP management, recent results from small randomized controlled trials demonstrated that neurological endpoints were not improved by an unindividualized “one-size-fits-all” approach.^41,42 In the Neuroprotect trial by Ameloot et al.,⁴¹ 112 patients that remained comatose after resuscitation from OHCA were randomized to be maintained at a MAP of 65 mmHg or 85–100 mmHg over the first 36 hours. Targeting a higher MAP improved NIRS and TCD measurements of cerebral perfusion, but ultimately no changes in MRI-identified anoxic brain injury or 180-day neurological outcome were observed. In the COMACARE trial by Jakkula et al.,⁴² 123 comatose OHCA patients were randomized to be maintained at 65–75 mmHg or 80–100 mmHg over the first 36 hours. Targeting a higher MAP was not associated with improved cerebral oximetry values or injury biomarkers including neuron-specific enolase and S100B. Additionally, 30-day mortality and neurological outcome at 6-months were not significantly different.

Cerebral Autoregulation in VA-ECMO and eCPR

No studies have examined cerebral autoregulation specifically in eCPR patients, and only one study investigated cerebral autoregulation in a preclinical model of eCPR.⁴³ Impaired autoregulation has been reported for adults on cardiopulmonary bypass,⁴⁴ but similar data in adult peripheral VA-ECMO are unavailable. Cerebral autoregulation during clinical VA-ECMO has been studied more extensively in the pediatric setting.^24,45 Notably, impaired cerebral autoregulation with reduced endothelial nitric oxide production was observed in healthy newborn lambs on peripheral VA-ECMO, raising the question of whether the nonpulsatile flow alone is responsible for impaired cerebral vasoreactivity.^46–48 This view is supported by reports from the cardiopulmonary bypass literature,⁴⁴ but refuted by two small observational studies (n=9, 15) which reported normal cerebral autoregulation in adults with a continuous-flow left ventricular assist device.^49,50

CONTROLLED REPERFUSION: ROLE OF HYPOCAPNIA AND HYPEROXIA

Initiation of VA-ECMO can cause rapid changes in arterial partial pressure of oxygen (PaO₂) and PaCO₂, although there are currently no recommendations guiding specific targets or correction rates.^51,52

Hypocapnia

Despite disturbances in cerebral autoregulatory mechanisms, cerebrovascular reactivity to carbon dioxide remains intact within the first 24 hours post-arrest⁵³ and is even enhanced during nonpulsatile blood flow⁵⁴—each potentially contributing to further ischemic injury in cases of hypocapnia following eCPR initiation. At the cellular level, lowering carbon dioxide quickly during reperfusion is associated with cell death and enhanced production of reactive oxygen species in post-ischemic chick cardiomyocytes.⁵⁵ In the general cardiac arrest population, observational studies have associated post-ROSC hypocapnia with increased mortality⁵⁶ and poor neurological function at discharge.⁵⁷ In an observational study which included 135 patients on veno-venous ECMO, rapid correction of PaCO₂ at ECMO initiation was associated with intracranial bleeding.⁵⁸

Hyperoxia

Hyperoxia in eCPR patients should likely be avoided.⁵¹ Mechanistically, the adverse effects of hyperoxia are attributed to enhanced production of reactive oxygen species, vasoconstriction, and inflammation.^59–61 Although an association between hyperoxia and mortality has not been demonstrated uniformly across the literature for the general cardiac arrest population,^62,63 recent evidence specific to eCPR patients suggests harm. In a retrospective single-center analysis by Chang et al.⁶⁴ which included 291 eCPR patients, PaO₂ was measured within 24 hours of eCPR initiation. PaO₂ between 77 and 220 mmHg was determined to be optimal, with values >220 mmHg associated with poor survival and worse neurological outcomes. A more conservative threshold was reported in a retrospective cohort study of the Extracorporeal Life Support Organization (ELSO) registry conducted by Munshi et al.⁶⁵ which included 412 eCPR patients. Moderate hyperoxia (PaO₂ 101–300 mmHg) on arterial blood gas 24 hours after initiation of eCPR was associated with increased mortality. Finally, in a retrospective cohort study by Halter et al.⁶⁶ which included 66 patients with OHCA, PaO₂ was measured by arterial blood gas 30 minutes after initiation of eCPR. Hyperoxemia (PaO₂ ≥300 mmHg) at this timepoint was significantly associated with 28-day mortality.

EPINEPHRINE

Use of Epinephrine in Cardiac Arrest and eCPR

Standard dose epinephrine is administered broadly for patients with cardiac arrest based on evidence that it improves ROSC rate and short-term survival.^67,68 Despite these benefits, there are several reports of a negative association between epinephrine and neurological outcomes.^69–71 In the PARAMEDIC2 trial, the largest randomized controlled trial for epinephrine in cardiac arrest (n=8,014), the use of epinephrine improved 30-day survival but did not improve survival to hospital discharge with a favorable neurological outcome.⁷¹ This finding is attributed to the fact that severe neurological impairment was more common in patients that received epinephrine.⁷¹ It remains unknown whether this association is due to epinephrine toxicity or rather a manifestation of improved survival in patients with longer arrest downtime—and by extension longer duration of cerebral ischemia. The latter hypothesis is supported by data demonstrating that late administration of epinephrine (>10–15 minutes post-arrest) is associated with worse neurological outcomes.^72,73

The safety and efficacy of epinephrine use in the eCPR population remains uncertain. Although current resuscitation guidelines prioritize rapid defibrillation over epinephrine for patients with a shockable rhythm,^67,68 patients with refractory arrest ultimately receive multiple doses of epinephrine prior to hospital arrival and cannulation for eCPR.¹ Current ELSO guidelines state that vasopressors and inotropes may be weaned rapidly after establishing VA-ECMO flow of at least 3 l/min.⁵¹

Only one study has examined the efficacy of epinephrine in preclinical eCPR. In a swine model of ischemic refractory ventricular fibrillation, epinephrine failed to improve 4-hour survival when administered during mechanical chest compressions prior to eCPR.⁷⁴ Epinephrine was associated with a greater increase in arterial lactate levels prior to eCPR, although direct measurements of cerebral perfusion were not conducted.

Effect of Epinephrine on Cerebral Perfusion

At the level of the brain, the α-adrenergic effects of epinephrine may impair cerebral microvascular flow as a result of platelet activation and thrombosis⁷⁵ or vasoconstriction.⁷⁶ Several studies in swine have shown negative effects of epinephrine on cerebral perfusion during or after CPR.^76–78 In a manner dependent upon α₁-adrenoceptor activation, epinephrine impairs cerebral microcirculatory blood flow and increases brain tissue hypoxia.⁷⁷ These findings are contrasted by other animal studies which demonstrate improvements in cerebral perfusion after epinephrine administration.^26,79,80 Several variables may account for these discrepancies including arrest duration²⁶ and dosing (i.e. repeated boluses vs. infusion).^79,80

Recent preclinical data suggests a positive effect of continuous epinephrine infusion in states of low flow. In a swine study by Putzer et al.,²⁶ a low-flow state representing conventional CPR was simulated by low-flow VA-ECMO (30 ml/kg/min) which was initiated after eight minutes of untreated ventricular fibrillation. Extensive invasive and non-invasive hemodynamic monitoring was conducted during arrest, during low-flow VA-ECMO, and during low-flow VA-ECMO plus continuous epinephrine infusion titrated to a MAP of either 40 or 60 mmHg. In this study, epinephrine quickly increased CBF, CePP, and cerebral oxygenation. These results provide insightful feedback into the effect of epinephrine on cerebral perfusion in low-flow states such as conventional CPR; however, the role of epinephrine prior to full VA-ECMO support was not examined.

TITRATION OF MAP AND VA-ECMO FLOW IN ECPR

An optimal MAP goal is not well established for eCPR patients. The empirical benefit of maintaining high MAP for promoting cerebral perfusion must be weighed against the risk of LV distension leading to LV stasis, thrombosis, and pulmonary edema.^51,81 Furthermore, a theoretical risk for cerebral hypertension leading to intracranial hemorrhage is a significant concern in VA-ECMO patients due to the requirement for systemic anticoagulation.⁶ Current ELSO guidelines recommend maintaining MAP between 60 and 80 mmHg by vasopressor titration while maintaining circuit flow at 3–4 l/min.⁵¹ Serial lactate monitoring may be conducted to monitor organ perfusion and inform further support.⁵¹ Importantly, an unindividualized approach to MAP management does not account for interpatient variability with respect to cerebral autoregulation which may be right-shifted or entirely absent.^23,38 In addition to interpatient variability, perfusion requirements may vary between organs and over time. Currently utilized indicators of perfusion (e.g. lactate, pH, blood gases) are reflective of systemic physiology and cannot provide organ-specific information in real-time. Therefore, while the term “optimal MAP” generally refers to a particular MAP target that is associated with improved outcomes within a given study, this term does not account for potential differences in organ-specific demands.

Only one study has evaluated optimal MAP in a cohort of patients specific to eCPR.⁸² In this single-center retrospective observational study, MAP was monitored continually over 96 hours post-arrest in 253 adults treated with eCPR. Patients with an average MAP around 75 mmHg had a low probability of poor neurological outcome at discharge, while patients with an average MAP below 60 mmHg demonstrated a high probability of poor neurological outcome. The probability of poor neurological outcome tended to increase in patients with an average MAP greater than 75 mmHg. Based on these findings, targeting a MAP of 75 mmHg may be an optimal yet unindividualized approach.

A considerable amount of work has recently been conducted to determine optimal hemodynamic support in swine models of ischemic arrest with eCPR.^83–85,43 In the case where MAP was titrated to 65 mmHg with norepinephrine over a resuscitation period of 6 hours, eCPR using a standard flow rate (65–70 ml/kg/min) compared to a low flow rate (30–35 ml/kg/min) led to faster lactate clearance and enhanced carotid artery blood flow.⁸³ A nonsignificant trend towards less norepinephrine required to maintain MAP was observed. In the case where circuit flow rate was held constant at 70 ml/kg/min, the use of norepinephrine to titrate MAP to 80–85 mmHg compared to 65–70 mmHg did not improve lactate clearance, microcirculation, or fluid requirements.⁸⁴ When the circuit flow rate was held constant at 40 ml/kg/min, the use of epinephrine to titrate MAP to 80–90 mmHg compared to 65–75 mmHg transiently improved cerebral hemodynamics during eCPR, but significantly worsened cerebral hemodynamics after ROSC.⁴³ Finally, in a head-to-head comparison between vasopressin and norepinephrine in which MAP was titrated to 65 mmHg over a constant circuit flow rate of 65–70 ml/kg/min, vasopressin was associated with faster lactate clearance, less pulmonary edema, and less fluid administration.⁸⁵ Further work is needed to determine how neurological outcomes are affected by different hemodynamic support strategies in eCPR.

NEUROPROTECTION STRATEGIES IN ECPR

Clinical Approach to Neuroprotection in eCPR

In the absence of concrete guidelines for post-arrest management of eCPR patients, optimization of neurological outcome must involve application of basic physiologic principles which are further informed by available data primarily from the VA-ECMO or resuscitation literature. For patients that meet eCPR inclusion criteria, the primary focus should be on maintaining short door-to-cannulation time, followed by rapid restoration of organ perfusion and prompt reversal of the underlying cause of arrest. Hyperoxia and rapid correction of PaCO₂ at ECMO initiation can be avoided through conservative oxygen reintroduction with a gas blender and low initial sweep gas followed by gradual titration. Organ perfusion may be optimized by limiting vasopressors and using higher circuit flow rates—thus requiring larger venous cannulas. Targeting a MAP of 75 mmHg is a reasonable strategy until more evidence is available in support of a more individualized approach. To minimize risk of premature withdrawal of life-sustaining treatment, definitive multimodal neuroprognostication must be delayed until 72 hours post-ROSC (or 72 hours post-rewarming for hypothermia-treated patients) after eliminating residual sedation as a confounding variable.^67,86 Sedation elimination may be prolonged in the presence of hepatic and/or renal impairment which are common occurrences in eCPR patients.¹ Prevention of post-ROSC brain injury should involve daily multimodal neuromonitoring and frequent neurological assessment by highly trained neurointensivists. Currently available neuromonitoring techniques include but are not limited to clinical examination, cranial CT, NIRS, EEG, somatosensory evoked potentials, TCD, and serum biomarkers. MRI may be conducted only after ECMO decannulation due to incompatibility with the ECMO circuit, although portable ECMO-compatible MRI is currently under evaluation (NCT05469139).

Synergistic Strategies for Enhancing Neuroprotection in eCPR

The provision of the ECMO circuit presents unique opportunities to apply synergistic therapies and to investigate the efficacy of neuroprotective therapies when administered prior to ECMO cannulation. For example, targeted temperature management is reasonable and readily achievable with eCPR, although efficacy and optimal temperature remain uncertain for eCPR patients.⁸⁷ Furthermore, the ability to deliver agents directly to the arterial circulation may enhance the effectiveness of certain therapies when compared to intraosseous/intravenous delivery. Considering that no drugs exist which improve long-term neurological outcomes following cardiac arrest, further work is needed to develop neuroprotective agents. Agents which may have failed to improve ROSC in non-eCPR studies deserve further investigation for their neuroprotective potential in the pre-eCPR setting. Experimental strategies for pharmacological neuroprotection in cardiac arrest have been reviewed in detail elsewhere.^88,89

CURRENT GAPS IN KNOWLEDGE

A paucity of evidence-based recommendations is available to guide clinical management of eCPR patients. While it is acknowledged that eCPR patients are generally sicker than the general VA-ECMO patient population,⁹⁰ current eCPR management recommendations are largely based on expert consensus and extrapolation from studies which focused on non-eCPR cardiac arrest patients or patients that were placed on VA-ECMO for various indications. We found no studies which investigated regional CBF or cerebral autoregulation in human cohorts specific to eCPR. Studies which assessed cerebral perfusion in clinical eCPR were small and limited to NIRS.^16–20 Furthermore, no studies prospectively or retrospectively evaluated the relationship between epinephrine and neurological outcomes in eCPR cohorts. In summary, OHCA eCPR patients represent a complex scenario in which neuropathology associated with VA-ECMO may interact significantly with the consequences of a prolonged low-flow state that typically precedes eCPR. The field currently lacks a comprehensive understanding of how regional cerebral perfusion and cerebral autoregulation are temporally modified by factors such as pre-eCPR low-flow duration, vasopressors, and circuit flow rate. Elucidating these critical relationships may inform future strategies aimed at improving neurological outcomes in patients treated with lifesaving eCPR.