Cardiopulmonary resuscitation

Incidence of cardiac arrest during anesthesia

This chapter provides an overview of incidence, etiologies and outcomes of pediatric perioperative cardiac arrest, the physiology of cardiopulmonary resuscitation, current therapeutic recommendations as well as physiologic and scientific underpinnings of those recommendations. The 2020 American Heart Association updates are presented thoughout the chapter and summarized in Box 57.1 . A perioperative cardiac arrest is an event that requires chest compressions while a patient is under an anesthesiologist’s care during the intraoperative or immediate postoperative period. The causes of perioperative cardiac arrest include factors related to the anesthesia, the surgical procedure, or patient comorbidities. Anesthesia-related cardiac arrest and mortality in children are infrequent events, and the reporting of available data is challenging to interpret. First, there is considerable variation in how reports define anesthesia-related cardiac arrest. The definitions vary from arrests in which anesthesia is the major causative factor to arrests in which anesthesia has any role. Second, the phases of care involved vary from including only the intraoperative period to including premedication through the first 24 hours (or longer) of postoperative care. One study looked at anesthesia-related deaths up to 30 days postoperatively ( ). Third is the possible gross underestimation of anesthesia-related cardiac arrest and mortality when voluntary reporting is involved. Fourth is that studies that rely on cardiac arrest as the inciting event for reporting may miss cases in which anesthesia-related cardiac arrest and mortality occurred postoperatively without capturing the responsible intraoperative events ( ). The final challenge in interpreting this literature is the variation in whether studies include arrests that occur during cardiac surgery. This absence of standardized methods of reporting combined with the infrequent occurrence of perioperative arrests contributes to a lack of precise information regarding the incidence and risk factors for anesthesia-related cardiac arrest and death in children.

Table 57.1 shows the reported incidence of pediatric perioperative cardiac arrest for all types of procedures, including cardiac surgery. The overall incidence of perioperative cardiac arrest for children of all age groups and undergoing all types of surgeries ranged from 5.1 to 22.9 per 10,000 procedures ( ; ; ; ; ; ; ; ). Notably, studies that are more recent have larger patient cohorts and lower incidences of perioperative arrest. Two reviews of >250,000 and >1,000,000 pediatric anesthetics both reported an incidence of 5 perioperative arrests per 10,000 anesthetics ( ; ). Studies that excluded cardiac surgery had lower overall incidences that ranged from 2.9 to 7.4 per 10,000 procedures ( ; ; ). Studies that reported on children undergoing cardiac surgery have the highest incidence of perioperative cardiac arrest, from 79 to 127 per 10,000 procedures ( ; ). When only anesthesia-related cardiac arrest is reported, the arrest incidence for all types of surgery (including cardiac) ranges from 0.8 to 4.6 per 10,000 procedures.

As Table 57.1 indicates, young age is a consistent risk factor for pediatric perioperative cardiac arrest. The highest risk is in infants <1 month of age, followed by those <1 year ( ; ; ; ; ; ; ; ; ). The Pediatric Perioperative Cardiac Arrest (POCA) Registry is a voluntary reporting registry formed in 1994 to investigate the causes and outcomes of pediatric perioperative cardiac arrest and death (≤18 years of age) in the operating room or postanesthesia care unit (PACU). A POCA registry analysis showed that 56% of anesthesia-related cardiac arrests were infants <1 year during 1994 to 1997, and that percentage fell to 38% during 1998 to 2004. They attribute this decrease to a switch from halothane to sevoflurane, which is associated with less bradycardia and myocardial depression ( ). Unfortunately, infants <1 year continue to have the highest rate of perioperative cardiac arrest. A large review found a threefold increase in perioperative cardiac arrest in children ≤6 months old compared with children >6 months old ( ). If adult perioperative arrests are included, infants have an incidence of perioperative cardiac arrest that is over four times the second highest age group (>80 years old). Fig. 57.1 shows the incidence and outcomes of perioperative cardiac arrest by age as reported to the National Anesthesia Clinical Outcomes Registry from 2010 to 2013 ( ).

The patient’s preoperative physical condition (stratified by their American Society of Anesthesiologists [ASA] physical status classification) also influences the risk of perioperative cardiac arrest. The risk of intraoperative cardiac arrest is increased in patients with an ASA physical status ≥3 ( ; ; ; ; ; ; ; ; ; ). Patients with an ASA physical status of 5 are often excluded because (by definition) they have a low likelihood of survival and it is difficult to determine whether their arrest is a result of their condition or related to anesthesia. Patients with an ASA classification ≥3 are nine times more likely, and those with an ASA classification of 4 and 5 are 30 to 300 times more likely, to have a perioperative cardiac arrest than those with an ASA classification of 1 or 2 ( ; ; ). Prematurity, congenital heart disease, and other congenital defects are comorbidities that increase the risk for children, especially in the neonatal period ( ; ; ; ).

The designation of emergency status to a patient’s procedure is a risk factor for both perioperative cardiac arrest and mortality. Emergency designation is associated with a 3.5- to 6.8-fold increase of pediatric perioperative cardiac arrest ( ; ). In addition to a higher incidence of arrests, emergency status is related to increased mortality after perioperative cardiac arrest in both children and adults ( ; ; ; ). The patient’s condition, the lack of optimal preoperative preparation and personnel, or both could be the reason for the increased risks of perioperative cardiac arrest and death in emergent procedures.

Etiology of cardiac arrest during anesthesia

Etiologies of cardiac arrest during anesthesia are categorized either by the organ systems involved or by the interventions applied. Table 57.2 shows a summary of the reported etiologies of cardiac arrest and the phase of the anesthetic. The POCA Registry uses a categorization system that involves both interventions and organ systems and reports arrests as medication-related, cardiovascular, respiratory, or equipment-related ( ; ). Some etiologies may be difficult to categorize because they fit into several groups. For example, succinylcholine-induced dysrhythmia can be classified as either a medication-related or a cardiovascular cause of cardiac arrest. A set of guidelines for reporting cardiac arrest data, known as the pediatric Utstein guidelines, suggested an organ system–based classification for etiologies: cardiac, pulmonary, and cardiopulmonary ( ). The Utstein guidelines have not been widely incorporated into the reporting of perioperative cardiac arrest. Typically, the anesthesia literature groups the etiology of cardiac arrest into medication-related, cardiovascular, or respiratory causes, with reports varying the contents within each classification. Box 57.2 shows general categories and etiologies of cardiac arrest during anesthesia.

Medication-related etiologies were historically the most frequent causes of anesthesia-related cardiac arrest in children (approximately 35%; range, 4% to 54%) ( ; ; ; ; ; ; ). Since the late 1990s, medication-related etiologies have decreased (5% to 28%), and cardiovascular and respiratory causes are the most frequently reported ( ; ; ; ). Fig. 57.2 shows the change in perioperative cardiac arrest etiology during the two POCA reports from 1994 to 1997 and 1998 to 2004 ( ). This decrease in medication-related arrests may be the result of reductions in inhalational agent overdoses, dysrhythmias, and myocardial depression that correspond with sevoflurane replacing halothane for inhalational induction. It is not clear if sevoflurane is less cardiotoxic than halothane or if the halothane vaporizer allowed delivery of greater multiples of the minimum alveolar concentration (MAC) of anesthetic than the sevoflurane vaporizer (1 MAC halothane = 0.8% with maximum vaporizer setting 5% versus 1 MAC sevoflurane = 3.3% with maximum vaporizer setting 8%) ( ). Similarly, succinylcholine-induced dysrhythmias during anesthesia decreased when a black box warning regarding the use of succinylcholine in children was issued in the late 1990s. One medication-related cause of anesthesia-related cardiac arrests that continues to be common is the relative anesthetic overdose. A relative anesthetic overdose involves the use of an appropriate dose of an intravenous (IV) or inhaled anesthetic that causes an unexpected hemodynamic effect (usually because of patient morbidity, e.g., sepsis or cardiac disease) and subsequent cardiovascular compromise or collapse requiring resuscitation ( ). Other medication-related causes of anesthesia-related cardiac arrest include syringe swaps (administration of the wrong medication) and the accidental IV administration of local anesthetic intended for the epidural space that results in local anesthetic toxicity. Medication-related causes of cardiac arrest that occur in the PACU usually include the inadequate reversal of muscle relaxant and opioid-induced respiratory depression.

Cardiovascular etiologies cause approximately 40% to 70% of the anesthesia-related cardiac arrests in children ( ; ; ; ; ). Hypovolemia-associated arrests reported in this category include inadequate volume administration, excessive hemorrhage, and inappropriate resuscitation with crystalloid and/or colloid ( ; ; ). The hyperkalemia-associated dysrhythmias reported in this category are caused by succinylcholine administration, rapid transfusion of blood products, reperfusion, myopathy, or renal insufficiency ( ). Vagally induced dysrhythmia or cardiovascular collapse (asystole) can result from procedural traction or pressure (insufflation) exerted on the abdomen, eye, neck, or heart. Anaphylaxis-associated cardiovascular collapse can occur from exposure to a wide variety of common operating room substances (latex, contrast, medications, or dextrans). Patients with underlying cardiac disease are more likely to arrest from a cardiovascular etiology (50%) than are patients without cardiac disease (38%) ( ). Venous air embolism is an important cause of cardiovascular collapse and cardiac arrest in patients under anesthesia. And finally, malignant hyperthermia is an infrequently reported cause of cardiac arrest in this group.

Respiratory etiologies cause approximately 35% (range: 15% to 71%) of anesthesia-related cardiac arrests in children and adults ( ; ; ; ; ; ; ; ; ). Respiratory etiologies of cardiac arrest have declined over the years as a source of malpractice claims, from 51% in the 1970s to 23% by 2000 ( ). This decline may be the result of the increased use of continuous pulse oximetry and quantitative end-tidal carbon dioxide monitoring in the perioperative period. Inadequate oxygenation and ventilation are broad categories often indicated as causes of cardiac arrest in this group. “Loss of the airway” is another broad category and may involve laryngospasm, bronchospasm, an anatomically difficult airway, or an endotracheal tube (ETT) that is misplaced, kinked, plugged, or inadvertently removed. Aspiration is a respiratory cause of anesthesia-related cardiac arrest, but although the aspiration occurs in the operating room, the arrest often occurs much later because of pneumonitis and hypoxia, and this etiology is infrequently reported.

Equipment-related etiologies cause approximately 4% (range, 0% to 20%) of anesthesia-related cardiac arrests in children and adults ( ; ; ). The most common causes of equipment-related cardiac arrest include bleeding or dysrhythmia related to central venous catheter insertion and breathing circuit disconnection. Other, unclassified, causes of anesthesia-related cardiac arrest reported include multiple events (3%) ( ), inadequate vigilance (6%) ( ), and unclear etiology (9%; range, 1% to 18%) ( ).

Determination that a perioperative cardiac arrest is attributable to the anesthetic or anesthesia care is a subjective decision. Variability in the interpretation of these definitions and in methodology makes it difficult to compare and reconcile the rates of anesthesia-related cardiac arrest and death ( ). Patient-related factors, procedure-related factors, and anesthesia care–related factors are the three most important determinants of etiology of intraoperative cardiac arrest. Attempts to determine the extent to which anesthesia care causes an arrest have resulted in the addition of terms such as “anesthesia-associated” and “anesthesia-attributable” cardiac arrest. The relative contribution of patient comorbidities and surgical or procedural factors further complicates this determination. To what extent does the anesthesia care contribute to an arrest caused by surgical bleeding in a coagulopathic patient? Some studies simply use the term anesthesia-related to describe a cardiac arrest that occurs after an anesthesiologist has been involved in the care of the patient.

It is important to know that anesthesia-related cardiac arrests are often preventable and frequently reversable. It is estimated that 53% of anesthesia-related cardiac arrests and 22% of anesthesia-related mortalities are preventable ( ). Human error is an important factor in deaths attributable to anesthesia and usually manifests not as a fundamental ignorance but as a failure to apply existing knowledge ( ). Poor preoperative preparation and inadequate vigilance are frequently reported avoidable errors. Proper preoperative preparation by pediatric anesthesiologists includes the identification of symptoms in patients with an undiagnosed muscular dystrophy, with coronary involvement from William syndrome, with prolonged QT syndrome, or with cardiomyopathy. Adequate vigilance includes early recognition of progressive bradycardia and rapid response to persistent hypotension. A relative anesthetic overdose can be avoided by using a “test dose” or divided dosing when administering medications that may cause hypotension in unstable patients. Transfusion-related hyperkalemia, local anesthetic toxicity, and inhalational anesthetic overdose are other important and preventable etiologies of “anesthesia-related” cardiac arrest ( ).

Non-anesthesia-related cardiac arrest is most often the result of the patient’s underlying condition or the procedure being performed. Trauma, exsanguination, and failure to wean from cardiopulmonary bypass are three of the most often reported causes of non-anesthesia-related cardiac arrest. Myocardial infarction, pulmonary embolus, sepsis, and ruptured aneurysm are other, less frequently observed patient-related causes of cardiac arrest. Procedure-related causes include technical problems, caval compression, vagal asystole related to traction or insufflation, and complications related to transplantation.

In summary, major risk factors for anesthesia-related mortality have been identified and include young age (especially <1 month of age but also <1 year of age), ASA physical status of ≥3, emergent procedures, and cardiac surgery. There is a need for a standardized definition of anesthesia-related cardiac arrest because, without a standard, it will be difficult to determine whether we are making progress in this area. Fortunately, the reporting of etiologies of anesthesia-related cardiac arrest is becoming more congruent, and common causes are recognized. Based on the knowledge of these etiologies, improvements in preparation and vigilance can help prevent or reverse anesthesia-related cardiac arrest.

Outcomes from cardiac arrest during anesthesia

What is the risk of a child dying during the perioperative period? Studies investigating this question report varied results depending on whether they include only “anesthesia-related” or all causes of perioperative cardiac arrest. Although survival is the outcome most commonly considered a measure of successful resuscitation, mortality is the variable most commonly reported in the literature. “Anesthesia-related” mortality during 2000 to 2018 is reported as 0.1 to 1.6 per 10,000 procedures ( ; ; ; ; ; ), which is down from the 2.9 per 10,000 reported between 1947 and 1958. Three studies between 2001 and 2006 reported no anesthetic-related deaths ( ; ; ). When all causes of perioperative cardiac arrest are included (anesthetic, surgical, and patient disease), the risk of mortality is 3.8 to 13.4 per 10,000 procedures in reports from 1990 to 2011 ( ; ; ; ; ). The largest report of pediatric perioperative cardiac arrest, encompassing over 1 million anesthetics from 2010 to 2015, found a mortality of 0.94 per 10,000 procedures ( ).

When reporting survival, there are inconsistencies in the duration of the survival (e.g., to 24 hours, to hospital discharge, to 6 months, etc.) and in the quality of the survival (neurologic status is missing). A patient may “survive” the initial resuscitation attempts but subsequently die in the intensive care unit (ICU) from persistent hemodynamic instability or be left with a devastating neurologic injury. The return of spontaneous circulation (ROSC) is defined as the restoration of a perfusing rhythm and blood pressure that persists for at least 20 minutes post arrest. The rate of ROSC is sometimes reported to indicate successful resuscitations but may not be a meaningful endpoint. The number of patients who achieve ROSC after cardiac arrest is greater than the number who have a longer, more meaningful period of survival. An assessment of the quality of survival should acknowledge whether the patient survives to discharge from the hospital and the presence and severity of a new neurologic deficit.

The overall survival rate from pediatric perioperative arrest is reported between 46% and 92% ( ; ; ; ; ; ; ; ). The later studies report better survival rates of 72% to 92% ( ; ; ; ). Two studies report ROSC and survival individually. A report of 27 cardiac arrests in 2008 found 85% had ROSC and 46% survival to hospital discharge ( ). A report of 42 cardiac arrests in 2016 found 95% had ROSC and 85% survival to discharge ( ). Reports of long-term sequelae from pediatric perioperative cardiac arrest are limited. A 2007 report from the POCA data found 61% of children had no neurologic injury after anesthesia-related cardiac arrest ( ). A 2016 review of 72 anesthesia-related cardiac arrests reported no sequelae in 88% ( ). As rates of ROSC and survival from perioperative cardiac arrest improve, hopefully, authors will include more detailed, long-term functional status of these patients.

Increases in survival from perioperative cardiac arrest have paralleled improvements in outcomes of “in-hospital” cardiac arrest (IHCA) in children. A 2019 report of survival to hospital discharge of pediatric IHCA (inclusive of both perioperative and nonperioperative cardiac arrest) was 32% for pulseless arrest and 63% for nonpulseless events. Additionally, survival from both pulseless and nonpulseless cardiopulmonary resuscitation (CPR) events significantly increased over the time of the study ( ). This is an increase from previous reports of pediatric IHCA survival to discharge of 8% to 42% ( ; ; ; ; ). The survival rates for children after a perioperative cardiac arrest, 72% to 92%, are higher than for IHCA, 32% to 63%. Potential explanations for a higher resuscitation rate from perioperative cardiac arrest include the resuscitation skills and training of the anesthesiologist, advanced preparation for emergencies, reversible causes of anesthesia-related cardiac arrest, and continuous hemodynamic and respiratory monitoring during anesthesia that provides early recognition of problems.

Several factors affect the likelihood of successful resuscitation and survival. Mortality is increased if the etiology of the cardiac arrest is hemorrhage or is associated with protracted hypotension ( ; ). Perioperative cardiac arrest occurring during nights or weekends has a doubling of mortality ( ). Patient comorbidities, as indicated by the ASA or emergency status, increase the risk of mortality from perioperative cardiac arrest ( ; ; ). Patients with underlying cardiac disease have a higher mortality rate after cardiac arrest than patients without (33% vs. 23%) ( ). The highest mortality is in cardiac patients with aortic stenosis, cardiomyopathy, and single-ventricle physiology prior to superior cavopulmonary anastomosis. These three groups accounted for over 75% of the deaths reported in children with congenital or acquired heart disease. In addition, more than half of patients with heart disease arrested in the general operating room (54%) as opposed to the cardiac operating room (26%) or cardiac catheterization laboratory (17%). Finally, factors related to CPR can affect outcome; these include the presenting cardiac rhythm, duration of resuscitation, and duration of “no-flow” and “low-flow” states during arrest and resuscitation.

In summary, the incidence of pediatric perioperative cardiac arrest is decreasing. The understanding of risk factors for both perioperative arrest and outcome has improved. Still needed are better definitions and more uniform reporting of survival parameters and the exploration of long-term outcomes beyond survival.

Recognizing the need for cardiopulmonary resuscitation

The primary goals of CPR are to minimize the no-flow time associated with cardiac arrest and maximize oxygen delivery to vital organs during the low-flow period of resuscitation. The potential for injury from low-flow or no-flow intervals can be reduced by the early recognition that a child’s vital signs are inadequate and the rapid initiation of quality CPR. It is difficult to give guidelines for the limit of each vital sign at which vital organ blood perfusion becomes inadequate for every child under anesthesia ( Table 57.3 ). These limits depend on many factors that include the patient’s general health and age, the type and depth of anesthetic, and the intensity and duration of vital sign decompensation. Ensuring adequate end-organ perfusion requires an understanding of the metabolic supply and demand of the vital organs in children of various ages and development. Advanced pediatric training and experience are valuable in these infrequent but critical situations to help practitioners recognize the threat of injury and rapidly initiate CPR.

CPR, including chest compressions, should be initiated when the anesthesia provider believes that perfusion is inadequate to deliver oxygen, substrates, or resuscitative medications to the heart or brain. The extensive intraoperative monitoring and continuous presence of anesthetic personnel are optimal for early detection of inadequate perfusion or ventilation in the operating room. In the absence of adequate monitoring, healthcare personnel can detect an abnormal heart rate or pulse quality by palpating the umbilical artery in the newly born, the brachial artery in the infant, and the carotid artery in the child. The analysis of anesthetized and slightly hypotensive (systolic pressure <70 mm Hg) infants revealed that detection of a pulse within 10 seconds was best with auscultation; the success of detection of a pulse using femoral, brachial, or carotid palpation varied with the training of the provider ( ; ). A reliable pulse cannot usually be detected when the systolic pressure is <50 mm Hg.

In the operating room, monitoring is readily available to help determine vital signs of an anesthetized child. When the monitoring is unavailable or the readings are in question, having one rescuer auscultate and another palpate may increase reliability and decrease the time needed to count a heart rate and determine the palpability of a pulse. The administration of anesthesia interferes with the utility of typical indicators for chest compressions such as unresponsiveness or apnea. Significant bradycardia by auscultation or lack of pulse by palpation may be valid indicators to start CPR in the operating room.

Physiology of cardiopulmonary resuscitation: Reestablishment of ventilation

Mechanisms of blood flow during cardiopulmonary resuscitation

proposed that external chest compressions squeeze the heart between the sternum and the vertebral column, forcing blood to eject. This assumption of direct cardiac compression by external chest compressions became known as the cardiac pump mechanism of blood flow during CPR. The cardiac pump mechanism proposes that the atrioventricular (AV) valves close during ventricular compression and that ventricular volume decreases during ejection of blood. During chest relaxation, ventricular pressures fall below atrial pressures, enabling the AV valves to open and the ventricles to fill. This sequence of events resembles the normal cardiac cycle and occurs with use of direct cardiac compression during open-chest CPR.

Several observations of hemodynamics during external chest compressions are inconsistent with the cardiac pump mechanism for blood flow ( Table 57.5 ). First, similar elevations in arterial and venous intrathoracic pressures during closed-chest CPR suggest a generalized increase in intrathoracic pressure. Second, reconstructing thoracic integrity of patients with flail sternums improves blood pressure during chest compressions (which is unexpected, because a flail sternum should allow more direct cardiac compression during closed-chest CPR). Third, patients who develop ventricular fibrillation produce enough blood flow by repetitive coughing or deep breathing to maintain consciousness; these are situations in which no direct cardiac compression occurs but only an increase in intrathoracic pressure. These observations suggest that changes in intrathoracic pressure contribute to the production of blood flow during chest compressions. The finding that changes in intrathoracic pressure without direct cardiac compression (i.e., a cough) produce blood flow epitomizes the thoracic pump mechanism of blood flow during chest compressions. Familiarity with the thoracic pump and cardiac pump mechanisms of blood flow during CPR contributes to an understanding of how continuous monitoring of CPR effectiveness may be important in patients with changing blood flow mechanisms.

Efficacy of blood flow during CPR

The level of blood flow to vital organs produced by conventional closed-chest CPR without pharmacologic support (BLS models) is disappointingly low. The range of cerebral blood flow in animal models during CPR is 3% to 14% of prearrest levels. Cerebral perfusion pressures are also low, at 4% to 24% of prearrest levels in animals and only 21 mm Hg in humans ( ). Myocardial blood flows in this basic CPR mode are also discouragingly low at 1% to 15% of prearrest levels in animal models. In preclinical models, MPP (measured in mm Hg) correlates with myocardial blood flow (measured in mL/min/100 g brain tissue) in a one-to-one relationship. Several factors affect cerebral and myocardial blood flow during CPR, and these disappointingly low levels during BLS can be improved with the addition of pharmacologic support.

Physiologic thresholds for minimal vital organ blood flow during CPR have been described. The inability to maintain blood flow above these thresholds during CPR results in organ malfunction. A myocardial blood flow of 20 mL/min/100 g or greater is necessary for successful defibrillation. A cerebral blood flow of greater than 15 to 20 mL/min/100 g is necessary to maintain normal electrical activity during CPR. Models of BLS often do not achieve these thresholds; the addition of ALS measures, such as the administration of epinephrine, is associated with increases in blood flow to levels above these thresholds. The importance of early administration of epinephrine has been shown in several studies of pediatric in-hospital and out-of-hospital cardiac arrest, and current recommendations are for initial epinephrine administration within 5 minutes in any setting ( ).

Maintenance of circulation

The goal of CPR is to minimize the no-flow or low-flow state by restoring and maintaining the best flow possible to the brain and heart until adequate spontaneous circulation is restored. Factors related to the patient, the ventilation technique, and the compression technique will contribute to restoration and maintenance of blood flow during CPR. The pediatric anesthesiologist needs to understand how these factors affect restoration and maintenance of blood flow during an intraoperative arrest.

Distribution of blood flow during CPR

Overall blood flow to tissues is decreased during CPR compared with the normal physiologic state. A redistribution of blood flow during CPR preferentially perfuses the heart and brain. This redistribution toward vital organ perfusion should enhance outcome despite a lower overall cardiac output, as the maintenance of myocardial blood flow during CPR is necessary for ROSC and the maintenance of cerebral blood flow determines the quality of neurologic outcome.

Distribution of blood flow to both the heart and brain during CPR is influenced by the development of regional gradients. Distribution of blood flow to the brain depends on development of three regional gradients: the suprathoracic gradient, the intracranial-extracranial gradient, and a caudal-rostral cerebral gradient. The intrathoracic-suprathoracic gradient provides flow of oxygenated blood from the chest to the upper extremities and head. Either venous collapse, secondary to elevated intrathoracic pressure, or closure of anatomic valves in the jugular system prevents the transmission of intrathoracic pressure to the suprathoracic venous system. When CPR is effective, arterial collapse does not occur and elevated intrathoracic pressure results in a gradient that promotes suprathoracic blood flow. The intracranial-extracranial gradient directs blood away from extracranial suprathoracic vessels and toward intracranial vessels. Alpha-adrenergic agonists constrict extracranial vessels but have little effect on intracranial vessels, resulting in increased intracranial blood flow. Animal models show that the use of the vasoconstrictor epinephrine increases intracranial blood flow while decreasing flow in the extracranial structures of the skin, muscle, and tongue. The cerebral caudal-rostral gradient involves intracranial vessels. The relatively low-flow state of CPR seems to increase the distribution of flow to caudal areas of the brain. Ischemia preceding CPR significantly increases the distribution of flow to these areas. This pattern of caudal redistribution of flow also occurs in other models of global ischemia and provides preferential perfusion of the brainstem. Although brainstem resuscitation is necessary for survival, this propensity for sparing of caudal circulation after either prolonged ischemia or prolonged CPR raises the concern for causing a survivor with severe neurologic injury or only brainstem function.

Myocardial blood flow does not have the advantage of the large extrathoracic pressure gradient that augments cerebral flow. The thoracic pump generates equal increases in all intrathoracic structures. This lack of a gradient can result in poor myocardial blood flow during external chest compressions. Several studies have shown much lower blood flow to the myocardium than to the cerebrum during external chest compressions. The type of CPR influences the production of myocardial blood flow. Methods that are more likely to cause direct cardiac compression, such as high-impulse or open-chest CPR, result in increased myocardial blood flow. Myocardial blood flow may be present only during relaxation of chest compression, correlating with a “diastolic” pressure, or in other methods seen during compressions correlating with a “systolic” pressure. Regional flow within the heart also changes during CPR, with a shift in the ratio of subendocardial:subepicardial blood flow from the normal 1.5:1 to 0.8:1. This ratio reverts to normal with epinephrine administration during resuscitation.

Blood flow to other organs during CPR is very much reduced compared with that in the brain and heart. The lack of valves in infrathoracic veins causes retrograde transmission of venous pressure and decreases the gradient for blood flow below the diaphragm in animals. Regional blood flows for infrathoracic organs (small intestine, pancreas, liver, kidney, and spleen) during CPR are usually less than 20% of prearrest rates and often close to zero. Administration of epinephrine during closed-chest CPR almost eliminates flow to the subdiaphragmatic organs (the exception is the adrenal glands, which are resistant to effects of epinephrine). Little information is available regarding blood flow to the lungs during CPR. Pulmonary blood flow occurs primarily at times of low intrathoracic pressure during closed-chest CPR. High extrathoracic venous pressure builds up during compression and results in pulmonary filling during relaxation as intrathoracic pressure falls. Resuscitation methods that lower intrathoracic pressure may augment pulmonary vascular filling. Leaning on the chest during relaxation of compression or maintenance of increased ventilation pressures may prevent the fall in intrathoracic pressure between chest compressions and decrease pulmonary venous return and blood flow. ETCO ₂ measurements are an indicator of pulmonary blood flow (discussed later in this section).

Monitoring the effectiveness of resuscitative efforts

The brain and heart are the organs most likely to suffer irreversible damage if resuscitation efforts do not provide adequate blood flow and oxygen delivery. Table 57.6 lists several methods that are useful to determine whether resuscitation efforts are effective in restoring adequate perfusion to these vital organs. If resuscitation efforts are inadequate, interventions to improve effectiveness include improving performance of compressions (i.e., increasing depth, ensuring adequate rate, preventing leaning, or replacing a fatigued compressor), administering fluid to improve intravascular volume, and administering vasoconstrictors to improve vascular tone. If improvement attempts fail and resuscitation efforts are determined to be ineffective (as happens with prolonged arrest before CPR), the decision can be made that continued resuscitation is futile and that efforts should be terminated.

The level of consciousness provides feedback about resuscitation efforts and can improve if resuscitation efforts are effective. Occasionally a patient with a nonperfusing rhythm or otherwise low cardiac output will regain consciousness during effective chest compressions only to become unresponsive when compressions are held. This cycle may recur repeatedly until a perfusing rhythm is restored. Although return of consciousness is evidence of highly effective resuscitative efforts, it is rare that the level of perfusion required can be accomplished and is more unlikely under anesthesia.

The ROSC is a more common indicator of adequate resuscitative efforts. It is a sign that perfusion to the heart muscle is enough to allow effective contractions. The temptation is often to hold resuscitation efforts when spontaneous ejection occurs, but spontaneous circulation may not be adequate or sustained and continued resuscitation efforts may be required. Continue compressions if spontaneous circulation does not adequately perfuse vital organs despite ROSC.

The palpation of a pulse can indicate that chest or cardiac compressions are generating significant arterial pressure. Unfortunately, feeling “a pulse” with compressions may represent only peak arterial pressure and not the presence of an adequate diastolic (relaxation) pressure necessary for coronary perfusion. An additional concern about reliance on palpation to determine the effectiveness of resuscitation is that the palpated artery is usually next to a large vein, and retrograde venous pulsations may occur in the absence of significant arterial blood flow. Retrograde venous femoral pulsations have been shown during CPR with a cross-clamped aorta and with ultrasound attempts at femoral venous cannulation. There are no data on when (or even if) palpation of a pulse during chest compressions correlates with ROSC or outcome.

One of the most useful ways to measure the effectiveness of chest or cardiac compressions in generating blood flow is to use quantitative ETCO ₂ monitoring . Fortunately, ETCO ₂ monitoring devices are readily available in anesthetizing locations. In low cardiac output states, including during CPR, the detection of ETCO ₂ during compressions is dependent on the adequacy of compressor-generated pulmonary blood flow. The level of ETCO ₂ increases as compressions become more effective at increasing pulmonary blood flow and the delivery of CO ₂ to the lungs. ETCO ₂ has been used to guide rate, depth, duty cycle, and hand position of chest compressions in an animal model ( ) and may result in better rates of ROSC and short-term survival than standard CPR ( ). These benefits were lost in the setting of prolonged (23-minute) asphyxial arrest prior to the onset of CPR ( ).

Low levels of ETCO ₂ generated during compressions correlate with decreased levels of blood flow and decreased likelihood of ROSC ( ). During CPR, ETCO ₂ measurements that are persistently <10 mm Hg predict an unlikely ability to achieve ROSC in adults. Some past studies have suggested that levels of ETCO ₂ greater than 15 mm Hg predict ROSC in adults and children ( ; ; ). However, a meta-analysis by Hartman et al. (2014) noted an average ETCO ₂ level of 25 mm Hg in patients who achieved ROSC. This value is significantly higher than the previously recommended goal of >10 to 15 mm Hg during CPR. The 2020 AHA recommendations for pediatric resuscitation are unable to recommend a specific target for ETCO ₂ during CPR; instead, they focus on the utility of its measurement ( ). Certainly, ETCO ₂ levels <10 to 15 mm Hg during CPR suggest that ROSC is unlikely and should prompt attempts to improve CPR (e.g., better compressions, fluid, or vasoconstrictor administration). The measurement of ETCO ₂ during CPR has also been used to detect (1) low levels of cardiac output during pulseless electrical activity ( ), (2) ROSC during compressions ( ), and (3) the presence of spontaneous circulation during cardiopulmonary bypass ( ).

ETCO ₂ measurement during CPR does not require an ETT. ETCO ₂ levels measured during CPR with bag-mask or laryngeal mask ventilations also correlate with likelihood of achieving ROSC ( ). Other important considerations when monitoring ETCO ₂ levels include the following: (1) The administration of bicarbonate can cause a transient elevation in ETCO ₂ without an increase in blood flow and may be misinterpreted as improving CPR. (2) Epinephrine administration has been associated with a transient decrease in ETCO ₂ despite an increase in MPP and may be misinterpreted as a worsening of CPR. The time frame of this decrease is usually short lived (<30 seconds) ( ). (3) The etiology of arrest influences the initial ETCO ₂ levels during resuscitation. Higher initial ETCO ₂ levels are seen after asphyxial arrest than fibrillatory arrest ( ). Fortunately, this bump in ETCO ₂ during CPR after asphyxia washes out with 30 to 60 seconds of CPR. Despite these considerations, ETCO ₂ monitoring during CPR is beneficial, and its use during CPR is recommended.

If invasive monitoring is in use at the time of an arrest, arterial pressure monitoring is also helpful to determine the effectiveness of resuscitative efforts. An arterial catheter is necessary to determine aortic diastolic pressure during the relaxation phase of compressions. If an arterial catheter is present, diastolic blood pressure (DBP) during CPR can be used as a surrogate for MPP when central venous access is unavailable; levels >15 to 20 mm Hg are necessary for, but do not guarantee, ROSC in adult patients ( ). An arterial catheter is also useful for measuring and providing real-time feedback for the effective rate of compressions. A low diastolic (relaxation) pressure (<15 to 20 mm Hg) noted on an arterial catheter waveform represents a low MPP and may help guide volume and vasopressor administration ( ; ). Suggested target levels for DBP during CPR are >25 mm Hg for infants and >30 mm Hg for children ( ).

A central venous catheter, if present, can be used to determine the central venous oxygen saturation during CPR. The level of venous blood oxygen saturation during CPR correlates with blood flow and likelihood of ROSC. Much like its use in patients with septic shock, venous oxygen saturation can determine the oxygen supply and demand balance. Patients with a mixed-venous oxygen saturation <30% are unlikely to have ROSC ( ).

The presence of both arterial and venous catheters allows simultaneous sampling of gases for calculation of the venous-arterial CO ₂ difference . This difference is approximately 5 mm Hg during native circulation and increases significantly as perfusion falls. During hypoperfusion, tissue CO ₂ increases, venous CO ₂ increases, pulmonary blood flow falls, and normal ventilation removes a greater percentage of CO ₂ from the reduced pulmonary blood flow and results in lower arterial CO ₂ . This increase in PvCO ₂ and decrease in Pa co ₂ creates a venous-arterial difference that is inversely proportional to the blood flow produced by CPR. This situation may make a venous blood gas more reflective of the actual acid-base status than an arterial sample. As the effectiveness of CPR improves, this venous-arterial gradient decreases.

The amplitude and frequency of ventricular fibrillation (VF) can be determined from the electrocardiogram (ECG) or by specific software built into AEDs and monitored to determine effectiveness of resuscitation. Typically, the VF waveform is coarse initially (high amplitude, low frequency) and deteriorates over time during ineffective CPR or prolonged arrest to fine VF (low amplitude, high frequency). As heart perfusion during CPR improves, VF reverts to a coarse pattern that indicates an improved metabolic state and that the heart is better primed for successful defibrillation. An index of amplitude and frequency, the amplitude spectral area (AMSA) has been correlated with the MPP, ETCO ₂ , and likelihood of ROSC in an animal model ( ). AMSA is calculated by software in the AED, and a value of ≥13 mV Hz has been regarded as a critical threshold for defibrillation success. VF waveforms may be used to guide CPR in this respect. One would expect a fine VF waveform to become coarser as effective chest compressions are delivered. If the VF waveform deteriorates, it may be helpful to reevaluate the effectiveness of chest compressions (depth, rate, duty cycle, hand position, etc.). Some newer-model defibrillators contain software that helps with rhythm analysis by extracting CPR artifact from the ECG waveform. The filtered signal allows continuous ECG analysis, avoids pauses for rhythm check, and improves preparation for defibrillation.

Transthoracic impedance monitoring (TTI) is possible during CPR using the gel-coated pads applied to the chest for defibrillation. The defibrillator software analyzes TTI continuously and provides feedback about the decrease in impedance related to increased blood movement through the chest during compressions. This technique can monitor the amount of blood flow generated by compressions and determine whether compressions continue to be effective or deteriorate with rescuer fatigue ( ). The amplitude change of TTI correlates with compression depth and MPP ( ). TTI has been used to detect ROSC in adults. Like a sudden spike in ETCO ₂ during CPR, changes in the TTI waveforms can signify ROSC. These methods can lead to earlier detection of ROSC and eliminate the no-flow period during a pulse check ( ; ).

An additional benefit of monitoring TTI during CPR is the ability to detect the presence or loss of ventilation. Gas entry into the lungs with ventilation causes an increase in TTI. The sudden loss of the cyclical changes in TTI with ventilation may indicate displacement of the ETT during CPR ( ). Although loss of ETCO ₂ is also used to indicate displacement of the ETT, ETCO ₂ levels may be markedly diminished by low pulmonary blood flow during CPR. The disappearance of impedance changes related to the failure of air entry with ETT displacement may be more reliable in situations of extremely low ETCO ₂ . TTI might eliminate the need to interrupt compressions to perform auscultation when ETT displacement is suspected.

Many methods are available to the anesthesiologist to determine the effectiveness of intraoperative resuscitation efforts. The equipment for quantitative determination of ETCO ₂ and arterial diastolic pressure monitoring is most likely to be available in perioperative situations.

Vascular access for medication and fluid administration

Intraosseous access

IO cannulation provides a rapid and safe route to vascular access via the bone marrow; this space is a noncompressible venous plexus and is reliably available when peripheral venous access is limited from dehydration or peripheral vasoconstriction. Trained providers can obtain IO access in 30 to 60 seconds, with a first-attempt success rate of 80%. All drugs, crystalloid, colloids, and blood can be administered via this route. The onset and duration of action of emergency medications are the same when given by IO, central access, or peripheral access during native circulation.

The preferred site for placement of an IO needle in a child is the anterior tibia. Alternative sites include the distal femur, medial malleolus, and iliac crest. In older children and adults, the distal radius, distal ulna, proximal humerus, and sternum (risk of cardiac laceration) are also considered appropriate sites ( Fig. 57.3 ). Specially designed IO needles should be available to the pediatric anesthesiologist for such emergencies. Rapid deployment devices for IO needles have been developed and may increase ease of IO placement ( ; and see Chapter 40 : Anesthesia for Pediatric Trauma, Figs. 40.6 and 40.7 A and B). The most common complication from IO access is displacement of the needle and extravasation of fluid and medication (12%). Other rare complications include bone fracture, compartment syndrome, osteomyelitis, and fat embolism.

Vasoactive medications