Critical Care Management of Patients After Cardiac Arrest: A Scientific Statement From the American Heart Association and Neurocritical Care Society

Post-CA brain injury is a common complication of CA and a major contributor to subsequent death and disability.⁶ Post-CA brain injury comprises both a primary injury caused by complete global brain ischemia and initial reperfusion and secondary injury occurring hours to days after ROSC. Secondary brain injury results from brain tissue ischemia and hypoxia, brain edema, elevated ICP, and seizures. Brain hyperoxia is also implicated as a cause of secondary post-CA brain injury and is discussed in the next section.

The overall incidence and degree of secondary brain hypoxia in patients after CA are unknown. Intracranial monitoring is not routinely performed, and indirect measures such as jugular venous bulb oxyhemoglobin saturation and near-infrared spectroscopy are limited because they measure oxyhemoglobin saturation rather than brain tissue oxygen tension. Key factors that affect brain oxygen delivery include cerebral perfusion pressure (CPP), partial pressure of arterial CO₂ (Paco₂), and arterial oxygen content. The discussion here focuses largely on perfusion and brain tissue oxygen delivery; systemic oxygen and CO₂ management are discussed later in the pulmonary section.

Although data are limited, preliminary evidence suggests that secondary brain tissue hypoxia occurs in the post-CA period.^21–23 Invasive brain tissue oxygen monitoring is increasingly used in patients with other causes of severe neurological injury such as brain trauma and subarachnoid hemorrhage. Using brain tissue oxygen as part of multimodal neurological monitoring paradigms to guide interventions after CA represents a promising new direction for individualized post-CA care. Additional approaches include monitoring blood oxyhemoglobin saturation from the jugular bulb; the jugular bulb blood oxygen content represents a surrogate for adequacy of brain oxygen delivery.²⁴ The optimal targets for jugular bulb oxygen saturation and parenchymal brain tissue oxygen are unknown in patients after CA and represent an opportunity for future research.

The literature on blood pressure targets during post-CA care in the intensive care unit (ICU) has demonstrated that low mean arterial pressure (MAP) is associated with worse outcomes,^25–28 although higher MAPs have been associated with better neurological outcomes.²⁹ Multiple retrospective and prospective observational studies have reported that MAPs >65 mm Hg are associated with higher survival rates and better neurological outcomes.^25–29 Because of the observational nature of these studies, it is difficult to determine whether these associations are driven by spontaneous blood pressures or by blood pressure targets achieved with vasopressors. Recent RCTs have compared MAP targets.^19,30,31 A multicenter trial randomizing patients to 65 to 75 mm Hg versus 80 to 100 mm Hg found that it was feasible to achieve these targets without worsened safety outcomes.³¹ This study was not powered to detect differences in survival or neurological outcome and used a surrogate biomarker as the primary outcome measure (neuron-specific enolase levels at 48 hours after arrest). A second trial randomized patients to an MAP of >65 mm Hg versus 85 to 100 mm Hg. Although the study did not detect any difference in efficacy end points, it further demonstrated feasibility and absence of harm when targeting higher MAPs.³⁰ A pooled analysis³² of these 2 studies^30,31 also showed that targeting an MAP between 80 and 100 mm Hg in patients with acute myocardial infarction and shock was associated with attenuated biomarkers of myocardial injury and no increase in rearrest. This result provides evidence that higher MAP targets are not associated with more adverse cardiac outcomes. A third trial randomized patients to an MAP of 77 mm Hg versus 63 mm Hg.¹⁹ Patients were also assigned to 1 of 2 oxygen targets. There was no difference in multiple outcome measures, including the primary outcome of composite of death resulting from any cause or hospital discharge with a Cerebral Performance Category of 3 or 4 within 90 days. There was also no difference in adverse events. The results of this study are difficult to generalize given the complex study intervention, the narrow patient population (>80% witnessed shockable rhythm with bystander CPR), the relatively small separation in blood pressure between the groups (10.7 mmHg [95% CI, 10.0–11.4]), and an opposite trend in treatment effect at the 2 participating institutions. It is also important to note that, in all of these studies, interventions to achieve blood pressure targets were not initiated until patients arrived in the ICU and therefore do not reflect potential benefits or harms of specific MAP targets in the first minutes to hours after ROSC.

Of additional concern in patients after CA is that compensatory cerebrovascular autoregulatory responses may be impaired. The range of MAPs in which cerebral blood flow is constant decreases and may even be lost altogether.³³ The cerebrovascular autoregulation curve may also be right shifted, necessitating higher MAPs to maintain adequate CPP.^22,34 Although there is variability between individuals and testing modalities, an MAP of 65 mm Hg is insufficient to provide adequate cerebral oxygen delivery in many patients.²² Moreover, because ICP is not routinely monitored in patients after CA, CPP is usually estimated, often assuming a normal ICP of 5 to 15 mm Hg. However, many patients after CA develop brain edema, which can result in elevated ICP (see the Elevated ICP and Brain Edema section). Taken together, these factors put patients at risk for inadequate brain perfusion at an MAP target of 65 mm Hg and make it difficult to empirically select a target MAP. A higher MAP target (eg, MAP >80 mm Hg) will increase the probability that CPP >60 mm Hg is achieved, assuming high normal or slightly elevated ICP. Alternatively, in patients who lack intact autoregulation, higher blood pressures may contribute to hyperemia and worsen intracranial hypertension. Thus, individual pathophysiology and the complex interplay between systemic and neurological physiology are important to consider (Figure 1). The limited data on minimal thresholds for cerebrovascular autoregulation, combined with the safety data supporting a higher MAP, led to the support for a higher MAP target.

A more nuanced approach to identify individualized MAP targets for patients after CA may be through the assessment of cerebrovascular autoregulation to identify the threshold MAP below which autoregulation is lost (lower limit of autoregulation) and the optimal MAP (MAP_OPT) at which cerebral perfusion is the least influenced by MAP fluctuations (Table 1). Several studies have offered preliminary evidence that individual determination of MAP_OPT in comatose post-CA survivors is a promising strategy to optimize cerebral perfusion and potentially minimize secondary brain injury.^22,26,35,36

Advances in invasive and noninvasive multimodal neuromonitoring may allow future approaches that optimize brain perfusion and oxygenation while balancing the potential concomitant cardiac and pulmonary effects of such interventions. For example, a lower MAP target may be beneficial in a patient after CA whose primary post-CA syndrome pathology is dominated by ongoing coronary ischemia and heart failure refractory to inotropic support. However, a higher MAP target may be beneficial in a patient after CA whose primary post-CA syndrome pathology is dominated by intracranial hypertension and brain hypoxia. Multimodal neuromonitoring may enable greater optimization of both neurological and systemic targets.

The role of neuromonitoring in titrating post-CA care to individuals is an important area for further research. Multimodal monitoring refers broadly to the evaluation and synthesis of data from multiple modalities, which include both invasive and noninvasive modalities, electroencephalography (EEG) and imaging, and clinical examination. How to combine these tools to synthesize and respond to this multimodal data to individualize care is an important focus for further research. Identifying mechanisms of autoregulation failure, brain hypoxia, and inadequate brain perfusion or brain hyperemia and developing strategies to treat these injuries in the days after CA are also important research opportunities.

Brain edema (both cytotoxic and vasogenic) is a common complication from hypoxic/ischemic brain injury and is associated with poor outcomes.^37–39 Brain edema is usually diagnosed through neuroimaging, and its presence is most studied as a prognostic imaging biomarker.^5,40 There is limited research investigating interventions focused on preventing and treating brain edema in patients with CA. Because the previously referenced NCS guidelines address general approaches to medical management of cerebral edema, the topic is not covered here. In patients after CA, diffuse brain edema is believed to be the primary mechanism that causes elevated ICP.

Elevated ICP is suspected when there is evidence of diffuse brain edema on imaging, when there are signs of elevated ICP on examination (eg, loss of pupillary reflex), or through invasive (direct) measures (eg, intracranial monitoring) or noninvasive (indirect) measures (eg, optic nerve sheath diameter, transcranial Doppler).^22,41–43 Early preliminary studies suggest that ICP elevations >20 mm Hg occur in a subset of patients, and they may occur during both hypothermia and rewarming and are associated with poor outcomes.^23,42,43 There is no consensus on what threshold constitutes an abnormal ICP after CA, and data from traumatic brain injury are often extrapolated to consider an ICP >22 mm Hg abnormal.⁴⁴

Although brain edema is a complication of the initial hypoxic/ischemic and ischemic/reperfusion injury, it can also be exacerbated or caused by other insults that are modifiable and potentially preventable.⁴⁵ Limited data support interventions to treat cerebral edema and elevated ICP after CA; thus, general approaches focus on using standard ICP-lowering therapies.^15,46

A promising area of research in CA is investigating how noninvasive measures that provide surrogate information about ICP such as optic nerve sheath diameter, transcranial Dopplers, and quantitative pupillometry may be used to guide therapies.^47–49 The longitudinal evolution of brain edema on imaging after CA and its relationship with ICP and compliance are other important areas for further investigation.⁵⁰

Seizures can develop early, during the first 24 hours after ROSC, or later, especially during or after rewarming in patients treated with hypothermic temperature control.^53,54,58,63 Limited guidance exists on how long and how often EEG monitoring should be performed after CA.⁶⁴ Continuous EEG (cEEG) monitoring has higher sensitivity than intermittent EEG monitoring; however, there are limited data on whether seizure treatment guided by continuous compared with intermittent EEG affects outcomes or is cost-effective.^65–67 cEEG is frequently used to guide treatment in patients with SE.⁶⁴ Cost and availability of technicians and clinicians with expertise in EEG interpretation are barriers to cEEG implementation, and studies have investigated the use of limited EEG montages, rapid EEG, quantitative EEG panel review (eg, spectrograms or amplitude-integrated EEG), and automated computational analysis of EEG signals.^63,66–68

Similar to seizures, many electrographic findings on the IIC are associated with poor outcomes; however, it is unknown whether IIC patterns represent an epiphenomena or a cause of brain injury that warrants treatment.^{56,57,69–71} With increased availability of cEEG monitoring, identification of EEG findings on the IIC became more common; however, the clinical relevance and management remain understudied. Summarization of IIC patterns under categories used in neuroprognostication such as malignant or highly malignant is imprecise, and these categories may inadvertently influence treatment decision-making despite not being developed for that purpose. Sporadic discharges (ie, nonrhythmic and nonperiodic spikes, polyspikes, or sharp waves) are common, but their association with neurological outcome is variable.^60,72

Treatment of post-CA seizures and SE lacks guidance from large-scale RCTs.^72–74 The overall body of literature is limited by retrospective design, early withdrawal of life-sustaining therapy, and heterogeneous ASM and anesthetic drug combinations, dosing, and treatment durations.⁷³ Many studies give limited information on how seizures or SE was managed.^55,71,75,76 Despite these limitations, a growing number of case series and 2 recent studies evaluating stepwise treatment protocols for recurrent seizures and/or refractory SE after CA with ASMs and anesthetic agents indicate that neurological recovery may occur when intensive seizure control is pursued.^{18,58,72–74,77,78} When seizures or SE is treated, studies commonly use initial regimens consisting of benzodiazepines, followed by valproic acid, levetiracetam, or phenytoin (ie, first- and second-line treatments).^72,73,79 Third-line treatments often consist of continuous anesthetic infusions such as propofol, midazolam, or ketamine. The side-effect profile of different agents is considered when an ASM is chosen, with additional consideration given to the potential confounding of the neurological examination by many classes of ASMs.^69,71,72 Barbiturates can be effective for seizures or SE treatment after CA as second- or third-line agents or to facilitate weaning of anesthetic agents in case of seizure recurrence during weaning.⁸⁰ However, barbiturates can confound neurological examination for several days, given their long half-lives and sedative effects. In the setting of SE, induction of a burst-suppression pattern on EEG with anesthetics is a common therapeutic target, but the efficacy and optimal depth and duration have not been established.⁶⁶ In the absence of EEG abnormalities, ASMs are not routinely given prophylactically, although additional studies are needed.⁷²

Clinical myoclonus after CA is characterized by repetitive, generalized, focal, or multifocal myoclonic movements, which may or may not have an electroclinical correlate on EEG.⁸¹ Clinical myoclonus for extended periods in the context of CA is often referred to in literature as myoclonic SE or status myoclonus (commonly defined as lasting >30 minutes). Association of clinical myoclonus with epileptiform activity on scalp EEG is not always present; however, when present, it is an electroclinical seizure, by definition, and can occur independently of electrographic seizures or electrographic SE. Therefore, the distinction among myoclonus observed exclusively on clinical examination (ie, clinical myoclonus), myoclonus in association with epileptiform activity (ie, electroclinical myoclonus), and myoclonus in association with electrographic seizures or electrographic SE (ie, electrographic seizure or electrographic SE with electroclinical myoclonus) may be relevant in management (Table 2).^82,83 Given the inconsistency in previous terminology and definitions, we define this terminology to differentiate the subtypes of myoclonus commonly seen after CA and to avoid the terms myoclonic SE and status myoclonus.

The certainty of evidence is limited concerning the treatment of clinical myoclonus, electroclinical myoclonus, or electrographic seizure or SE with electroclinical myoclonus. Although myoclonus has been historically associated with poor outcome, neurological recovery has been reported more recently in patients with post-CA SE associated with electroclinical myoclonus who undergo aggressive therapy with ASM or anesthetics.^78,84,85 Aggressive treatment of clinical myoclonus without an electroclinical correlate has limited efficacy and can be associated with side effects and impairment of the neurological examination.⁸²

In patients with EEG findings on the IIC, treatment is highly variable, and its impact on patient outcomes is poorly understood.⁸⁶ Studies in other patient populations with acute brain injury indicate that persistent and high-frequency IIC patterns might cause secondary brain injury, and treatment with ASMs has been advocated.⁸⁷ In comatose CA survivors, stepwise treatment with ASMs targeting suppression of IIC patterns was not associated with improved functional outcome at 3 months compared with the group randomized to no ASM treatment in a large RCT.¹⁸ However, this study involved predominantly patients with generalized periodic discharges at 0.5 to 2.5 Hz and did not include subgroup analyses to discern whether different treatment strategies or frequency thresholds, beyond those that evolved or qualified as SE, might result in different outcomes. Thus, treatment of IIC patterns is individualized and takes into consideration other EEG background features indicative of potential for neurological recovery such as a continuous background or EEG reactivity.^72,74 Diagnosis and management of IIC patterns guided by intracranial multimodality monitoring or advanced neuroimaging (eg, computed tomography or magnetic resonance imaging perfusion and positron emission tomography) are under investigation, and these methods may offer additional insights into physiology and may help determine the impact of treatment.⁸⁸

Future research may focus on the use of advanced seizure monitoring techniques such as intracranial EEG and advanced brain imaging (eg, computed tomography perfusion, magnetic resonance imaging perfusion, or positron emission tomography) to offer additional insights into diagnosis and to help guide treatment, especially when IIC patterns are seen on scalp EEG. The approach to and impact of treating post-CA SE are important areas of future research. Which specific IIC findings merit treatment with ASMs and the impact of treatment on outcome are other important areas for further research.

As discussed in the section, “Neurological Management in the Intensive Care Unit: Brain Oxygenation, Perfusion, Edema, and ICP, ” the optimal blood pressure target to improve CA outcomes and how to achieve said targets have been of research interest over the past decade. Recent studies have focused on the association between blood pressure and neurological function, specifically evaluating surrogates of brain perfusion and oxygenation, under the premise that using physiological data points that assess cerebral oxygen delivery and perfusion may help guide blood pressure thresholds and improve neurological outcomes.²² Individual patient characteristics, including the degree of myocardial dysfunction, underlying cause of shock, baseline prearrest blood pressure,¹⁰⁵ and concomitant brain injury, are evaluated when post-CA MAP targets are set.

Patients with post-CA syndrome have hypotension and shock 50% to 70% of the time,¹⁰⁶ which complicates acute hemodynamic management approaches and strategies to maintain systemic, coronary, and cerebral perfusion. Shock may be due to a combination of cardiogenic, vasoplegic, hypovolemic, and septic causes, and it can often be difficult to determine whether the shock physiology was precipitated by the arrest or was present before the arrest. Serum lactate and central venous oxygen saturation are often used to guide resuscitation in shock states, but these markers may be less reliable in patients with the mixed shock after CA.^107,108 Persistent poor perfusion, shivering, sepsis, alterations in metabolism during hypothermic temperature control, and liver dysfunction are all causes of persistently elevated lactate in patients after CA. Thus, when other markers of perfusion have improved, aggressive treatment of clinical scenarios contributing to persistently elevated lactate occurs.

Echocardiography is an important tool to evaluate and manage patients with post-CA myocardial dysfunction.¹⁰⁹ Echocardiography is noninvasive and provides valuable information about ventricular function, cardiac output, fluid status, and underlying causes of CA.¹¹⁰ The literature on patients after CA shows that hemodynamic monitoring with echocardiogram is not robust, and there are many caveats to its use. Echocardiography can also assist in determining the most appropriate pharmacological treatment of shock. Many patients after CA have myocardial stunning that can be profound within the first several hours after ROSC. It is important to consider inotropy, vasodilatory state, and volume status when choosing the correct pharmacological treatment for hypotension and shock after CA. Invasive monitoring with a pulmonary artery catheter may also provide helpful information, particularly among those with severe post-CA cardiogenic shock requiring mechanical circulatory support.^111,112 Alternatively, noninvasive measurement of cardiac output is available, although its usefulness in post-CA shock is less certain.

The use of mechanical circulatory support continues to expand in patients after ROSC to maintain organ perfusion.¹¹³ A host of devices are available, including intra-aortic balloon pumps, percutaneous ventricular assist devices, and extracorporeal membrane oxygenation. No robust randomized trials exist to guide patient and device selection. Two trials evaluating the use of extracorporeal membrane oxygenation specifically in refractory arrest were recently published. The ARREST trial (Advanced Reperfusion Strategies for Refractory Cardiac Arrest) showed profound benefit in the extracorporeal membrane oxygenation–facilitated resuscitation arm; however, the trial included only 30 participants and was limited to those with shockable rhythm.¹¹⁴ The Prague out-of-hospital CA study was a large study (256 participants) also designed to assess the benefits of extracorporeal membrane oxygenation over standard advanced cardiac life support in refractory arrest; however, this trial failed to show benefit, possibly because of significant crossover between groups.¹¹⁵ In the absence of definitive data, it is important for the team caring for post-CA patients to understand individual device characteristics and to select a specific device or no device, as dictated by the patient’s unique physiology.

Important areas for future research include evaluation of which vasoactive drugs, inotropes, and fluid strategies are appropriate to attain these targets and evaluation of cardiac and neuromonitoring modalities to guide individualized targets. In addition, evaluating systems of care for how and when to transfer patients for advanced mechanical circulatory support is an important area for further research.

There is reasonable consensus that early coronary angiography (performed within the first 6 hours) is safe and beneficial for post-CA care, although the majority of such data are from nonrandomized cohort studies.^116–118 Patients with ST-segment elevation on their ECG after resuscitation have an acutely occluded coronary at early coronary angiography nearly 80% of the time,¹¹⁹ which is similar to those with ST-segment–elevation myocardial infarction not complicated by out-of-hospital CA. The current AHA guidelines recommend early coronary angiography for patients with ST-segment elevation on their ECG after CA, even if comatose with an unknown neurological prognosis.¹²⁰ Nonrandomized registry reports suggest that both short-term and longer-term outcomes are improved in patients with CA and ST-segment elevation undergoing early coronary angiography and intervention.^121–125

There is less agreement about the utility of early coronary angiography after out-of-hospital CA in patients without ST-segment elevation. This is an important issue because the majority (>70%) of all resuscitated individuals with out-of-hospital CA do not have ST-segment elevation.¹¹⁹ A substantial amount of literature on nonrandomized cohort studies in post-CA patients without ST-segment elevation exists, as it does for those with ST-segment elevation. At least 10 000 patients have been included in such before-and-after cohort reports.^{116,121–128} The overall results suggest that early coronary angiography is both safe and beneficial, including for improved survival and functional neurological status, among post-CA patients without ST-segment elevation. However, the proportion of post-CA patients without ST-segment elevation having an acutely occluded coronary culprit vessel is ≈20% to 30%.^{119,126–128} This lower proportion has led to a less emphatic AHA recommendation.¹²⁰

Five randomized studies evaluating the value of early coronary angiography in those without ST-segment elevation have been published, including 2 in 2019 and 1 in 2020. The first was the limited ARREST pilot study of 40 patients.¹²⁹ As anticipated, no outcome differences were found, and the main conclusion was that a larger, definitive RCT to examine this issue was feasible, acceptable, and safe. Of note, 15 of 32 (47%) undergoing coronary angiography within 3 hours of admission had a culprit lesion found and revascularized. The COACT trial (Coronary Angiography After Cardiac Arrest)¹³⁰ randomized 552 patients without an ST-segment–elevation myocardial infarction to either immediate or delayed coronary angiography. No difference in 90-day survival was seen between the groups. However, acute unstable coronary lesions were found in only 14% of patients undergoing immediate coronary angiography. The DISCO pilot study (Direct or Subacute Coronary Angiography in Out-of-Hospital Cardiac Arrest), a small pilot randomized trial,¹³¹ did not report >24-hour survival but did find a culprit lesion in 37% of patients undergoing early coronary angiography. The PEARL study (Pilot Randomized Clinical Trial of Early Coronary Angiography Versus No Early Coronary Angiography for Post–Cardiac Arrest Patients Without ST-Segment Elevation), although underpowered to show any outcome differences, found a culprit vessel in 47% of patients undergoing early coronary angiography, including 1 of every 7 patients demonstrating an acutely occluded culprit vessel.¹³² The true prevalence of coronary culprit lesions in those without ST-segment elevation is not yet resolved. The effect of early coronary angiography and percutaneous coronary intervention on outcome will probably be directly related to the prevalence of lesions. A fifth RCT, the Angiography After Out-of-Hospital Cardiac Arrest Without ST-Segment Elevation, was published in 2021.¹³³ Randomizing a total of 530 patients to early or delayed/selective coronary angiography, those investigators found no difference in 30-day mortality. They reported a culprit vessel incidence of 38% in those undergoing immediate coronary angiography but did not provide any data on acutely occluded culprit lesions. Last, the EMERGE trial (Emergency vs Delayed Coronary Angiogram in Survivors of Out-of-Hospital Cardiac Arrest) was published in 2022.¹³⁴ EMERGE randomized 279 patients to emergency or delayed coronary angiography (141 and 138 patients, respectively) and again showed no difference in their primary end point of 180-day survival with Cerebral Performance Category of 1 or 2. At least 2 more trials are evaluating the role of emergent coronary angiography patients in this population (NCT02309151 and NCT02641626).

Early coronary angiography for patients resuscitated after out-of-hospital CA is safe and reasonable,¹³⁵ and although it did not improve outcomes in these RCTs, it can identify potential coronary culprit lesions that could benefit from timely reperfusion and revascularization. The incidence of such acutely occluded culprit lesions has varied by study, and the final answer remains to be determined. This is an important area for future research. On the basis of existing data, survival may not be the only important end point. Timely reperfusion of an acutely occluded culprit coronary artery is also key because it salvages at-risk myocardium and may prevent rearrest and the subsequent development of heart failure in survivors. Quality of life is improved with preserving left ventricular function and ultimately avoiding heart failure, with its life-altering limitations.

During CA, the brain tissue oxygen tension falls rapidly to 0.¹⁴⁰ At the onset of reperfusion, there is commonly a hyperemic phase in which tissue oxygenation becomes supranormal. Administration of high concentrations of oxygen during this phase may expose ischemic brain tissue to significant hyperoxemia, potentially exacerbating the formation of reactive oxygen species and increasing postischemic oxidative injury and cellular death.⁴⁵ Human studies that address this topic are limited. Postischemic neurons are likely to be most prone to any harmful effects of hypoxemia immediately after ROSC. In out-of-hospital CA, this implies the need for a titrated oxygen therapy intervention to be started before arrival to the hospital. The data on prehospital oxygen titration are beyond the scope of this document, although several studies have addressed prehospital oxygen titration, and the topic is covered in the 2020 AHA guidelines for CPR and emergency cardiovascular care.⁵ Concerns about the physiological basis for harm from hypoxemia drive the opinion that titration of Fio₂ should not be attempted if a reliable measure of oxygenation (pulse oximetry or arterial partial pressure of oxygen) is not available. It can be difficult to obtain an accurate pulse oximetry measure in many patients after CA, especially in the prehospital setting but also in some patients in severe shock in the ICU, although it should still be done when possible. Investigators have also reported that occult hypoxemia (normal pulse oximeter oxygen saturation [Spo₂] when Pao₂ is abnormally low) is more common in patients with darker skin pigmentation, supporting extra caution when oxygen is titrated down on the basis of Spo₂ in patients with dark skin.^141,142 Arterial blood gas measurement can be helpful when reliable pulse oximetry is difficult to obtain.

Three RCTs have compared different oxygen value targets in patients admitted to an ICU after CA.^143–145 In the first, there was no significant difference between the groups in neuron-specific enolase values or secondary clinical outcomes. In the second, of the 965 enrolled patients in the ICU, a subgroup of 166 had suspected hypoxic/ischemic encephalopathy (ie, post-CA brain injury). Post hoc analyses of this subgroup showed higher mortality at day 180 in the usual-oxygen group, and there was a nonsignificant trend toward unfavorable outcome on the Extended Glasgow Outcome Scale in the usual-oxygen group. However, a secondary analysis of this subgroup, which adjusted for baselines differences, showed no significant differences in outcomes.¹⁴⁶ A third RCT randomized patients to Pao₂ of 68 to 75 mm Hg or 98 to 105 mm Hg.¹⁴⁵ There was no difference in the primary outcome, which was a composite of death resulting from any cause or hospital discharge with a Cerebral Performance Category of 3 or 4 within 90 days. There was also no difference in adverse events. The results of this study are difficult to generalize given the complex study intervention, the narrow patient population (>80% witnessed shockable rhythm with bystander CPR), the relatively small separation in Pao₂ goals between the groups (both of which were well below levels thought to cause oxidative injury), and an opposite trend in treatment effect at the 2 participating institutions. It is also important to note that, in all of these studies, interventions to achieve Pao₂ targets were not initiated until patients arrived in the ICU and therefore do not reflect potential benefits or harms of specific Pao₂ targets in the first minutes to hours after ROSC.

Many observational studies have compared the outcome of patients after CA who have been exposed to a period of hypoxemia or hyperoxemia with the outcome of those who have been normoxemic throughout. All of these studies are significantly limited by residual confounding due to lack of sufficient information or adjustment for CA characteristics and other patient factors, as well as a lack of clarity on the exposure. Definitions of hyperoxemia have also varied. In addition, these studies are comparing partial pressure of oxygen or oxygen saturation values rather than oxygenation strategies. Especially in the case of hypoxemia, these values are likely determined by patient severity of illness and not by management strategy. Even among studies that adjusted for potential confounders, results are inconsistent.^147–152 Despite the lack of clarity on this question, because hyperoxemia has not been found to be beneficial and some studies suggest harm, the available evidence supports a strategy of titrating the Fio₂ down when oxygenation is measurable and adequate.

There are conflicting physiological rationales for targeting hypercapnia and hypocapnia in the post-CA setting. CO₂ is known to affect the cerebral vasculature, with hypercapnia leading to vasodilation and hypocapnia to vasoconstriction. Hypocapnia induced by hyperventilation can cause cerebral ischemia.^153,154 One RCT found higher cerebral oxygenation in patients treated with mild hypercapnia compared with patients with low-normal Paco₂ targets.¹⁵⁵ However, hypercarbia could be harmful in patients with elevated ICP (Figure 1). Both existing RCTs comparing different ventilation strategies (targeting normocapnia or mild to moderate hypercapnia) in comatose patients after CA excluded those with suspected or confirmed elevation in ICP or cerebral edema and found no significant difference in clinical outcomes between groups, but they were not powered for clinical outcomes.^143,155 Observational studies have essentially been split among finding benefit,^151,156 no difference,^152,157 and harm^149,158 from hypercapnia. Limited observational data on hypocapnia support either harm or no benefit.^{149,151,157,158} It is possible that there are complex interactions among ICP, CPP, Paco₂, and the presence or absence of cerebral autoregulation. In addition, how Paco₂ targets should be adjusted for in patients with chronic hypercapnia is unclear, but there is no evidence to suggest that decreasing Paco₂ below someone’s baseline is beneficial in the post-CA setting. All these considerations make it difficult to determine the optimal Paco₂ for any individual patient after CA. Arterial blood gas analysis remains the most accurate way to determine the PaCO₂.

Given the paucity of research in this area and the impact of oxygenation and ventilation on all critical organs, this is an important area for additional research. Oxygenation targets at different time points after arrest, approaches to oxygen titration and ventilation, and the use of systemic and cerebral monitoring to guide these approaches are all important research opportunities.

Hemoglobin is a key determinant of tissue oxygen delivery. Thus, anemia may be particularly harmful in postanoxic brain injury. Anemia is frequent in critically ill patients. Among CA survivors, a strong linear relationship between hemoglobin and cerebral brain oxygen saturation was observed, suggesting that anemia may play a role in the development of secondary brain injury.¹⁵⁹

Several retrospective studies have found an association between hemoglobin values and outcome among survivors of CA.^160–163 However, results from these studies cannot be compared or generalized because of significant methodological variability across studies, including when hemoglobin values were assessed; patient inclusion criteria, including location of CA (in hospital versus out of hospital or both); thresholds associated with poor outcome; and transfusion practices. Indeed, even any benefits of red blood cell transfusion (RBCT) to treat anemia are balanced by the risk of a transfusion-related complications. RBCT has been associated with increased mortality and a higher occurrence of organ dysfunction in observational studies that included general critically ill patients.¹⁶⁴ Additional transfusion triggers, either systemic (ie, oxygen saturation or lactate levels)¹⁶⁵ or cerebral (ie, noninvasive oxygen saturation),¹⁵⁹ have been investigated, although data supporting their role in this setting are limited.

Given the lack of post-CA–specific data, transfusion strategies follow general guidelines for critically ill patients.¹⁶⁶ An individualized approach for RBCT based on the clinical status of the patients rather than on the isolated hemoglobin level is prudent, and all interventions should balance the risks and benefits.

Further understanding of the role of anemia and RBCT on neurological and cardiac function in patients after CA is important. Data from physiological and prospective studies could help to design interventional studies and/or to individualize RBCT in this setting. A well-designed clinical trial evaluating hemoglobin targets and transfusion practices specifically for patients resuscitated from CA would be of important value.

Ensuring adequate caloric intake in critically ill patients is a challenge, and multiorgan dysfunction, infection, hyperthermia, and preexisting malnutrition can further compromise the nutritional status of patients after CA.^175,176 Malnutrition results in a significant reduction of muscle mass and strength and in the development of ICU-acquired weakness, which are independent predictors of ICU mortality.¹⁷⁷ Nutritional support refers to the provision of calories, proteins, electrolytes, vitamins, minerals, trace elements, and fluid. There are no dedicated clinical trials focusing on nutritional support in patients after CA in the ICU. Details related to the daily nutritional goals and formulations of nourishment are beyond the scope of this scientific statement.

Early (within 24–48 hours of ICU admission) initiation of enteral nutrition (EN) is relatively low cost and is important to maintain the intestinal barrier integrity and to provide important nutrient and caloric support^178,179; however, because intestinal ischemic injury and concomitant shock are frequent after CA^180,181 and the use of sedatives and analgesic drugs during temperature control might reduce intestinal motility,¹⁸² the tolerance of EN could be limited. One study showed that patients undergoing temperature control after CA were able to tolerate a substantial proportion of their EN, and the routine use of prokinetic drugs may increase the success of EN delivery.¹⁸³ Moreover, in 2 other studies, early EN after CA during temperature control appeared safe and was even associated with better neurological outcomes.^184,185 However, a large published study on early EN versus early parenteral nutrition in patients with shock¹⁸⁶ suggested a possible harmful effect of the enteral route, so EN regimens should be carefully evaluated in patients after CA who have hemodynamic impairment. A trophic regimen (ie, 10–20 mL/h) could be the initial prescribed dose, which could be increased if well tolerated. Randomized studies in the general critical care population showed no clear differences between short-term, early EN and early parenteral nutrition.^182,186

Randomized trials on nutrition tolerance and effects on outcome could be relevant in this field. Calculating energy expenditure and caloric needs remains to be further evaluated in CA survivors, especially if they were treated with temperature control to a hypothermic target temperature. The role of vitamins and trace elements on organ function and outcome in these patients needs to be further evaluated.

Critically ill patients are at high risk of mucosal damage in the upper gastrointestinal tract in the acute phase of their disease. This damage can potentially result in deep and focal mucosal damage penetrating the submucosa, with a high risk for gastrointestinal bleeding.¹⁸⁷ Many risk factors¹⁸⁸ are common in CA survivors. Clinically relevant gastrointestinal bleeding can cause hemodynamic instability and increase the need for RBCTs, potentially contributing to increased ICU stay and mortality.¹⁸⁷ Intestinal injury has been described in patients after CA, could potentially be one determinant for endotoxemia, and could also worsen post-CA shock and organ failure.^180,181

Acid-suppressive agents such as histamine H2-receptor antagonists and proton pump inhibitors are used largely to prevent mucosal damage and to reduce the risk of gastrointestinal bleeding. In addition, EN may be protective against stress-related mucosal bleeding in high-risk patients.¹⁸⁹ However, whether specific post-CA gastric injury is present and which is the most effective method for ulcer prophylaxis between early enteric feeding and administration of a histamine H2-receptor antagonist or proton pump inhibitor remain unknown. One specific consideration in post-CA patients that remains to be further demonstrated is whether some proton pump inhibitors such as omeprazole could attenuate the antiplatelet effect of clopidogrel,¹⁹⁰ an oral antiplatelet drug that is frequently administered in patients after CA who are undergoing percutaneous coronary angioplasty.

Further studies are needed to define the appropriateness and timing of stress ulcer prophylaxis and the role for pharmacological prophylaxis compared with, and in the setting of, EN. Prospective assessment of the incidence of gastrointestinal bleeding is critically important.

Volume resuscitation can influence outcomes in septic shock and ARDS. There is conflicting evidence about whether a goal-directed approach for fluid administration is superior to standard volume loading.^202–205 There is also evidence suggesting that excessive volume resuscitation is harmful and that relative volume restriction is safe after the initial resuscitation phase.^206–208 After initial volume resuscitation, additional fluid administration should be guided by markers of end-organ perfusion, including central venous oxygen saturation, pulmonary arterial catheter, or point-of-care ultrasound.

It is currently unclear whether volume balance should be managed differently in patients after CA compared with other ICU populations because of the lack of evidence specific to the post-CA phase. Prehospital cold saline during induction of hypothermic temperature control was associated with no benefit and with increased adverse events.²⁰⁹ These findings cannot be extrapolated to the inpatient setting because a majority of the patients did not have a secured airway, and shivering resulting from cold fluid was inconsistently treated. Because of the overwhelming heterogeneity in patients with CA, a single approach to volume resuscitation cannot be applied to all patients.

There are no trials evaluating balanced solutions compared with normal saline specifically in the post-CA population. However, several trials in other critically ill cohorts are informative.^210–212 The largest RCT, the SMART trial (Isotonic Solutions and Major Adverse Renal Events Trial), demonstrated a lower rate of the composite outcome of death, renal replacement therapy (RRT), or persistent renal dysfunction in patients treated with balanced solutions compared with normal saline.²¹³ Numerous systematic reviews evaluated the potential benefit of balanced crystalloid solutions.^214–219 These analyses had mixed results, which may be due to the end points chosen (eg, a composite end point in the SMART trial), the trial designs permitted (ie, prospective versus observational), and the patient populations included.

A single approach to fluid selection may not apply to patients with post-CA syndrome because of the associated neurological injury. Although data exist describing the potentially harmful effects of hypotonic solutions on cerebral edema, data comparing individual balanced solutions are limited. Clinicians must consider the fact that balanced solutions are slightly more hypotonic than normal saline and could aggravate cerebral edema. Additional prospective studies are needed to evaluate the effect of hyperchloremia on kidney failure and mortality.

Few data describe sodium bicarbonate use specifically in the post-CA period. Most studies investigate the role of sodium bicarbonate for metabolic acidosis in heterogeneous critically ill populations, and the available data do not support routine administration of sodium bicarbonate for metabolic acidosis.²²⁰ Two recent studies, however, suggest benefit with bicarbonate in patients with extreme acidosis and acute kidney injury (AKI).^221,222 The lack of clear effect in the overall population could be due to the multitude of causes for metabolic acidosis and the possible benefit for some causes (ie, AKI) but not others (ie, lactic acidosis). Further prospective research is needed that is specific to patients with post-CA syndrome or focuses on subgroups most likely to have physiological benefit.

RRT is used to support critically ill patients who are in renal failure. Receipt of RRT is generally a marker of higher illness severity and therefore is associated with worse outcomes. A notable concern for patients after CA is the potential for dialysis disequilibrium syndrome, especially in intermittent hemodialysis, in which RRT can cause fluid shifts and rapid changes in solute concentration, resulting in increased cerebral edema and seizures.

AKI after CA is common (37% of patients) and sometimes leads to RRT (33% of patients with AKI). When patients with AKI are supported aggressively, renal recovery is possible, and most patients do not require long-term RRT.^223–225 AKI after CA is associated with shock, older age, comorbidities, and illness severity but, it is important to not, not with iodinated contrast exposure during the first 24 hours after resuscitation.^224,226

Lower cortisol levels after CA were associated with the degree of shock and with mortality during the first 48 hours after CA, with up to 75% of nonsurvivors having low cortisol.^227,228 Supplemental steroids have inconsistent associations with outcome after CA, but much of these data precede the era of temperature control. Two nonrandomized studies compared patients treated with steroids and those not treated with steroids after CA and found no significant difference in the incidence of survival or regaining consciousness.^229,230 Other nonrandomized studies have correlated steroid use in patients after CA with higher rates of survival to discharge.^231,232

Clinical trials of steroid administration during and after CA have found no overall improvement in outcome but suggest subgroups of interest. Hydrocortisone administration in the emergency department for patients during CA increased the ROSC rate in the hydrocortisone group but did not change hospital complications, survival, or hospital discharge rates.²³³ Two trials had a complex design comparing one group who received vasopressin, epinephrine, and supplemental corticosteroids initiated during CPR and up to 7 days with a group who received epinephrine and no corticosteroids. Both trials found that the vasopressin-epinephrine-steroid group did better than the comparator group.^234,235 Post hoc analysis of these trials found increased benefit of steroids in patients with septic shock.²³⁶ A blinded RCT of 50 patients after CA found no difference between the hydrocortisone and placebo groups in time to shock reversal, shock reversal, good neurological outcome, or survival to discharge.²³⁷

Glucose control in the general ICU population affects complication rates and mortality. Conventional glucose control (140–180 mg/dL) is recommended over intensive glucose control (81–108 mg/dL) on the basis of a large randomized trial showing decreased mortality in the conventional group.^179,238 Hyperglycemia is common in patients after CA and can be influenced by phase of temperature control.²⁴⁰ Hypoglycemia is also harmful and can worsen secondary brain injury. Because patients have dynamic physiology in the early hours and days after CA and because interventions like hypothermic temperature control can alter insulin resistance, short-acting drugs that are easier to titrate may be best for glycemic control. Hyperglycemia, glucose variability, and delay in the correction of hyperglycemia are associated with unfavorable neurological outcomes; however, it is unclear whether this is due to the hyperglycemia itself, the underlying severity of illness, or the aggressiveness of ICU care.^241–243 One small randomized trial in patients after CA comparing strict (72–108 mg/dL) with moderate (108–144 mg/dL) glucose control demonstrated no benefit of strict control on mortality at 30 days.²⁴⁴

Given the low certainty of evidence, endocrine management and fluid resuscitation are important topics for further investigation. Studies investigating the timing and methods of RRT, especially in the setting of cerebral edema, are of importance. Investigations into optimal glucose management and the use of steroids are also important.

The concept of a CA center of excellence is included in statements from the AHA and the International Liaison Committee on Resuscitation.^103,245,246 In patients with multiple diagnoses, including stroke, trauma, and ST-segment–elevation myocardial infarction, outcomes are improved in systems of care that cohort patients.^247,248 In patients with CA, the data on triaging and transporting patients to specialty centers are inconsistent and influenced by local factors.

A retrospective cohort study of CA care at 37 hospitals in Ontario found that higher-volume centers were more likely to successfully implement hypothermic temperature control, but there was no impact on premature withdrawal of life-sustaining therapy or survival with good neurological function.²⁴⁹ A large database study of academic medical centers contributing to the Vizient clinical database compared characteristics of academic medical centers with higher-than-average CA survival rates with those with lower-than-average survival rates.²⁵⁰ Centers with higher survival rates had higher CA and surgical case volumes and were more likely to offer cardiac catheterization or cardiac surgery. These centers were also more likely to be located in catchment areas with higher household incomes and lower poverty rates. A retrospective analysis of >44 000 patients who survived CA in Japan found that patients receiving care at high-volume centers were more likely to be discharged with a favorable outcome.²⁵¹ However, a retrospective analysis of the 48 hospitals contributing to the Great Paris Registry found that characteristics of receiving hospitals (including size and cardiac catheterization availability) were not associated with survival rate at hospital discharge.²⁵² This lack of association was explained, in part, by the presence of a prehospital dispatching protocol that orients prehospital services to centers with appropriate resources.

Although some data suggest a trend toward improved outcomes when patients receive care at specific centers, these studies may reflect regional factors such as the presence or absence of prehospital destination protocols for emergency services, socioeconomic status of the patient population, characteristics of the destination hospital, and services offered at the destination hospital. In addition, the studies are variable in their assessment for post-CA care. We lack clear data that describe which hospital characteristics are most directly associated with improved outcomes. These studies should be considered in the design of future experimental studies. Given the lack of data and the potential implications, a general statement guiding the diversion or transfer of patients to certain centers during or immediately after a CA could not be broadly supported. However, it was agreed that certain post-CA services be offered at centers routinely providing post-CA care and even advertising expertise in CA care. Centers routinely providing care for patients after CA should collect data on their outcomes and routinely engage in quality improvement activities guided by data. Additional research is needed to understand the role of legislated or mandated systems of care based on designated or certified CA centers. The variable outcomes reported in studies by hospital destination combined with the resources necessary to develop a system of care based on CA centers warrant further study.

Protocol-guided care is associated with improved adherence to key processes and a decrease in practice variability, as well as improved patient outcomes. However, studies suggest that the presence of a protocol in protocol-guided care is often more important than the individual components of the protocol.²⁵³ In addition, lack of agreement on what constitutes a protocol complicates researching the impact of a protocol on patient outcomes. However, the management of a patient’s post-CA care is interdisciplinary and often complex. Integration of multiple disciplines to determine roles, responsibilities, and the overall scope of care is a reasonable best practice in the post-CA patient population.

Multiple single-center observational and experimental studies demonstrate improved outcomes when post-CA care is organized by the use of protocols.^254–256 Limitations of the studies included lack of homogeneity in the definition of protocol, components of the protocol, and evaluation of outcomes. However, the studies consistently demonstrated improved patient outcomes, including improved processes for cardiac catheterization and temperature control, as well as improved outcomes, including mortality and measures of functional recovery, when care was organized by the use of a protocol. A 2019 meta-analysis of 6 studies including 1422 patients evaluated the impact of a structured care pathway for the management of early coronary angiography, temperature control, and ICU care.²⁵⁷ The meta-analysis found that patients who were treated with a structured care pathway had a higher likelihood of favorable functional outcomes than those treated with standard care.

Supporting the family during the crisis of post-CA intensive care is also an important component of bundled care. Studies report high rates of distress and posttraumatic stress after a loved one is hospitalized with critical illness.^258,259 Family support structures in the ICU, including social workers, transparent and available staff, and frequently scheduled updates with a consistent message by a stable care team, are important aspects of comprehensive care.

Although comprehensive care delivered with an organized protocol is good practice and should be considered at centers routinely providing post-CA care, the components of each protocol are strongly influenced by local contextual factors. Protocols that address the aspects of care (Table 3) are important. Additional research to identify the importance of individual bundle and protocol elements is needed. Systems-of-care research that better identifies which patients should be transferred for care and at what point in their care spectrum they should be transferred is also important (Figure 2). Patients and families are central to the care provided in ICUs to patients after CA. Thoughtful management of complex physiology requires comprehensive care and expertise.