The four phases of cardiac arrest

There are four phases of cardiac arrest (figure 1).

Figure 1

The four phases of paediatric cardiac arrest. CPR, cardiopulmonary resuscitation.

Paediatric arrest in numbers

It is estimated that the incidence of atraumatic out of hospital cardiac arrest (OHCA) is 8.04 per 100 000 person-years2 and an approximate rate of ROSC in 36% of cases. Worldwide survival rates of children who arrested in emergency departments vary between 12.8% and 33.8%,3 whereas cardiac arrests that occurred outside the hospital had a lower survival rates, reported as 2.6%–14.7% in various regions of the USA.4 Of those surviving, favourable neurological outcomes have been documented to be as low as 10%5 as and as high as 71%.6

Severe respiratory compromise and circulatory failure, in that order, are the most common causes of cardiac arrest in childhood. Both conditions can cause respiratory arrest with ensuing hypoxia, which subsequently results in cardiac arrest. The majority of cardiac arrests in children present with non-shockable rhythms. Children with chronic comorbidities, such as asthma, congenital heart disease or neurodisability, are more likely to suffer a cardiac arrest than age-matched patients without such disease burden.

In the adult population, cardiac arrest is usually due to long-term illness, such as heart disease and chronic obstructive pulmonary disease. The chronicity of these conditions leads to myocardial ischaemia, which is further impacted by acute myocardial insult that is more likely to manifest in a shockable rhythm. This basic difference in pathology explains why the CPR ratio in children is 15:2, in contrast to the adult ratio of 30:2. While the former is focused in reversing the causes of respiratory compromise (the likeliest cause of arrest) and allow oxygenation and reperfusion, the latter is designed to give more compressions in order to improve coronary circulation. Good-quality CPR can address hypoxia and maintain adequate perfusion to coronary and cerebral circulation, regardless of the age of the patient.

Postcardiac arrest syndrome (PCAS)

The period following ROSC is known as PCAS (table 1). During this phase, the risk of reperfusion injury and arrhythmias is at its highest. PCAS severity depends on the cause of cardiac arrest as well as the duration.

Prolonged ischaemia and subsequent rapid reperfusion tissue injury are postulated as the main pathophysiological mechanisms.7 This dyad is thought to disrupt homeostasis, contribute to free radical formation, activate proteases and cause a systemic inflammatory response syndrome-like response.

The four stages of PCAS are:

• Immediate post arrest: first 20 min.

• Early post arrest: 20 min to 6–12 hours.

• Intermediate phase: 6–12 hours up to 72 hours.

• Recovery phase: from 72 hours onwards.

Neuroprotection and neurological assessment

Evidence suggests that high-quality CPR can, at most, achieve around 50%–60% of normal cerebral blood flow (CBF).8 The greater the reduction of CBF in this setting, the lesser likely it is for the patient to survive and the worse the long-term neurological sequelae.

Following a cardiac arrest, the hypoxic and ischaemic insult to the brain contributes to a cascade of events that leads to secondary neurodegeneration. Both cerebral oedema and hyperaemia ensue due to impaired vascular reactivity as the brain attempts to reperfuse. The course of these events and the extent of the secondary brain injury can be augmented, to an extent, by proactive neuroprotection that needs to commence at the point of ROSC.

It is vital that we frequently assess neurological status following ROSC, this can aid in prognostication. Aim to perform a very quick assessment before the administration of anaesthetic and paralysing agents. Clearly document the GCS, the presence of any respiratory effort (or gasping), pupillary size and reactivity. Minimise delays between assessment and commencement of adequate anaesthesia and paralysis. This will enable better neuroprotection and improve patient comfort during ventilation and the insertion of central/arterial line.

Temperature control

The topic of therapeutic hypothermia following cardiac arrest has generated lots of research and academic discourse in recent years. While evidence suggests that application of therapeutic hypothermia correlates to better neurological outcomes in some patient groups, namely adult patients with VF arrest and newborns with birth asphyxia, evidence of its application in children following ROSC is less clear. Targeted therapeutic hypothermia, especially at temperatures around 32°C, are associated with worse survival, in part due to the increased risk of arrhythmias, coagulopathy and immunosuppression with subsequent infections.9

There is clear evidence, however, that fever in the post-ROSC phase worsens neurocognitive outcomes in cardiac arrest survivors, so while there is significant variability on the application of therapeutic hypothermia among paediatric intensive care units in the UK, there is gradually a shift towards targeting normothermia and the avoidance of fever. Fever increases metabolic demand and worsens secondary injury to the ischaemic brain whose autoregulatory processes fail to acclimatise to the postarrest imbalance between metabolic demand and CBF. In the context of traumatic injury to the brain, it has been estimated that every 1°C increase in temperature increases the risk of adverse outcomes by 220%. A further 0.5°C rise in temperature can cause neuronal death and exacerbate secondary injury.10

There is consensus that patients (irrespective of age) should be rewarmed above the commonly accepted ‘danger zone’ for the occurrence of ventricular fibrillation (28°C–30°C) as quickly as possible.11 This mainly (but not solely) applies to patients that arrest secondary to cold water drowning.

The rate at which a hypothermic patient suffering a cardiac arrest secondary to cold water drowning should be rewarmed depends on the core temperature on presentation. Any temperature under 30°C needs a fast rewarm. The choice of rewarming therapy hinges on device availability and accessibility. Plumbed garments, warm intravenous fluids and heated mattresses are easy and effective methods. There is not any high-quality evidence to differentiate which modality or combination thereof is safer or faster. The key is not to overwarm past the 30°C mark.

After the patient has reached 30°C, it is crucial that we rewarm slowly. Evidence suggests that the ‘sweet spot’ lies between 0.5°C and 2°C an hour so generally aim for the median of 1.2°C/hour. Rewarming too fast with rates of greater than 5°C/hour is associated with higher in-hospital mortality.12

Therapeutic hypothermia is no longer routinely recommended for survivors of paediatric out-of-hospital cardiac arrest due to drowning. The THAPCA-OH trial found that hypothermia did not result in a statistically significant benefit with respect to survival with good functional outcome at 1 year when compared with normothermia. Mortality risk at 28 days and 12 months did not differ between treatment groups either13 but a small benefit in neurocognitive outcomes has been suggested by posthoc analysis of the study.14

Sedation, paralysis and seizure control

Both adequate sedation and paralysis are vital components in managing patients following ROSC. Paralysis is necessary in optimising the patient for a safe intubation.15 It is also useful in improving mechanical ventilation and minimising muscle metabolic demand and shivering, especially in cases where the patient is hypothermic.16 It is important to realise that relying on paralysis alone for intubation is no longer recommended practice, as it increases the chances of accidental extubation and postcardiac arrest psychological sequelae due to patient awareness.17 Using paralysis and some sedation during intubation is the preferred practice. Sedation is important as it helps patients tolerate life-saving invasive procedures, aids against seizure control but also slows down metabolic demand of the brain, thereby aiding against secondary brain injury. The selection of sedation depends on the patient’s pathology and cardiovascular status, but in the UK, a combination of morphine and midazolam is the most common with the occasional use of fentanyl, ketamine or propofol as continuous intravenous infusions.

The incidence of seizures in children following cardiac arrest has been reported to be between 10% and 47%.18 The nature of seizures is variable and vary from refractory status epilepticus to subclinical seizures. Electroencephalographic (EEG) features, seizure resolution and response to anticonvulsants are important in the context of prognosticating neurological outcomes and survival, following cardiac arrest. It is, however, unlikely that EEG or cerebral function analysing monitoring is available prior to transfer to PICU. In the clinical setting where a patient is intubated, sedated and paralysed following a cardiac arrest, the ability to accurately diagnose postarrest seizures is blunted. Good neuroprotection is important; if we are to optimise the chances of a favourable neurological outcome pre-emptive control of potential seizure activity is vital. In the UK, it is a common practice in to maintain anaesthesia using a continuous infusion of benzodiazepines, usually midazolam. It is also good practice to load these patients with a long-acting anticonvulsant. Levetiracetam has been found to be non-inferior to phenytoin in two well-conducted randomised control trials.19 Given that arrhythmias are a potential complication of myocardial ischaemia following a cardiac arrest, phenytoin is generally avoided in this setting due to its arrhythmogenic properties and a 40 mg/kg loading dose of levetiracetam is preferred.

Oxygenation

Hypoxia is proven to worsen outcomes following the successful ROSC.20 Avoidance of hyperoxia (with the exception of cases of CO poisoning or severe anaemia) is equally important. Hyperoxia during the postresuscitation phase has been shown to amplify free radical production, thereby leading to neuronal damage.21 Moderate hyperoxia has been shown to improve long-term organ function and neurological outcomes.22 However, studies focusing on the effects of severe hyperoxia (PaO₂ over 40 kPa) following resuscitation have consistently shown an association with poor outcomes following cardiac arrest.23

It is reasonable to have high concentration oxygen therapy during the low flow resuscitation and early postresuscitation phases as the most common causes of cardiorespiratory arrest in children are of respiratory origin. In the subsequent phases, we should target oxygen saturations between 88% and 92%24 and be proactive in reducing FiO₂. It is clinically appropriate to aim for higher oxygen saturation in cases of severe anaemia or carbon monoxide poisoning.

Blood pressure (BP) control

It is well evidenced that hypotension following ROSC worsens the clinical prognosis as physiologically this exacerbates cerebral ischaemia.25 Invasive BP monitoring is the gold standard of care for post-ROSC care in children as non-invasive BP monitoring during this period is likely to be inaccurate. If invasive monitoring capabilities locally are limited then non-invasive monitoring set at 1–2 min intervals can be used as a proxy until arterial cannulation is achieved.

During the first few minutes following ROSC, poor cardiac function ensues due to a combination of acidosis and systemic vasoconstriction. Adequate fluid resuscitation and inotropic support to target an age-appropriate BP are the main ways in which the myocardium can be supported. Repeated bedside examination after each intervention, especially focusing on the presence of hepatomegaly or rales, can help avoid fluid overload. When used in conjunction, central venous and arterial pressure monitoring as well as point of care ultrasound can be extremely useful in titrating fluid management.

An exhaustive review of inotrope choice and physiological rationale is beyond the scope of this paper and has been addressed in another article by the authors.26 Broadly, the selection of inotrope largely hinges on whether the patient requires more inotropy or more vasoconstriction. Epinephrine achieves the former, while norepinephrine achieves the latter. Following return of spontaneous circulation, there is frequently a degree of myocardial dysfunction. It is therefore good practice to start an epinephrine infusion in order to support the myocardium and achieve the desired BP. Intraosseous access allows us to give central concentration strength infusion until central access is obtained either by the local or paediatric critical care team. If attempts at intraosseous access fail, peripheral administration using a dilute concentration is also effective. Gold standard practice dictates that we should monitor the patient’s arterial BP. This would allow us to titrate and wean epinephrine proactively. Until arterial access is obtained, peripheral BP should be monitored automatically on a 1 min cycle. This allows for the commencement of peripheral inotropes early, without delay, regardless of whether or not the clinical team can obtain central and arterial access on a child.

Decisions in this aspect of clinical care are usually made in tandem with the regional critical care retrieval team. Current practice based on evidence from adult studies suggests that in cases of traumatic cardiac arrest we should target a ‘low normal’ BP27 and in atraumatic cardiac arrest for some benefit may be conferred by maintaining a ‘high normal’ BP28 in order to maintain adequate cerebral perfusion in the face of anticipated cerebral oedema.

It is worth considering that extreme acidosis (pH<7) gives a physiological effect of catecholamine resistance whereby exogenous and endogenous catecholamines become less effective. While routine correction of acidosis with alkalinising agents appears to neither worsen nor improve clinical outcomes, administering bicarbonate when the pH is <7 may help inotropes work better.29 There is of course clinical merit in administering alkalinising agents if the arrest has been caused by a tricyclic antidepressant overdose30 or if the patient is hyperkalaemic.30

Positioning

In general, positioning that can achieve optimum cerebral perfusion is with the patient at a 30° angle31 with the head in the midline. This widely accepted approach to positioning, albeit without a strong evidence base, is thought to maximise cerebral arterial perfusion without compromising cerebral venous drainage. Any implements that can impede cerebral venous drainage such as central venous lines in the neck or a cervical collar should, if possible, be avoided.

Eucapnia

Arterial partial pressure of carbon dioxide (PaCO₂) is a marker of the homeostatic equilibrium between CO₂ clearance and CO₂ production. Fluctuations of pH and PaCO₂ can both affect CBF by their effects on arterial vascular tone.

CO₂ has systemic vasodilatory effects, while hypocapnia causes cerebral vasoconstriction.32 The effects of pH alone on vascular tone are outwith the purposes of this paper, they are complex to understand and largely depend on whether the pH is intracellular or extracellular and whether the patient is hypocapnic or hypercapnic.33

While hyperventilation strategies during acute spikes of ICP are used to cause hypocapnia and a rapid reduction in ICP,34 prolonged and sustained hypocapnia in patients with brain injury has been shown to worsen clinical outcomes35 due to a reduction in CBF.

Physiologically, hypoventilation can lead to hypoxia and a secondary increase in ICP due to hyperaemia, which in turn exacerbates cerebral oedema.36 High PaCO₂ can also worsen right ventricular function.37

Similarly, hyperventilation can increase intrathoracic pressure, thereby impeding cardiac venous return, which in turn causes a BP drop and a worsening of cerebral perfusion.

It is also worth noting that some studies suggest that CO₂ has antioxidant anti-inflammatory and anticonvulsant properties.38 These properties suggest that CO₂ may have a neuroprotective role in secondary brain injury. What we do know is that extremes of PaCO₂ on either end of the normal physiological range are associated with significantly worse outcomes following ROSC.39

It seems prudent therefore that avoidance of hypocapnia or hypercapnia is key in preventing secondary brain injury in the children following ROSC. The best available evidence would suggest our target should be between 4.5 and 6.0 kPa.40

Glycaemic control

Hypoglycaemia or hyperglycaemia are not uncommon following paediatric ROSC. Brain excitability and glycaemic control are closely associated. Glucose is the main energy substrate of neurons. Glycaemic fluctuations can promote instantaneous hyperexcitability and subsequent seizures.

Hypoglycaemia is remarkably epileptogenic, more so in newborns (due to high metabolic demand), patients with neurometabolic disorders and in children with diabetes mellitus. There is evidence to suggest that hypoglycaemia negatively impacts neurological outcomes and can predispose to the development of neurocognitive challenges in childhood.41 Vigilant glucose monitoring and correction are vital. Overt reliance on dextrose boluses should prompt consideration for a continuous glucose infusion.

Hyperglycaemia is usually an appropriate physiological response to the stress experienced by the arrested patient. Hyperglycaemia can also cause acute symptomatic seizures although rarely and usually as acute complications of diabetes mellitus. Evidence suggests that aggressive control of hyperglycaemia in the post-ROSC period confers no added benefit to mortality or morbidity.42 The exception to this is diabetic patients who still require tight glycaemic control as, in that group of patients, hyperglycaemia has been found to worsen prognosis.43 In non-diabetic patients, patience and watchful deliberation are enough to allow the glucose levels to drop down to normal levels without the need for pharmacological intervention.

Adjunctive investigations

Diagnosis and ongoing management following paediatric ROSC will be dictated by history and clinical picture, the latter would be incomplete without a set of investigations. Renal function, electrolytes, liver function tests, full blood count and clotting are always needed as a basic starting point. In cases of severe metabolic acidosis, lactaemia or a high anion gap, checking ammonia levels and toxicology, is necessary. The fastest and most efficient means of checking and correcting electrolyte abnormalities and adequacy of ventilation/oxygenation is arterial blood gas (ABG) analysis. ABGs can also help in the diagnosis of carbon monoxide poisoning.

Imaging also plays a role in tailoring ongoing management. X-rays following intubation are important in order to determine endotracheal tube position and lung pathology. X-Rays can also give diagnostic clues if cardiac pathology is suspected. In cases of traumatic cardiac arrest or if NAI is suspected, CT of the head is important. CT of the head should be done early in these situations, ideally before the arrival to PICU as that would minimise the time taken for diagnosis and appropriate, urgent, neurosurgical management if indicated.

In the UK, unexpected childhood deaths and near-miss cases are subject to a sudden unexpected death in infancy investigation. It is important for the clinicians involved to undertake referral to the child protection team, police and social care so that due process is followed.

Preparation for transfer to PICU

Following ROSC, patients will need to be stabilised, packaged for transport and transferred to the nearest PICU for ongoing management. In certain time-critical cases, general or neurosurgical intervention is needed prior to the admission to PICU. Whether the child is transferred by a team mounted by the local hospital or a regional critical care transport team depends on local network agreements and policies. Adequate neuroprotection, by applying the simple principles outlined above, is feasible even during the retrieval process. Target euthermia and treat hypoglycaemia aggressively. Maintain a low threshold of loading with long-acting anticonvulsants if seizures are suspected. Aim for eucapnia and avoid extremes of oxygen saturations by monitoring EtCO₂ and SpO₂ and adjusting ventilation accordingly. Aim for a high-normal BP in atraumatic cardiac arrest and low-normal in trauma and use a combination of volume and inotropes depending on the pathophysiology to achieve that target. If inserting an arterial line will delay the transfer significantly then use a non-invasive BP monitoring set at 1 min cycles. Similarly, avoid task-fixation when trying to insert a central line, an intraosseous line is just as effective and easier to insert. Ensure that adequate history and examination is done and have a low threshold of suspicion for non-accidental injury in infants or if the clinical picture does not quite match the history given.

Finally, do not forget about the family. Ensure a team member updates them regularly, ideally every 10–15 min. This may be a challenging day at work for us, but it is probably the worst day of their lives so use compassion liberally.