Part 2: Stabilization and Management of the Acutely Ill Infant and Child

INTRODUCTION

Children with serious illness or injury with the potential for rapid deterioration invariably require close observation to detect changes in function or state. Careful bedside assessment provides the basis for appropriate triage and early awareness of clinical deterioration. Electronic monitoring complements this assessment by providing (1) repetitive or continuous assessment that does not disturb the patient, (2) a means for detecting the effectiveness of interventions, and (3) warning signals for physiological disturbances that permit staff to observe multiple patients simultaneously. The ideal monitoring system has been described by many as one that is accurate, able to predict deterioration, reproducible with a rapid response, operator independent, easy to use, safe, continuous, inexpensive, and integrative with other medical systems. To date, no such monitoring currently exists in clinical use.

RESPIRATORY MONITORING

Respiratory Rate

The respiratory rate (RR) and tidal volume (V_T; volume of each breath) together determine minute ventilation (V̇_E; volume of air movement into and out of the lungs in 1 minute). Minute ventilation is product of RR × V_T. Since arterial blood carbon dioxide levels (PaCO₂) are inversely proportional to V̇_E, decreases in RR and/or V_T will lead to increases in PaCO₂. Therefore, determination of RR is an essential component for the evaluation of respiratory function.

There are several methods for determining RR, ranging from counting visual chest movement and auscultated breaths to using devices that measure changes in electrical bioimpedance, air movement at the nares, or inductance plethysmography. These methods each have advantages and limitations but none measure carbon dioxide (CO₂) levels. Auscultation provides the best estimation of RR and depth of breathing. It provides the practitioner the opportunity to listen for air movement within the lungs and to differentiate between problems related to airway obstruction (upper vs lower) and problems with poor lung/chest wall compliance. The drawback is that auscultation is time consuming and a static measure.

Transthoracic impedance monitoring is the technique most commonly used to continuously measure respiratory rate. A small current is passed between 2 electrodes across the chest (similar to those used for electrocardiography). During inhalation and chest rise, the electrodes move further apart, increasing the impedance between them. During exhalation, the electrodes move closer together and the impedance decreases. The sinusoidal changes in impedance are displayed as the RR. Transthoracic impedance monitoring is safe, relatively cheap, and readily available to all patients. It is important, however, to recognize its limitations. First and foremost, the interpreted signal is dependent on chest movement and not breathing, per se. Therefore, chest movement unrelated to breathing may be mistaken for breathing, overestimating RR. Next, chest movement may negate impedance measurements and also may result in the underestimation of RR. Finally, transthoracic impedance monitoring does not accurately measure the depth of breathing nor does it rely on actual air flow. Thus, patients who have airway obstruction may have significant chest movement but little to no ventilation of the lungs. This may lead to under-recognition of airway obstruction.

Other methods of monitoring RR are respiratory inductance plethysmography, airflow thermistor, and capnography (discussed below). Respiratory inductance plethysmography tends to be used in specialized settings, such as during polysomnography (sleep studies). This method utilizes elastic bands that encircle the chest and abdomen, and assumes that the thoracic cavity is a cylinder that expands and constricts circumferentially during inhalation and exhalation, respectively. Electrodes within the bands measure the changes in circumference of the chest and abdomen in order to determine a RR as well as to estimate a tidal volume. When used in conjunction with other measures of respiration, inductance plethysmography can be useful to differentiate different disorders of breathing, such as central versus obstructive sleep apnea. Caveats to using this device are the relatively specialized equipment needed for accurate measurement, correct placement of the bands, and tolerability of wearing the device.

Respiratory thermistors are devices that detect changes in air temperature at the nares in order to estimate RR. The thermistor is placed at the patient’s nares and detects a difference between the warm gas breathed out during exhalation versus the cooler inhaled air. The changes in temperature are then used to determine RR. The relative changes in temperature do not relate to the volume of each breath, however; thus this device can only detect RR. Limitations in its use include inaccuracies during mouth breathing and underestimation of hypopnea.

Capnography

V̇_E is indirectly proportional to PaCO₂. Although RR is important, it is often inadequate to assess ventilation due to the inability to accurately take into account depth of breathing and physiologic changes within the lung including “dead space” ventilation (V̇_D) which refers to air movement without the exchange of gases between the lung and the blood. Therefore, during situations in which accurate measurements of ventilation are desired, measurement of CO₂ levels may be necessary. The gold standard for measurement of ventilation is the arterial blood gas (discussed below). Although it is the truest measure of ventilation, it is invasive and is usually not continuous. In situations where continuous monitoring is necessary, capnography has become a standard of care, especially in the emergency department, intensive care units (ICUs), and operating rooms, and during procedural sedation.

Quantitative capnography is based on the ability of CO₂ to absorb infrared light. When gas is sampled within the path of airflow (mainstream) or from the side (sidestream), it passes through a detector that emits infrared light and detects the absorbance by CO₂. Alternatively, some devices use spectrometry to detect CO₂ levels. As the detector measures the exhaled CO₂ (P_ECO2), a tracing appears on the monitor (Fig. 99-1); this is the standard time-based capnograph. During a respiratory cycle, the partial pressure of carbon dioxide (PCO₂) of inhaled air is 0 mm Hg, accounting for the minimal amount of atmospheric CO₂ in relation to nitrogen and oxygen. This air enters the trachea, bronchi, and then the respiratory lung units (alveoli and respiratory bronchioles). Carbon dioxide quickly diffuses from the blood to the respiratory lung units, but not into the dead space. This dead space ventilation can be areas of normal, anatomic dead space (ie, trachea, mainstem bronchi, and nonrespiratory bronchioles) or pathologic areas of the lung that are poorly perfused (eg, pulmonary embolus and during cardiac arrest). During exhalation, the first volume of gas is from the anatomic dead space (trachea and bronchi); therefore, the P_ECO2 is 0 mm Hg (phase 1). As gas moves from the respiratory lung units into the larger airways, the P_ECO2 increases (phase 2). In normal lungs, this increase in expired CO₂ is rapid. In the latter phase of exhalation, the P_ECO2 plateaus (phase 3), resulting in the typical square-shaped waveform. At the end of exhalation, the end-tidal CO₂ (P_ETCO2) is quantified and displayed. As inhalation begins again (phase 4), the P_ECO2 quickly falls due to the movement of atmospheric CO₂ through the detector. Therefore, based on the capnograph waveform and P_ETCO2 value, the adequacy of ventilation can be continuously assessed in a reliable fashion.

Quantitative time-based capnography is used in situations where the adequacy of ventilation needs to be continuously monitored. The most common scenario includes one in which there may be disordered control of breathing, such as during procedural sedation and anesthesia or if patients have depressed sensorium or neuromuscular weakness. Additionally, capnography should routinely be used as a safety device in endotracheally intubated patients since disconnection from the ventilator, accidental extubation, or other mechanical problems will be detected by a loss of capnographic signal. To that end, qualitative (or colorimetric) capnography is used to quickly confirm endotracheal tube placement since the litmus paper within the device changes color when the expired CO₂ reaches a certain threshold. Capnography is also useful for making adjustments to a patient’s ventilator settings, especially if rapid titration is necessary (such as to manage increased intracranial pressure) or if a blood gas measurement is not possible or warranted. Finally, an increasing recommendation is the use of quantitative capnography to assess the quality of chest compressions during cardiopulmonary resuscitation. Since the P_ETCO2 depends on blood flow through the lungs in addition to air flow, during cardiac arrest, P_ETCO2 will drop toward 0 due to lack of pulmonary blood flow. As chest compressions are instituted, P_ETCO2 will start to rise. As the effectiveness of chest compressions improve, the P_ETCO2 should also increase, given the same V̇_E, due to improved pulmonary blood flow (Fig. 99-1E).

The utility of quantitative capnography lies not only in the P_ETCO2 value, but also with the shape of the waveform. Typically, the P_ETCO2 value is less than the PaCO₂ (~0–5 mm Hg) in normal lungs, owing to the anatomic dead space. In patients with higher degrees of lung disease, the difference between PaCO₂ and P_ETCO2 (dead space ventilation; V̇_D) increases and can be followed to determine disease progression in critically ill patients, especially if they are tracheally intubated with a cuffed endotracheal tube. One of the most useful attributes of the capnographic waveforms is the ability to detect the degree of airway obstruction and bronchospasm. In patients with bronchospasm, CO₂ from the alveoli enters the large airways at variable rates due to increased resistance in the small airways. This is manifested as a slow, upward ramping of the capnographic waveform instead of the normal square waveform. If the obstruction is severe enough, the P_ETCO2 may never reach the true level of the PaCO₂ since inhalation begins prior to complete exhalation.

Capnography, like other devices, has its limitations and caveats. In nonintubated patients, capnography typically uses sidestream detectors via nasal cannula. In these cases, a decrease in P_ETCO2 may be caused by other environmental factors, such as dislodgement of the cannula, mouth breathing, or entrainment of atmospheric air (such as from the concomitant use of a high-flow nasal cannula for oxygenation). Additionally, appropriate interpretation of the capnographic waveform and P_ETCO2 can be confusing, at times. As an example, decreased ventilation will increase P_ETCO2, whereas decreased perfusion to areas of the lungs will decrease P_ETCO2. If a patient has both decreased ventilation and perfusion, then it will be difficult to interpret the true nature of the findings. The interpretation of changes in mainstream quantitative time-based capnography also assumes consistent (or no) breath-to-breath leaks in the system.

Pulse Oximetry

Pulse oximetry is a mainstay in the monitoring of oxygen saturation. To recognize the benefits and limitations of pulse oximetry, it is important to understand the principles on which it is designed. Pulse oximetry depends on light absorption in a tissue bed, depending on the relative ratio of oxyhemoglobin (HbO₂) and deoxyhemoglobin (Hb) at 2 different wavelengths of light (Fig. 99-2). At red wavelengths, Hb has a higher light absorption than HbO₂; whereas at infrared wavelengths, HbO₂ has a higher light absorption than Hb. For standard pulse oximeters, light in the red and infrared wavelengths is emitted from a light emitting diode (LED) and travels through a tissue bed (eg, finger, earlobe, toe) to a detector on the opposite side. The ratio of the emitted (known) and the detected (measured) light intensities is calculated for each wavelength to determine the relative level of absorbance of each. Based on correlation curves using blood O₂ saturation levels from healthy human subjects, the O₂ saturation of the tissue is estimated. To estimate the arterial saturation, the pulse oximeter uses the pulsatility of blood in the tissue to its advantage. During systole, more blood is in the arterial system; therefore, more light is absorbed as there is more Hb and HbO₂ in the path of light. During diastole, there is less blood to absorb light in the pathway; therefore, the absorption of light is caused by the combination of tissue, venous blood, and non-pulsatile arterial blood. The difference in absorption during systole and diastole is then assumed to be caused by the contribution of arterial blood alone. The device then displays the oxygen saturation of pulsatile blood (SpO₂).

An understanding of the basis of pulse oximetry helps clinicians to recognize its benefits and limitations as well as how to troubleshoot problems. Clearly, the main benefit of pulse oximetry is the ability to estimate arterial O₂ saturation in a continuous and noninvasive manner. For the vast majority of patients, SpO₂ measurements have replaced arterial oxygen saturation (SaO₂) measurements. Although newer devices have improved the reliability of SpO₂ measurements, there are still limitations to the accuracy of current pulse oximeters. It is important to remember that pulse oximeters average the signal over the preceding 20 beats; therefore, the displayed SpO₂ is actually reflective of changes that occurred up to 20 cardiac cycles ago. This is important during critical periods, such as during resuscitation or intubation. Since pulse oximetry relies on the device to distinguish between pulsatile blood and non-pulsatile blood, there are a number of conditions that may reduce the accuracy and/or reliability of measurement: patients with poor perfusion to the extremity used for pulse oximetry (eg, in severe shock, arterial occlusion, or hypothermia), excessive movement, and venous pulsation (eg, in severe tricuspid regurgitation). Since standard pulse oximeters typically rely on only the red and infrared wavelengths of light, certain dyshemoglobinemias, such as carboxyhemoglobinemia (overestimates SpO₂) and methemoglobinemia (overestimates SpO₂ and trends toward 85% as MetHb levels rise), may result in inaccurate measurements. In these instances, co-oximeters, which utilize multiple wavelengths of light, should be used in place of pulse oximeters. Finally, excessive ambient light interference or alterations in absorption (eg, from fingernail polish, dyes, pigments) can affect the accuracy of measurements. Another caveat is that there is a lower limit for accurate SpO₂ measurements with most devices. At lower O₂ saturations (below 80–85%), SpO₂ readings are considered to reflect a trend rather than an accurate saturation. This is of particular relevance in patients with cyanotic congenital heart disease, or in the setting of severe hypoxemic respiratory failure, when permissive hypoxia is a part of routine intensive care.

Blood Gas Measurement

As described above, the gold standard to determine oxygenation and ventilation in patients is through the measurement of arterial blood gases (ABG), which characterizes efficiency of gas exchange and also disturbances in the metabolic and respiratory control of acid–base balance.

All modern blood gas analyzers take 4 measurements: sample temperature, pH, PO₂, and PCO₂. The pH, PO₂, and PCO₂ are directly measured, whereas most analyzers calculate the following using various equations and nomograms: bicarbonate (HCO₃^-), O₂ saturation (SO₂), and acid–base excess (ABE). This is important to take into consideration when there is a need for exact SO₂ values, such as in the presence of dyshemoglobinemias and when used to estimate cardiac output using the Fick principle. It is also important to understand the different areas of sampling available for blood gas determination. Although arterial sampling is the gold standard, blood gases can be obtained from the venous and capillary vessels, especially for outpatient monitoring of ventilation, but they cannot be used as a guide to adequacy of oxygenation. Blood venous samples tend to be approximately 0.05 to 0.1 pH units lower and have PCO₂ levels approximately 5 mm Hg higher than simultaneous arterial samples owing to the diffusion of CO₂ from the tissues into the capillary beds prior to returning to the venous system. In patients with poor perfusion due to shock, hypothermia, or the use of a tourniquet, these discrepancies may be greater. Therefore, when obtaining samples from these areas, it is important that the blood is free-flowing without the use of a tourniquet because restricted blood flow during sampling may artificially decrease the pH and increase the PCO₂. Venous PO₂ levels are a reflection of cellular O₂ uptake, and high levels may be indicative of impairment in mitrochondrial O₂ uptake from processes such as cyanide toxicity. The utility of capillary samples is entirely questionable, and many practitioners believe they do not have any role in blood gas monitoring. Finally, when obtaining any blood sample for blood gas analysis, it is important to ensure that there are no air bubbles in contact with the sample as this will affect the measurements, due to diffusion of gas between the blood and air.

The arterial pH (normal 7.35–7.45) reflects the acid-base balance within the system. If the arterial pH is less than 7.35, the blood is deemed acidemic. If the arterial pH is greater than 7.45, the blood is alkalemic. The 2 main determinants of pH in the blood are the PCO₂ (normal 35–45 mm Hg; a volatile acid released by the lungs) and serum bicarbonate levels (HCO₃^-; normal 22–26 mM beyond early postnatal life; primary buffering system), which are related to each other by the following chemical reaction:

and are related to pH by the Henderson-Hasselbalch equation:

Other determinants are the levels of various proteins, ions, and other weak buffers within the blood and are out of the scope of this chapter. Since PCO₂ is a volatile acid that is exhaled by the lungs, as PCO₂ increases, the blood becomes more acidic and pH falls. For acute changes, if PCO₂ changes by 10 mm Hg, pH will change in the opposite direction by 0.08. Conversely, changes in HCO₃^-, a base, result in changes in pH in the same direction, such that an increase in HCO₃^- by 10 mM increases the pH by 0.15. Of note, the HCO₃^- result on standard blood gas analyzers is a calculated result based on the Henderson-Hasselbalch equation and the value should be compared to the serum total CO₂ (tCO₂; normal 23–30 mM) that is measured by serum chemistry. Taking these factors into account, basic acid–base disorders can be determined by proceeding in a step-wise fashion (Fig. 99-3), with more complex evaluations discussed in other resources.

The other component use of the arterial blood gas analysis is PaO₂ (normal 75–100 mm Hg). Although the majority of oxygen is carried by hemoglobin in the erythrocytes and reliably monitored using pulse oximetry, determination of PaO₂ is useful in patients with cyanotic heart disease in which SpO₂ is lower and less reliable, as well as when differentiating low SpO₂ in neonates using the hyperoxia test. Finally, PaO₂ levels are useful for characterizing the degree of lung injury in acute lung injury and acute respiratory distress syndrome for prognosis and research.

CARDIOVASCULAR MONITORING

Heart Rate and Rhythm

The heart rate and rhythm and their trends over a finite period of time are a fundamental part of the evaluation and treatment of critically ill patients. Electrocardiograms (ECGs) can be monitored continuously from electrodes placed on the chest, and a cardiotachometer determines heart rate from the frequency of the QRS signal. There are important sources of error to consider with these methods, however. When the signal is large, the T wave and QRS may both be detected, and the cardiotachometer may show a value that is twice normal. Failure to detect a signal can occur when the QRS amplitude changes—for example, with respiration, repositioning of the leads, or changes in posture. It is quite easy to misinterpret rhythm disturbances or misdiagnosis ST-segment abnormalities when reviewing a single lead rhythm or rhythm strip. Whenever a dysrhythmia or ECG abnormality is suspected based on the monitor, a 12 to 15-lead ECG should be obtained for true diagnostic purposes. Finally, one must remember that ECG only measures electrical activity and not actual cardiac muscle contractions, which is an important distinction in patients with pulseless electrical activity (PEA). Therefore, a pulse check is imperative in these patients to assess their systemic perfusion.

Bedside arrhythmia analysis and bed-to-bed switching (so that a central monitoring station was not required) allowed widespread use of electrocardiographic monitoring in adult ICUs in the 1960s and 1970s. Widespread adoption in pediatric ICUs did not occur until the late 1980s with the advent of modular units that allowed input from various devices. At this time, the potential for arrhythmia in postoperative cardiac surgical patients and other organ dysfunction states was becoming appreciated. Focused attention to ST-segment changes has been commonplace after operations that involve manipulation or reimplantation of the coronary origins such as the arterial switch operation and the Ross operation. The theoretical advantage of automated ST-segment analysis is balanced by the limitations of algorithms used by currently available software for this purpose in the pediatric populations.

Blood Pressure

Noninvasive Blood Pressure

Most commonly, noninvasive blood pressure assessment is performed using intermittent methods. Determining pressure using a sphygmomanometer with a cuff that occupies about two-thirds of the length and almost completely (80–100%) encircles the circumference of the limb segment remains the standard approach for initial patient assessment. The cuff is inflated to a pressure well above that occluding the arterial systolic pulse and is gradually deflated while listening for Korotkoff sounds. The first sound corresponds to arterial systolic pressure, and the muffling of the second sound corresponds to the arterial diastolic pressure. The systolic pressure can also be determined by palpating the artery distal to the cuff and noting the pressure at which the pulse is first felt when the cuff is deflated.

Alternatively, most automated blood pressure cuffs utilize oscillometry to determine blood pressure. These devices detect pressure changes caused by turbulence in arterial blood flow as the cuff is deflated. These devices record the peak oscillations as the mean arterial blood pressures (MABP) and then extrapolate the systolic and diastolic blood pressure measurements. The most important factor that determines the accuracy of these oscillometric techniques in measuring MABP is cuff width–to–arm circumference ratio which ideally should be in the range of 0.44 to 0.55. A smaller cuff will overestimate the true MABP. It is also important to note that the mean bias of noninvasive methods become unreliable at both the higher and lower limits of physiologic blood pressure measured invasively (ie, severe hypertension or clinical shock, respectively).

Principles of Invasive Pressure Monitoring

Briefly, once a vascular compartment of interest has been accessed by a catheter (eg, arterial, central venous, left atrial), the pulsations generated by the space are transmitted via a continuous column of fluid to a transducer that converts the analog signal into a digital one. When setting up a direct pressure monitoring system, it is important to ensure that the system has a baseline reference value; this is usually set to atmospheric pressure at the anatomic level of interest (such as the right atrium for vascular pressures). Other facets of calibration, such as damping, are important but out of the scope of this text.

Central Venous Pressures and Atrial Pressures

Central venous pressure (CVP) is used as a proxy for central vascular volume. In general, as the volume in the vasculature increases, the CVP rises. More accurately, the measured CVP reflects the back pressure of venous return from the peripheries (this driving pressure from the peripheries is known as mean circulatory filling pressure). Changes in intravascular volume status, ventricular compliance, or intrathoracic pressure will affect the value of the CVP. The pressure from a catheter in the right atrium or a large central vein is used to monitor cardiac filling or, more precisely, right ventricular preload. Many assumptions are made when using CVP as a measure of intravascular volume, including lack of downstream obstruction, normal lung compliance, and a consistent relationship between pressure and volume (known as ventricular elastance). A more accurate measure of this is the left atrial pressure, which can be directly measured with a left atrial line, or indirectly using a Swan Ganz catheter with measurement of pulmonary artery wedge pressure. Left atrial lines are still placed in selected patients undergoing open heart surgery. However, the use of Swan Ganz catheters in pediatric patients, is now at most, rare and, in many units, obsolete.

With these constraints in mind, CVP can serve as a useful guide for determining how interventions are affecting cardiac filling in patients in whom responses are not very predictable (eg, the child with congestive circulatory failure in whom positive pressure ventilation is initiated or adjusted) or in patients in whom there is a narrow margin of safety (eg, the child with elevated intracranial pressure and low cardiac output). Normal values for CVP range from –3 to +3 mm Hg in the infant to 5 to 10 mm Hg in the child and young adult. These values, however, can be greatly affected by the use of positive-pressure breathing. Accordingly, it is usually more useful to judge relative changes rather than to rely on absolute targets for therapy. In addition to pressure monitoring, additional information regarding the adequacy of systemic oxygen delivery is provided by central venous saturations (ScvO₂) drawn from an appropriately positioned central catheter with the tip in the superior vena cava–right atrial junction and without the confounding effects of a significant left-to=right atrial level shunt. By using the oxygen extraction ratio (OER = [SaO₂ − ScvO₂/SaO₂]; normal 0.2–0.3), an estimation of adequacy of cardiac output can be made, with OER > 0.3 as concerning for inadequate cardiac output for the current oxygen consumption state.

Continuous venous pressure tracings, and the appearance of its component a, c, and v waves, can also provide useful information as to nature of the underlying cardiac rhythm and changes therein (Fig. 99-4).

Arterial Pressures

The use of invasive arterial pressure monitoring is reserved for intraoperative, perioperative, or intensive care monitoring. The most common reasons for arterial line placement are for frequent blood gas measurement in patients with respiratory failure, for frequent blood sampling in unstable patients, and the need to monitor hemodynamic changes during periods of critical illness and in response to interventions and titrations of therapies. The choice of location of invasive arterial pressure monitoring is based on ease of access and limitation of complications, with important exceptions. These include not attempting catheterization of an artery that is known to have been completely or partially occluded, or in a circulation where there may already be compromised flow (a systemic-to-pulmonary shunt affecting flow to that arm, or dialysis fistula) or poor potential collateralization. In the case of surgery on the aortic arch, a right arm arterial line (typically radial) is used to guide cerebral perfusion both prior to and during cardiac surgery. In smaller infants and neonates, especially in the context of shock, cannulating peripheral arteries can be challenging. In these situations, the femoral artery may be the next choice. As a rule, the longer term consequences of vascular occlusion should always be considered when deciding on the optimal location of an arterial line, and the operator performing the procedure.

It is important to recognize that there are regional variations in blood pressure and that these can be accentuated in pathological states. Normally, mean blood pressures are equal throughout the large arteries, although systolic pressure is higher and diastolic is lower when comparing the large arteries of the legs to the brachial or radial artery. However, in states of low cardiac output, the systolic and mean blood pressures from a large central artery often are higher than values from more peripheral arteries depending on the degree of impairment of the perfusion, and peripheral vasoconstriction. The central systolic blood pressure may be sustained near normal levels even in the presence of shock, owing to severe vasoconstriction of peripheral arteries and arterioles.

SUGGESTED READINGS

INTRODUCTION

Rapid establishment of vascular access is necessary for aggressive fluid resuscitation and administration of medications such as catecholamines, antibiotics, narcotics, and sedatives during emergencies. However, attaining vascular access in a child during a life-threatening illness is difficult and often consumes precious time. An organized approach to vascular access can minimize this potentially life-threatening delay in treatment. This chapter discusses the priorities in vascular access during emergent, urgent, and stable situations. It also reviews various techniques for achieving vascular access and covers the relative indications and potential complications.

PRIORITIES OF VASCULAR ACCESS

Time is critical when attaining vascular access in life-threatening emergencies such as cardiopulmonary arrest or shock. Of course, any pre-existing intravenous catheter should be utilized in the initial resuscitation efforts, regardless of how small such a catheter might be. If such access is not available during a life-threatening emergency, intraosseous access should be attained as rapidly as possible, especially in children under 6 years old. A practical approach is to pursue this and peripheral venous access simultaneously. Time should not be wasted waiting for attempts at peripheral venous catheterization before attempting intraosseous access, which can be attained more rapidly and more reliably. Similarly, skilled clinicians may attempt placing central venous catheters during life-threatening emergencies, but such attempts should not preclude simultaneous attempts at intraosseous access. After attaining intraosseous access for initial fluid resuscitation and infusion of emergency medications, peripheral or central venous catheterization is the next priority in order to ensure a more reliable, long-lasting vascular access.

For urgent situations, such as fluid resuscitation of a child with compensated shock or dehydration, the risk–benefit ratio shifts. Generally, it is most appropriate to initially insert a peripheral venous over-the-needle catheter. If multiple attempts are unsuccessful, or if the child requires fluids or medications that cannot be given safely in a peripheral vein, central venous catheterization should be attempted by a qualified individual. Of course, if the child’s clinical condition deteriorates prior to achieving vascular access, priorities should be reassessed and it may be necessary to attain intraosseous access.

Relatively stable children may need vascular access for maintenance fluids or intravenous medications. Generally, peripheral venous cannulation with an over-the-needle catheter is adequate. If vascular access is necessary for more than 2 to 3 weeks, or if solutions to be infused can cause serious tissue injury if extravasated, a central venous catheter may be necessary. However, central venous catheterization entails added risks, and its inherent risks and benefits deserve consideration (Table 100-1).

PERCUTANEOUS PERIPHERAL VEIN CANNULATION

Small-caliber plastic catheters have allowed for increasingly easy and reliable peripheral venous cannulation. Small over-the-needle catheters (22–24 gauge) are available to cannulate even the small veins of the hand, foot, or scalp of neonates and premature babies. Such peripheral venous catheters are generally all that are necessary for short-term delivery of intravenous fluids or medications.

The most common complication of peripheral venous cannulation is catheter displacement and infiltration of the tissues with the infusing fluid. Most intravenous solutions are isotonic or hypotonic and easily absorbed from the tissues. Infiltration with these solutions is generally treated by removing the catheter and elevating the limb. However, solutions that are very hypertonic (eg, those containing greater than 12.5% dextrose, 3% sodium chloride, or 8.4% sodium bicarbonate) or that contain irritating substances such as calcium or potassium salts, some antibiotics (eg, erythromycin), or medications with an extreme pH (eg, phenytoin) can cause substantial tissue injury leading to necrosis of the skin or subcutaneous tissues. Importantly, extravasation of vasoconstricting agents such as dopamine, epinephrine, and norepinephrine can result in profound local vasoconstriction and subsequent substantial tissue injury. Even without extravasation of the previously noted medications and fluids, local vascular injury can result in aseptic thrombophlebitis. This condition may serve as a nidus for suppurative thrombophlebitis. The risks of aseptic and suppurative thrombophlebitis increase with the duration of the indwelling catheter.

Selection of catheter size and peripheral venous site are important issues. For a patient in shock, the widest and shortest catheter is optimal, because longer, narrower catheters result in more resistance to flow. For the trauma victim, it is preferable to insert the catheter into an uninjured limb or at least a limb apparently free from major vascular trauma. The greater saphenous vein, median cubital vein, and external jugular vein are often used, because they are relatively large and consistent in their anatomy. In older children, veins in the back of the hands and forearms are commonly used. Avoiding the dominant hand or the hand with the finger or thumb that the infant prefers to suck allows for improved patient comfort and function, although such choices are not always available.

Before the vein is cannulated, the operator should wash his or her hands well and use universal precautions, including protecting the operator’s hands with gloves. The extremities should be adequately immobilized, and the site should be cleansed with alcohol or iodine-containing solutions and allowed to dry. A tourniquet should be applied in order to engorge and distend the vein. The skin can be stretched taut with the operator’s nondominant hand in order to immobilize the vein. The operator should puncture the skin at a 15° to 30° angle, 5 to 10 mm distal to the expected entrance site into the vein. Next, the needle punctures the vein, usually resulting in blood return into the catheter hub. When this blood return occurs, the catheter is advanced a few millimeters to ensure that the catheter tip and the needle are in the lumen of the vein. The catheter is then advanced over the needle into the vein, and the needle is removed. The tourniquet is released, and saline is flushed intravenously to ensure patency of the catheter and vein. After the needle is removed, it should generally not be reinserted while the catheter is in the subcutaneous space or in the vein, because this could shear the tip of the catheter and generate a small plastic embolus.

Adequate immobilization of the extremity is an important aspect of successful peripheral vein cannulation. Moreover, it is important to adequately secure and protect the catheter after successful cannulation. Notably, ultrasound or devices using near-infrared have been used for difficult peripheral vein access.

INTRAOSSEOUS INFUSION

Although emergency intraosseous (IO) infusions were common in the 1940s and 1950s, the arrival of butterfly needles and plastic catheters and the widespread use of venesection led to a virtual extinction of this technique for several decades. However, there has been a resurgence in the use of intraosseous infusions since the early 1980s. The principal value of intraosseous access is that it can be performed rapidly and reliably. Medical personnel can usually attain access in less than 1 minute during emergencies. The bone marrow cavity is effectively a noncollapsible vascular space, even in the setting of shock or cardiac arrest. Therefore, intraosseous access is the initial vascular access site of choice in patients with life-threatening problems such as cardiopulmonary arrest or decompensated shock.

Almost any medication that can be administered into a central or peripheral vein can be safely infused into the bone marrow, including crystalloid solutions, colloid solutions, blood products, and hypertonic solutions. In particular, all the medications the American Heart Association recommends for pediatric advanced life support can be safely and effectively administered via this route, including catecholamine infusions. Pharmacokinetic variables, such as onset of action and plasma concentrations, are similar with intraosseous or peripheral venous administration in cases of circulatory shock or cardiac arrest with cardiopulmonary resuscitation (CPR). Emergency medications should be followed by a saline flush to ensure rapid delivery into the circulation. In addition, the initial bone marrow aspirate is a reliable specimen for venous pH and PCO₂, blood typing and cross-matching, serum glucose, electrolytes, and blood cultures. Due to stasis in the bone marrow, the results of such studies may be less reliable after administering drugs through the intraosseous needle during CPR.

The greatest obstacle to successful intraosseous cannulation is psychological and originates from the natural reluctance by any inexperienced provider to force a needle into a child’s bone. Experienced providers may at times wrongly forfeit this quick and reliable alternative to undertake venous cannulation via rapid percutaneous central venous catheterization or peripheral vein cutdown, sometimes resulting in dangerous delays in starting therapy. Intraosseous cannulation is generally a safe procedure but can result in complications, including osteomyelitis, fractures, subcutaneous or intramuscular extravasation of toxic medications (eg, epinephrine, calcium), and compartment syndrome. Compartment syndrome is easily avoided by monitoring the insertion site for swelling and discontinuing the infusion if significant swelling occurs. Microvascular pulmonary fat and bone marrow emboli have been demonstrated but do not appear to be a clinically significant problem. The risk of such complications is acceptable in the dire circumstances of hypotensive shock or cardiac arrest. Such risks, and the pain involved, may not be acceptable when the clinical indications are less stringent.

The most commonly utilized site for intraosseous access is the medial surface of the tibia, 1 to 3 cm below the tibial tuberosity (Fig. 100-1). Alternative sites include the distal tibia above the medial malleolus, the distal femur, and the anterior superior iliac spine. There are various styles of intraosseous needles. They are all designed to easily penetrate the bone cortex, and they all have a stylet to keep bone core from obstructing the needle during insertion. Aseptic technique and universal precautions should be followed. The needle should be twisted into, rather than pushed through, the bone marrow. Evidence for successful entrance into the marrow includes (1) the lack of resistance (or a “give”) after the needle passes through the cortex, (2) the ability of the needle to remain upright without support, (3) aspiration of the bone marrow into a syringe, and (4) free flow of the infusion without significant subcutaneous infiltration. Aspiration of bone marrow into the intraosseous needle is not always possible, especially in a very dehydrated patient. An alternative method of IO insertion involves using a battery-operated drill that allows for quick and effective insertion of the IO needle. Regardless of the insertion method used, infiltration of fluid into tissues around the bone is common when this route of infusion is used for a prolonged period of time or if the fluid is infused under great pressure. Infiltration will manifest as an enlarging leg circumference, fluid leaking out from the skin insertion site, or resistance to flow of the infusate.

CENTRAL VENOUS CATHETERIZATION

Central venous catheterization provides more reliable and longer term vascular access than peripheral venous catheterization. In addition, central venous catheters permit hemodynamic monitoring and laboratory sampling of central venous blood. However, the convenience of central venous catheters must be balanced against the added risks.

Central venous cannulation is particularly suited for administering irritating or vasoconstrictive medications, which are diluted in the high blood flow of central veins, thereby limiting their contact with the vascular endothelium. Central venous cannulation may also offer a more expeditious alternative to peripheral venous access, particularly for patients who are in circulatory shock or cardiac arrest. Especially in these cases, a central venous catheter provides a means to monitor central venous pressure. In some circumstances, specialized central venous catheters may be used to monitor cardiac output, mixed venous oxygen saturation, pulmonary artery pressure, and pulmonary artery occlusion pressure.

In critically ill infants and children, polyurethane central venous catheters are generally inserted percutaneously through a large central vein such as the internal jugular, the subclavian, or the femoral vein. For chronic vascular access, particularly in patients requiring long-term chemotherapy, total parenteral nutrition, or antibiotics, Silastic catheters are often preferred, because they are less thrombogenic and have lower infection rates. It is now widely recommended that these catheters are placed using ultrasound guidance if this is available.

Peripherally inserted central catheters (PICC) have gained enormous popularity, because they can be inserted easily at the bedside, usually through a peripheral vein (eg, brachial vein), and advanced into a central location, using ultrasound and in some cases fluoroscopic guidance. The Broviac and Hickman catheters are placed in a central vein and tunneled out through a distant exit site in the skin. The Hickman catheters have Dacron cuffs that are very fibrogenic. The resultant fibrous scar around the Dacron allows for the catheter to be more securely anchored and decreases the risk of catheter infection from the skin. A last category of Silastic catheters is equipped with a proximal port that is embedded in the subcutaneous tissue (Port-A-Cath). These totally implantable venous access devices reduce the chance of microbial propagation through the catheter track and are most useful when only intermittent infusions are necessary.

There are unavoidable risks associated with all central venous catheters. During attempts to cannulate the internal jugular or subclavian veins, inadvertent needle puncture of the lungs can result in pneumothorax. Perforation of a large vessel with resultant leak into the pleural space can lead to hemothorax, and mechanical damage to the thoracic duct or thrombosis of the left brachiocephalic vein downstream from the insertion of the thoracic duct can lead to chylothorax. Central vein thrombosis and its attendant risks (venous congestion and edema, chylothorax) are particularly common in newborns and small infants because of the small size of their central vessels. In addition, injuries to nerves and arteries (including inadvertent cannulation) near the insertion site have been well documented.

When the central venous catheter is open to the atmosphere and the patient creates negative intravascular pressure with a spontaneous breath, it is possible for air embolism to occur. This can be minimized by keeping the needle or catheter exit site in a dependent position (eg, Trendelenburg position during internal jugular or subclavian catheterization). The Seldinger technique, or guidewire technique, is commonly used to simplify catheter placement. A needle or trocar is used to puncture the vessel, followed by passage of a wire through the trocar. The trocar is then removed, and the catheter is then introduced over the wire. This technique is convenient but is associated with both an increased risk of vessel perforation by the guidewire, and catheter malposition because of the potential for the guidewire to take an unexpected course in the vascular system.

The most significant risk of central venous catheterization is infection. The risks of both thrombosis and suppurative thrombophlebitis increases with time and is especially high in children who suffer from hypercoagulability and/or immune deficiencies. The possibility of a hypercoagulable state should be considered in critically ill children who develop catheter-related central venous thrombosis. Newer catheter materials (polyurethane, Silastic), heparin bonding, and low-dose heparin administration through the catheter may minimize the risks of thrombosis. Similarly, antibiotic bonding has been used to decrease the risk of infection. However, the most effective preventative measures are based on the conventional wisdom of placing central venous catheters only when indicated and removing them as soon as possible.

Local hemorrhage is common when removing the central venous catheters. Applying direct pressure as the catheter is removed can minimize the size of the hematoma. Air embolism is a rare, life-threatening complication associated with catheter removal. As during insertion, the risk of air embolism can be minimized by placing the patient in the Trendelenburg position, removing the catheter at the end of deep inspiration, and immediately covering the exit site with petroleum jelly gauze.

The most commonly used sites for central venous catheterization (Fig. 100-2) in acutely ill children are the internal jugular (less frequently, the external jugular vein is cannulated) or femoral veins. The subclavian veins are also often used by experienced clinicians, although the risk of pneumothorax or hemothorax is higher than with the other techniques. The femoral vein is the safest approach for less-experienced providers and, unlike in adults, does not seem to carry a higher risk of infection. Femoral cannulation offers clear anatomical landmarks, ease of hemostasis, and a safe distance from vital organs and sites of activity during resuscitation. Prior to catheterization of the neck veins, the child should be placed in the Trendelenburg. Conversely, if the femoral vein is used, a slight reverse Trendelenburg position will engorge the veins and improve the success rate.

Ultrasound-guided central venous access is now the preferred approach to clarify anatomy of nearby arteries and veins, identify the vein for cannulation, determine that the vein is patent, assure that the insertion needle tip is in the vessel prior to advancing a wire through the needle, and decrease the risk of complications, such as inadvertent arterial cannulation. The technique requires additional training and equipment but increases safety and effectiveness. The overall procedure is otherwise similar to the traditional approach utilizing landmark identification.

Local anesthesia with 1% lidocaine should be provided subcutaneously unless the patient is anesthetized. A syringe and needle are used to locate the vein and aspirate blood. For femoral vein catheterization, the vein is located immediately medial to the artery, 1 to 2 fingers’ breadths below the inguinal ligament. The introducer needle is inserted into the vein at a 45° angle. When blood is aspirated freely, the syringe is removed and the hub of the needle is covered with the thumb of the nondominant hand. A guidewire is then placed through the needle, and the needle is removed (Seldinger technique). A dilator is advanced over the wire to dilate the subcutaneous space and the entrance into the vein. The dilator is then removed, and the catheter is placed over the guidewire. After the catheter is introduced into the vein, the guidewire is removed. The electrocardiogram should be monitored during the procedure, because the guidewire can trigger arrhythmias when advanced into the heart.

Once the procedure is complete, it is essential to aspirate all of the lumens of the line to check that blood flows freely, and then check the position of the central line using 1 or more of the following confirmatory tests (depending on availability): x-ray to confirm the position of the line tip; a blood gas to distinguish from an inadvertent arterial cannulation (although this can be confounded by the presence of arterial hypoxemia); and observation of the pressure waveform of the cannulated vessel after zeroing the line (again, to rule out arterial cannulation). Continuous monitoring of the pressure during the procedure is essential in the setting of pulmonary artery catheterization.

Central line-associated bloodstream infections (CLABSI) are key quality indicators for not only intensive care units, but for hospitals as a whole and are associated with increased morbidity and increased mortality across multiple populations. Many institutions have central line checklists, or “bundles” that incorporate key interventions to minimize the risk of any adverse events related to central line insertion, in particular CLABSI. These include proper hand hygiene; preparing the skin with chlorhexidine; using a full-body sterile barrier, surgical masks, gowns, caps, and gloves for those performing the procedure; and removing unnecessary catheters. The introduction of central line bundles has been associated with a marked decrease in the rate of CLABSI.

Transduction of vascular pressure during the procedure provides a more effective means of confirmation and is essential if the catheter is to be inserted into the pulmonary artery. A radiograph can also be used to confirm position, but sometimes a single projection is insufficient to determine the precise distal site of the catheter (eg, distinguishing right atrium from right ventricle).

ANALGESIA AND SEDATION FOR PROCEDURES

An important yet frequently neglected issue for any invasive procedure on a child is that of pain management and sedation. In the nonemergency situation, topical or local anesthetics should be used, and oral sucrose has been shown to lessen distress in younger infants. The psychological milieu during the procedure should also be optimized. Music, toys, and interaction with family members can be important tools for improving comfort, thereby increasing the likelihood of cooperation and technical success. Moreover, minimizing discomfort during the procedure generally encourages better cooperation during the next potentially painful procedure for that child. However, when sedation is provided, personnel qualified to handle sedation-induced respiratory failure or shock should be present.

SUGGESTED READINGS

INTRODUCTION

Children with acute or critical illness may be hypovolemic, euvolemic, or hypervolemic. They may have lost blood (eg, hemorrhagic shock), plasma (eg, severe burns), extracellular fluid (eg, gastroenteritis or diarrhea), or electrolytes. They may have internal compartment fluid redistribution (eg, ascites or capillary leak from sepsis or Dengue shock syndrome) and despite being edematous may still have intravascular volume depletion. Children with acute or critical illness may have renal impairment or high antidiuretic hormone (ADH) levels that results in fluid retention. Decisions about fluid management need to be based on clinical features and understanding of pathophysiology, thus systematic and repeated clinical evaluations are essential to good quality care. In this chapter, several acute clinical syndromes that require careful fluid management will be discussed.

SHOCK AND FLUID MANAGEMENT

Cells require oxygen for the production of adenosine triphosphate (ATP), the principal cellular energy source. In pathophysiological terms, shock occurs when cardiac output is insufficient to provide oxygen delivery to tissues for ATP production, leading to cellular dysoxia, with resulting anaerobic metabolism, lactic acidosis, and cellular dysfunction. The assumption behind fluid bolus therapy for shock is that it will improve cardiac output and in turn augment oxygen delivery to tissues. There are several clinical definitions of shock. Hypovolemic, low cardiac output shock manifests as “cold shock” with vasoconstriction, cold limbs, and tachycardia. This type of shock requires urgent fluid resuscitation. Cold shock occurs in severe dehydration (the classic examples being from cholera or gastroenteritis), hemorrhagic shock, and in some (generally late) stages of septic shock.

Fluid resuscitation for children with fever and signs of shock has been controversial. This is partly because many clinical definitions of shock encompass a continuum from adaptive physiological changes to fever, to states of severe hypotension and dysoxia. Fluid therapy for shock has also been controversial because fluid therapy alone will not deal with shock apart from that which occurs solely from extracellular fluid losses. Therefore, in settings where intensive care support is unavailable, fluid therapy alone in some forms of shock will be insufficient. Many children with fever and 1 or 2 clinical cardiovascular signs of shock do not have hypovolemia or dysoxia. They have high levels of adrenaline and renin-angiotensin, leading to tachycardia and vasoconstriction, and raised levels of ADH, which leads to fluid retention, thus protecting them from hypovolemia and shock. The assumption that fluid boluses will improve cardiac output and in turn augment oxygen delivery to tissues is not certain if the cardiac output is normal or the child is not hypovolemic. In this situation, the benefits of bolus fluid therapy are less and the costs of excess fluid will be tissue edema and perhaps a blunting of the adaptive cardiovascular responses. A transient increase in stroke volume after fluid bolus therapy, because of an increase in venous return and left ventricular end-diastolic (filling) pressure, is often accompanied by a reduction in heart rate. Thus, the effect on cardiac output, which is the product of the 2, in some circumstances will be small and transient. Large amounts of intravenous (IV) fluid will tend to accumulate in children with high ADH levels, and edema, reduced lung compliance, and other negative consequences may occur. The effect of fluid boluses on dampening adaptive responses like the hyperadrenergic state has been proposed but needs further study.

Revised guidelines for the management of shock, incorporating the World Health Organization’s new fluid guidelines for shock, are in Appendix 101-1.

SEVERE DEHYDRATION AND FLUID MANAGEMENT

Children with hypovolemic shock from severe dehydration should be given parenteral fluid immediately: administer 20 mL/kg of 0.9% sodium chloride (NaCl) or Ringer’s lactate (or similar fluid) repeatedly until plasma volume is restored, and pulses are present (20–100 mL/kg may be required). Fluid can be given intravenously, or if an IV cannula cannot be inserted, into the bone marrow through an intraosseous needle. Nasogastric fluid is effective in severe dehydration, but once shock has developed, the splanchnic circulation is less able to absorb fluid. Intraperitoneal fluid is not effective in shock. Any child with shock who requires more than 40 mL per kg as a bolus should be urgently reviewed to consider the need for vasopressor or inotropic support, and to consider the adverse effects of more fluid on lung function, and where available intensive care review should be considered.

In calculating fluid requirements for children with dehydration, 3 fluid types need to be considered:

• Existing fluid deficit

• Ongoing fluid losses

• Maintenance fluid requirements

Fluid Deficit

The fluid deficit is most reliably estimated from the loss of body weight if a recent pre-illness weight is available. The loss of weight is equal to the fluid deficit. (For example, if 600 grams have been lost, there is a fluid deficit of 600 mL, which represents 6% dehydration in a 10-kg child). The deficit may be overestimated in infants whose most recent weight was months ago, as the weight difference does not take account of normal growth, which may be 100 g to 150 g per week.

Clinical signs may be used to determine approximate fluid deficit. These are useful but not precise.

1. Mild dehydration (< 4%)

   • Usually no clinical signs; the child is thirsty

2. Moderate dehydration (4–6%)

   • Delayed capillary refill time

   • Increased respiratory rate

   • Mildly decreased tissue turgor

3. Severe dehydration (≥ 7%)

   • Very delayed capillary refill, > 3 seconds; cold, mottled skin

   • Other signs of shock (irritable or reduced consciousness level, hypotension, tachycardia)

   • Deep, acidotic breathing

   • Decreased tissue turgor

In children, hypotension is a late sign, but is often preceded by a narrowing of the pulse pressure from vasoconstriction. Other markers of mild to moderate dehydration include the following:

• A history of oliguria

• Lethargy

• Sunken eyes

• Dry mucous membranes

• Sunken fontanelle

• Absence of tears

In hypernatremic dehydration, the clinical signs of dehydration are less marked because the extracellular compartment is relatively preserved, and the degree of dehydration is underestimated. A child with hypernatremic dehydration may look only mildly unwell until they suddenly collapse with shock.

In severe dehydration, the existing deficit should be replaced over the first 12 to 24 hours, using the IV (or nasogastric) route. Faster deficit replacement has been shown to be safe in many children, but confers no advantage apart from earlier discharge from an emergency department. For replacement of deficit, an isotonic fluid should be used.

In severe dehydration electrolytes should be checked; hypokalemia is common in dehydration from diarrhea, but hyperkalemia can occur if there is prerenal failure resulting from hypovolemia, or severe metabolic acidosis. Hyper- and hyponatremia, hypoglycemia, and hypomagnesemia are commonly seen in children with severe dehydration.

Hypernatremic dehydration is most frequently caused by a water deficit (increased losses or reduced intake). In some cases, it is caused by an increase in total body sodium (such as ingestion of large quantities of sodium, incorrect preparation of milk formula, ingestion of hyperosmotic feeds in the setting of diarrhea). For hypernatremic dehydration, after the correction of shock if present, the aim is to lower the serum sodium slowly at a rate of no faster than 12 mmol/L in 24 hours (0.5 mmol/L per hour). This rate of normalization of serum sodium should be even slower if hypernatremia is chronic (weeks). If sodium is corrected too rapidly, cerebral edema, seizures, and permanent brain injury may occur. In moderate hypernatremic dehydration, Na⁺ 150 to 169 mmol/L after initial resuscitation, replace the deficit plus maintenance over 48 hours. Nasogastric oral rehydration salts should be used whenever possible, remembering that these have a low sodium concentration (60–90 mmol/L). Monitor electrolytes every 4 to 6 hours. If requiring IV rehydration, then Ringer’s lactate +5% dextrose, 0.9% NaCl + 5% dextrose, or PlasmaLyte with 5% dextrose should be given, adding maintenance potassium chloride once urine output has been established. In severe hypernatremic dehydration, Na⁺ > 169 mmol/L, the child should be admitted to intensive care. After initial resuscitation from shock if present, the fluid deficit should be replaced using 0.9% NaCl + 5% dextrose over 72 to 96 hours. In severe hypernatremic dehydration, electrolytes and glucose should initially be checked hourly. If serum sodium is falling faster than 0.5 mmol/h, then the infusion rate should be slowed down by 20%, and the serum sodium should be rechecked after 1 hour. If, after 6 hours of rehydration therapy, the sodium is decreasing at a steady rate, then the electrolytes and glucose can be checked every 4 hours. If there are persistent neurological signs, cerebral imaging should be considered. Venous sinus thromboses can occur from severe hypernatremic dehydration.

Ongoing Losses

It is vitally important to document fluid balance carefully and, ideally, hourly in patients during active fluid resuscitation. Fluid balance charts should accurately record—whenever possible—output volumes, including urine, gastric aspirates, vomitus, diarrhea, drainage from fistulae, and other fluid losses. These will determine the volume and type of fluid and electrolyte replacement. In practice, it can be difficult to accurately measure all losses. A rule of thumb is to add 100 mL for every diarrheal stool in the previous 12 hours to make up the ongoing losses that are added to the total fluid intake for the next 12 hours.

Maintenance Fluid

For over 5 decades, the standard maintenance IV fluid for children in hospitals throughout the world was hypotonic saline and glucose. A commonly used solution was 4% dextrose and 0.18% NaCl (containing 30 mmol/L sodium and chloride), and another fluid used in some countries was 3.3% dextrose and 0.3% NaCl. The recommended flow rates of IV maintenance fluid followed the 100/50/20 rule: 100 mL/kg/day for the first 10 kg body weight, 50 mL/kg/day for the next 10 kg, and 20 mL/kg/day for each kg thereafter. This roughly equates with the 4/2/1 rule, which is in mL/kg per hour. This is approximately the amount of water that a well, active child who has a normal metabolism and normal renal function requires to excrete nitrogenous wastes in an isosmotic urine, that is, urine that is neither overly concentrated nor overly dilute. The composition of the recommended IV fluids was based on the calculation that, if a child receives “normal maintenance volumes” for weight using 4% dextrose and 0.18% NaCl, he or she would receive 2 to 3 mmol/kg/day sodium (the amount needed for metabolism and growth), and 3.5 mg/kg/min of glucose (enough to avoid hypoglycemia in a patient who is nil orally).

The problem lies in the assumptions underlying these recommendations. Sick children are not the same as well, active children. Hospitalized children with infectious processes, such as pneumonia, bronchiolitis, sepsis, or central nervous system infections, and those with other conditions including surgical emergencies and those who are postanesthesia, commonly have increased levels of ADH, which causes water retention. Depending on how sick they are, such children may also have impaired renal function, increased capillary permeability (including in the brain, lungs, and soft tissues), poor lung compliance, and hypoproteinemia. The retention of water by ADH is a protective physiological mechanism for fasting sick patients. However, the administration of “normal maintenance IV fluids” to a child who cannot excrete a free water load risks fluid overload and edema of tissues, lungs, brain, and other organs. The administration of an IV solution with a sodium content substantially lower than plasma in a child who has impaired ability to excrete free water leads to hyponatremia, and this can lead to brain cell edema. On an unknown number of occasions, this has resulted in seizures, cerebral edema, and death. While some risk factors for these complications are known, it is not easy to predict to which patients this will occur, and hyponatremia and cerebral edema can also occur in patients without known factors. Hyponatremia is not always preventable.

Probably even more frequent than neurological complications are those arising from excessive volumes of IV fluid relative to true requirements for fluid homeostasis. Tissue edema leads to poor lung compliance and worsening hypoxemia in children with respiratory infections, soft tissue edema, immobility and pressure areas in children in coma, and impaired kidney function. There are now many case series of iatrogenic hyponatremia and its consequences; patent safety alert warnings by national and coronial authorities from the UK, Northern Ireland, Canada, and elsewhere; and randomized trials showing a lower risk of hyponatremia in children who receive isotonic fluids.

In 2013, the World Health Organization (WHO) recommended that “children who require IV fluids for maintenance should be managed with Ringers lactate with 5% dextrose, or 0.9% normal saline with 5% dextrose or half-normal saline (0.45% NaCl) with 5% glucose.”

Fluid management of critically ill children requires careful assessment of disease process, pathophysiology, hydration, and intravascular volume status and responses to administered fluid. Both standardized guidelines and an individualized approach are needed for optimal fluid management. Guidelines for prescribing maintenance fluid are detailed in Appendix 101-2, and the water and electrolyte contents of commonly used fluids are given in Table 101-1.

SUGGESTED READINGS

INTRODUCTION

Successful closed-chest cardiopulmonary resuscitation (CPR), the fundamental life-saving skill, was first reported in 1960, and the American Heart Association (AHA) has published pediatric and neonatal resuscitation guidelines since 1980. Although CPR has been widely taught to healthcare providers (HCPs) and the public, it is only recently that survival from cardiac arrest has improved. This improvement has occurred in association with an emphasis on teaching, monitoring, and improving CPR quality; widespread implementation of lay-rescuer CPR and automated external defibrillation (AED) programs; increased frequency of dispatcher-guided lay-rescuer CPR; and improvements in postcardiac arrest care, including targeted temperature management. Published studies from both out-of-hospital and in-hospital registries have provided additional information about the epidemiology, presentation, and outcome of pediatric cardiopulmonary arrest (CPA) at different ages and in different settings. The AHA Guidelines for Pediatric Basic Life Support (PBLS) and Pediatric Advanced Life Support (PALS) are now based on a continuous, structured international evidence-based review process. The reviews, sponsored by the AHA and the International Liaison Committee on Resuscitation (ILCOR), are posted on ILCOR’s Web site (http://www.ilcor.org./home/), in the Systematic Evidence Evaluation and Review System.

Neonatal resuscitation guidelines are jointly created by the AHA and the American Academy of Pediatrics. These guidelines target resuscitation in the delivery room, emphasizing establishment of effective ventilation, because inadequate airway and ventilation are the most common problems requiring newborn resuscitation. Neonatal resuscitation guidelines are typically applied during the initial hospitalization of the newborn. The AHA infant Basic Life Support (BLS) and CPR guidelines typically apply to the infant after the initial hospitalization until 1 year of age. If a newborn remains hospitalized for treatment of heart disease, it is appropriate to apply infant CPR guidelines to provide a higher compression-to-ventilation ratio consistent with focus on establishment of adequate blood flow as well as oxygenation and ventilation. The AHA PBLS and CPR guidelines apply to children from 1 year of age to puberty. Adult BLS guidelines apply to adolescents. For ease of teaching and practice, providers in a unit such as a pediatric cardiovascular intensive care unit (ICU) may elect to apply the infant and child CPR guidelines to all patients in the unit.

In 2010, the AHA recommended a change in the initial sequence of CPR from ABC (airway, breathing, circulation/compressions) to CAB (circulation/compressions, airway, breathing). This change was made because the initial steps of opening the airway and delivering breaths are relatively complicated and often created long delays to initiating chest compressions. Although support of airway, oxygenation, and ventilation is especially important during pediatric cardiac arrest, this change in sequence should delay ventilation only by a few seconds, and is designed to shorten the overall time to initiation of CPR for all victims of cardiac arrest.

Every step in resuscitation is important, including treating prearrest conditions when possible; immediately identifying the arrest itself and providing high-quality CPR; and seamlessly integrating shock delivery with CPR when needed. Skilled, evidence- and protocol-supported postcardiac arrest care, including targeted temperature management is critical.

PATHOGENESIS AND ETIOLOGY OF CARDIAC ARREST

The causes and types of cardiac arrest during infancy and childhood vary by age and arrest location and will affect the priorities of resuscitation. In the past, much of the evidence used to characterize pediatric cardiac arrest was based on extrapolation from adult series; however, in recent years, several pediatric cardiac arrest registries have provided a wealth of information about many characteristics of pediatric CPA.

Data from Registries

Multiple large pediatric out-of-hospital cardiac arrest series have been published from registries such as the Cardiac Arrest Registry to Enhance Survival (CARES; https://mycares.net/), the Resuscitation Outcomes Consortium (ROC; https://roc.uwctc.org) registry, and the all-Japan Utstein Registry. In the United States, the Get with the Guidelines-Resuscitation (GWTG-R) registry, formerly the National Registry of Cardiopulmonary Resuscitation (http://www.heart.org/HEARTORG/Professional/GetWithTheGuidelines-Resuscitation/Get-With-The-Guidelines-Resuscitation_UCM_314496_SubHomePage.jsp) has published multiple pediatric series. These registries collect data based on the reporting templates known as the Utstein criteria. Publications from registries such as these contribute substantially to our knowledge of pediatric prearrest, arrest and postarrest conditions, further informing the emphasis of the AHA PBLS and PALS Guidelines.

Asphyxial Versus Sudden Cardiac Arrest

There are 2 major types of CPA: those that occur secondary to progression of respiratory failure or shock (so-called asphyxial or hypoxic arrest) and those that occur as an abrupt cessation of cardiac function, called sudden cardiac arrest (SCA). These types of CPA can be further identified by the terminal cardiac rhythm. Bradycardia and pulseless electrical activity (PEA) are the most common terminal rhythms associated with asphyxial arrest; these rhythms are called “nonshockable” because shock delivery is not indicated to treat the arrest. Pulseless ventricular tachycardia (pVT) and ventricular fibrillation (VF) are the terminal rhythms associated with SCA; these rhythms are known as “shockable” rhythms because shock delivery is required. The incidence of each type of CPA varies with the age and condition of the child, and with the location and circumstances of the arrest. Resuscitation priorities vary with arrest type, the number of rescuers, and the location of the arrest.

At birth, approximately 5% to 10% of newborns require some resuscitation, ranging from simple stimulation or clearing of the airway to support of ventilation and occasionally to chest compressions. Most neonates in distress demonstrate bradycardia, and most will respond to the establishment of effective ventilation.

Asphyxial arrest is the most common type of CPA in infants and children. However, this type of arrest can occur in patients of any age with conditions such as respiratory failure, trauma, drowning, drug overdose, poisonings, metabolic disorders, and shock. In the United States, the most common causes of infant death include congenital malformations, disorders related to prematurity, sudden unexpected infant death (including sudden infant death syndrome [SIDS], suffocation, and strangulation) and injury. Injuries are the most common causes of childhood and adolescent death. See also Chapters 113 and 116.

The infant or child often develops bradycardia as a preterminal event after asphyxia arrest. Bradycardia in children may also result from vagal stimulation (eg, induced by suctioning or gagging), from other causes of hypoxia, or from heart block. In the minutes before an asphyxial or bradycardic arrest, oxygen delivery to the tissues is compromised by low arterial oxygen content, by inadequate blood flow, or both, and major organ ischemia is likely to be present even before the arrest occurs. Resuscitation requires conventional CPR because support of effective oxygenation and ventilation, and compressions are needed to support oxygen delivery.

The infant or child with SCA has pVT/VF that causes the abrupt cessation of cardiac function. This type of cardiac arrest is most common in adults but is less common in children, particularly before adolescence. SCA has been reported in young athletes following a blow to the chest (commotio cordis) and in infants and children with genetic anomalies of impulse formation and conduction (eg, channelopathies such as prolonged QT syndrome), congenital heart disease, coronary artery abnormalities, cardiac inflammation, cardiomyopathy, and drug toxicity. Because arterial oxygen content is normal at the time of SCA, immediate support of blood flow with chest compressions may maintain oxygen delivery for the first few minutes of the arrest. However, beyond the first minutes of arrest, arterial oxygen content declines rapidly, so conventional CPR is needed.

Characteristics of Out-of-Hospital Arrest

Estimates of the incidence of asphyxial arrest versus SCA are based on extrapolation from mortality data collected by the Centers for Disease Control and Prevention and from published case series and registry data. In general, two-thirds of out-of-hospital pediatric cardiac arrests are asphyxial, about two-thirds of victims are infants who arrest at home, most are unwitnessed, and about half of the victims receive bystander CPR. These characteristics provide important information for exploring causes and prevention of CPR, targeting CPR education to parents and caregivers, and establishing resuscitation priorities for dispatcher-guided CPR.

Cardiac arrest in schools, particularly among student athletes, has stimulated grassroots efforts to establish CPR and AED programs in schools. Extrapolation from voluntary US databases suggests that approximately 1 per 100,000 high school athletes suffers from SCA annually, with the incidence higher in males and lower in females, and lower among nonathletes. In the United States, more than half of the SCA deaths in athletes are attributed to hypertrophic cardiomyopathy, commotio cordis (a blow to the chest that triggers ventricular tachycardia or fibrillation), or coronary artery anomalies. Some athletes or children with arrhythmias such as those caused by channelopathies may have syncopal episodes prior to the cardiac arrest event. Vigorous exercise can trigger lethal ventricular arrhythmias; many episodes of SCA in athletes occur during sporting events and are witnessed.

Characteristics of In-Hospital Arrest

Publications from the nearly 10,000 pediatric arrests in the GWTG-R registry have enabled the characterization of in-hospital arrest in children. Most children with in-hospital cardiac arrest have prearrest respiratory insufficiency, shock, or congestive heart failure. The most common intermediate causes of arrest are hypotension, acute respiratory insufficiency, and arrhythmias. PEA or asystole is the terminal rhythm in two-thirds of the arrests. Of the children with initial pVT/VF, approximately half have a cardiac illness or surgical condition, and prearrest vasoactive infusions have been identified as proarrhythmic factors. Although pVT/VF is not a common initial arrest rhythm, VF is present at some time during resuscitation in approximately 25% of pediatric in-hospital arrests, so providers must be facile at integrating shock delivery with high-quality CPR.

RESPIRATORY AND CARDIAC ARREST: CLINICAL MANIFESTATIONS

Respiratory Arrest

Respiratory arrest is defined as the cessation of spontaneous respiratory activity in the presence of spontaneous circulation (palpable central pulses). The development of hypoxemia and tissue hypoxia will produce bradycardia and poor perfusion. Detection and immediate treatment of respiratory arrest and symptomatic bradycardia often prevent progression to CPA.

Cardiopulmonary Arrest

Cardiopulmonary arrest is defined as the cessation of effective cardiac mechanical function and systemic perfusion. The victim is unconscious/unresponsive, with apnea or agonal gasps and with no palpable central pulses. Agonal gasps may be mistaken for effective spontaneous breathing, delaying the identification of cardiac arrest. They are much more common in the first minutes of SCA but are uncommon in children with asphyxial cardiac arrest. Agonal gasps can increase pulmonary gas exchange if the airway is patent. Gasps can also enhance venous return and create some coronary and carotid perfusion during the first minutes of arrest.

Once cardiac arrest develops, oxygen delivery ceases. Unless CPR is provided and spontaneous rhythm and perfusion are restored, the myocardium, brain, and all tissues will become progressively ischemic, and the likelihood of successful resuscitation diminishes with every passing minute. Generation of lactic acid, increased cell membrane permeability, production of free oxygen radicals, and activation of inflammatory mediators contribute to progressive organ and tissue destruction.

Pulseless Electrical Activity and Asystole

Pulseless electrical activity is a term used to describe cardiac electrical activity that fails to produce sufficient mechanical function to generate a palpable central pulse. Although this term could encompass agonal (eg, bradyasystolic) rhythms, it is typically reserved for patients with narrow QRS complexes that fail to produce a palpable pulse. Reversible causes of PEA in children include tension pneumothorax, cardiac tamponade, hypovolemia, and, rarely, pulmonary embolus. If PEA is not promptly treated, the rhythm will ultimately deteriorate to asystole.

Asystole is the ultimate terminal rhythm, characterized by electrical silence. Asystole is likely to be present for any unwitnessed or prolonged arrest, and it is the most common rhythm reported in pediatric prehospital and in-hospital arrest. Although overall survival with asystole or PEA is poor, survival is higher among children than adults who present with these rhythms in in-hospital arrests.

Ventricular Tachycardia and Fibrillation

Just as in adults, children in cardiac arrest who present with pVT/VF in the out-of-hospital or in-hospital setting typically have higher survival rates than those presenting with PEA or asystole. Untreated pVT will rapidly progress to VF, so treatment of these rhythms is identical and the rhythms are considered together in treatment algorithms.

During the first few minutes of VF, the amplitude of the VF will be high (so-called good VF), indicating that the myocardium initially has adequate oxygen and substrates. High-quality CPR can help maintain myocardial oxygenation and the amplitude and duration of VF (before it deteriorates to asystole). Rapid shock delivery is likely to result in elimination of the VF, enabling return of spontaneous cardiac rhythm and return of spontaneous circulation (ROSC). It is important to note that typically there is a delay between elimination of VF and ROSC, so CPR is needed until the shock can be delivered and after shock delivery until ROSC occurs.

If the VF remains untreated (ie, no CPR or shock delivery) for several minutes after arrest, the VF amplitude will gradually decrease (so-called bad VF), indicating severe myocardial ischemia. At this point, shock delivery is less likely to eliminate the VF; multiple shocks may be required. Even if VF is eliminated, return of spontaneous rhythm and return of spontaneous cardiac rhythm are less likely.

Untreated VF will ultimately progress to asystole. Once asystole is present, shock delivery will not be effective. High-quality CPR may result in the reappearance of VF (so-called secondary VF) that may respond to shock delivery, although the survival of patients who present with secondary VF is much lower than that of patients who present with VF as their initial rhythm.

RESPIRATORY AND CARDIAC ARREST: MANAGEMENT

Respiratory Arrest and Bradycardia

The child with respiratory failure, shock, or respiratory arrest typically develops bradycardia with poor perfusion before developing pulseless arrest. If rescuers detect and treat respiratory arrest and bradycardia before the development of pulseless arrest, survival is typically 75% or higher in both out-of-hospital and in-hospital settings.

Symptomatic bradycardia (ie, bradycardia with signs of poor perfusion) and respiratory arrest are both treated with support of airway, oxygenation, and ventilation through provision of bag-mask ventilation with oxygen. If the heart rate and perfusion improve but ventilation remains inadequate, mechanical ventilation is indicated. If the child’s heart rate remains at or below 60 beats per minute with signs of poor perfusion despite support of adequate oxygenation and ventilation, chest compressions are added to the bag-mask ventilation (ie, CPR is provided).

When symptomatic bradycardia persists despite bag-mask ventilation and compressions, epinephrine is the drug of choice. Rescuers should consider using atropine for symptomatic bradycardia with increased vagal tone or when primary heart block is present. Cardiac pacing may also be needed.

Opiate overdose has emerged as a common cause of death in young adults. If the lay first responder or HCP encounters a victim of respiratory arrest with suspected drug overdose, naloxone administration is added to interventions such as bag-mask ventilation and CPR as indicated.

Cardiopulmonary Arrest

Once cardiopulmonary arrest is present, immediate high-quality CPR is needed because CPR maintains a small but critical amount of blood flow (estimated at 10% to 33% of normal) and oxygen and substrate delivery to the heart, brain, and other organs. Performing CPR can prolong VF and can maintain or increase VF amplitude. This increases the window of opportunity for shock delivery and increases the likelihood that the shock delivery will be followed by return of spontaneous cardiac rhythm and ROSC.

Cardiopulmonary Resuscitation Sequence

The appropriate sequence for pediatric resuscitation is determined by the type and location of the arrest and the type of rescuers and equipment available. A lone HCP must choose a sequence of actions, while multiple rescuers are able to accomplish several resuscitation steps simultaneously.

For the out-of-hospital arrest of a child or likely victim of asphyxial arrest, CPR focuses on initiation of high-quality conventional CPR. The lone HCP (with no nearby help or mobile phone) delivers about 5 cycles or about 2 minutes of CPR before leaving the victim to activate the emergency response system and retrieve an AED. The rescuer then returns to the victim to resume CPR and use the AED. If multiple bystanders are present during the out-of-hospital arrest, 1 rescuer remains with the victim to provide high-quality CPR while others activate the emergency response system and retrieve an AED.

During an in-hospital arrest, multiple rescuers are generally present to accomplish many tasks at the same time. The first HCP to identify cardiac arrest calls for nearby help and initiates resuscitation. As additional rescuers arrive with equipment, they will perform tasks such as providing bag-mask ventilation, attaching the adhesive defibrillator electrode pads and operating the defibrillator, establishing vascular access, serving as an alternate compressor, recording the resuscitation events, and functioning as team leader.

High-Quality Cardiopulmonary Resuscitation Technique

The elements of high-quality CPR are compressions of adequate rate and depth, complete chest recoil after each compression, minimal number and duration of any interruptions in chest compressions, and avoidance of excessive ventilation.

During CPR, the child should be placed supine on a firm, flat surface. A backboard is generally used as this surface.

Chest Compressions of Adequate Rate and Depth

Chest compressions create blood flow by increasing intrathoracic pressure and by directly compressing the heart. Myocardial blood flow is determined by the coronary perfusion pressure present between compressions. Coronary perfusion pressure is calculated as the difference between aortic end-diastolic pressure (during CPR, this is the aortic relaxation pressure) and the mean right atrial pressure; the mean CVP may be used as a surrogate for right atrial pressure. In adults with cardiac arrest, a coronary perfusion pressure greater than 15 mm Hg has been linked with ROSC. Although a similar threshold has not been established in children, if arterial and right atrial or central venous monitoring catheters are in place, rescuers should attempt to optimize coronary perfusion pressure by maximizing aortic relaxation pressure and minimizing right atrial and central venous pressure (see “Monitoring During Cardiopulmonary Resuscitation,” below).

During CPR, blood flow, mean aortic relaxation pressure, and mean coronary perfusion pressure are increased by increasing the rate and force of chest compressions, and these variables are reduced when compressions are of either inadequate or excessive rate or inadequate depth. If the compression rate is too slow, blood flow generated by the compressions is inadequate. If the chest compression rate is too rapid, the depth of compressions may be compromised and chest recoil between compressions may be incomplete. The AHA-recommended chest compression rate for infants, children, and adults is 100 to 120 per minute.

The sternum is depressed approximately one-third the depth of the chest in the newborn and at least one-third the depth of the chest (about 1.5 in, or 4 cm) in the infant. The lone rescuer uses 2 fingers to compress the newborn or infant sternum, just below the intermammary line. When 2 HCPs are delivering CPR to the newborn or infant, the compressor uses the 2-thumb-encircling hands chest compression technique. The compressor’s 2 thumbs are placed together on the sternum to depress the sternum (the thumbs can be placed side by side or they may overlap), while the fingers of both hands spread to encircle the thorax. The 2-thumb-encircling hands technique generates better coronary perfusion pressure and enables creation of more accurate depth of compression, and it may produce higher systolic and relaxation pressures with less rescuer fatigue than the 2-finger compression method.

Chest compressions in the child 1 year of age to puberty may be accomplished with 1 or 2 hands, whichever allows compression of the lower third of the sternum to a depth of at least one-third the depth of the chest and about 2 inches (5 cm). For chest compressions in the adolescent (after puberty) the traditional adult 2-hand technique is used to compress over the lower third of the sternum to a depth of 2 to 2.4 inches (5–6 cm). While excessive compression depth in adults can cause injuries such as rib fractures, inadequate compression depth is the more common problem and clearly reduces blood flow, coronary perfusion pressure, ROSC, and even survival from cardiac arrest.

Many CPR feedback devices are commercially available to provide visual and auditory indication and documentation of rate and depth of chest compressions and other resuscitation variables. These devices may be used for training and in actual resuscitations (see “Continuous Quality Improvement,” below).

Complete Chest Recoil after Each Compression

If rescuers fail to allow the chest wall to recoil completely after each compression, venous return to the heart is decreased, elevating right atrial pressure and reducingcoronary perfusion pressure. In addition, it prevents venous return to the heart and reduces the blood flow generated by the next compression. Rescuers who become fatigued while providing compressions are more likely to lean on the chest, failing to allow complete chest wall recoil. For this reason, during CPR, rescuers should rotate compressors about every 2 minutes (or more often if the compressor becomes fatigued or compression quality declines), and should monitor for and reduce leaning.

Minimal Interruptions in Chest Compressions

Aortic relaxation pressure and coronary perfusion pressure fall whenever chest compressions are interrupted (eg, when breaths are provided or when compressions are interrupted to check rhythm or deliver a shock). Frequent or prolonged interruptions in chest compressions reduce mean coronary perfusion pressure, ROSC, and survival. Interruptions should, therefore, be minimized during CPR and rescuers should monitor the frequency and duration of any interruptions and strive to reduce them. The chest-compression fraction (CCF) is the total CPR time spent delivering compressions. The CCF should be at least 60% of total CPR time and the resuscitation team should strive for a CCF goal exceeding 80% of total CPR time.

Avoid Excessive Ventilation

It is important to avoid excessive ventilation during CPR. When cardiac arrest is present, pulmonary blood flow, carbon dioxide delivery to the lungs, and oxygen uptake from the lungs are approximately 10% to 33% of normal, so victims need only a fraction of normal minute ventilation to match ventilation to perfusion during CPR. Excessive ventilation is harmful because it will reduce coronary perfusion pressure and blood flow during CPR. It will create positive intrathoracic pressure, reduce venous return to the heart, and reduce coronary perfusion and systemic and cerebral blood flow and survival.

The optimal compression-to-ventilation ratio for pediatric resuscitation is unknown. For pediatric CPR, the AHA recommends a 30:2 compression-to-ventilation ratio for single rescuers and a 15:2 ratio for 2 or more rescuers. Once an advanced airway is placed, rescuers should no longer deliver cycles of compressions and breaths. Instead, the rescuer performing compressions should deliver continuous chest compressions, and the rescuer providing ventilation should deliver approximately 10 breaths per minute (1 breath every 6 seconds).

Continuous Chest Compressions versus Conventional CPR

Since 2008, the AHA has recommended Hands-Only/compression-only CPR for treatment of adult sudden witnessed cardiac arrest, particularly if the bystander is untrained or unwilling to provide mouth-to-mouth ventilation. The purpose of this recommendation was to increase the portion of victims of out-of-hospital cardiac arrest who receive some bystander CPR before the arrival of EMS providers. Because most pediatric cardiac arrest is asphyxial in origin, the AHA continues to recommend conventional CPR for infants and children with out-of-hospital cardiac arrest. This recommendation has been supported by pediatric data from out-of-hospital registries, with conventional CPR associated with higher survival than compression-only CPR, particularly in infants and particularly when asphyxial arrest is present.

Advanced Cardiopulmonary Resuscitation Techniques

Open-chest CPR can produce nearly normal blood flow but is impractical for most resuscitation situations. This technique is often used in the immediate postoperative period for pediatric cardiovascular surgical patients, when the chest is left open or the sternotomy can be opened quickly.

Extracorporeal cardiopulmonary resuscitation (ECPR) is the use of extracorporeal membrane oxygenation (ECMO) during attempted resuscitation. Extracorporeal cardiopulmonary resuscitation has been successful for pediatric in-hospital resuscitation, particularly among children with heart disease (see Chapter 108). The timely implementation of ECMO may also be effective for preventing cardiac arrest among children with low cardiac output (eg, following cardiovascular surgery), septic shock, myocarditis, and cardiomyopathy.

Monitoring during Cardiopulmonary Resuscitation

Cardiopulmonary Resuscitation Feedback Devices

Several elements of CPR quality may be monitored and recorded, with devices that provide visual and auditory signals. These feedback devices range from relatively simple accelerometers to more complicated feedback defibrillators. The thresholds provided on the defibrillator screens and in recordings are typically consistent with a minimum 2-inch (5-cm) depth of acceptable chest compressions. These feedback defibrillators can be used to monitor the ECG, shock delivery, depth and rate of compressions, frequency and volume of breaths, and frequency and duration of any interruptions in chest compressions. This information is critical for team debriefing following every arrest as well as to set goals and monitor progress for improving CPR quality and survival in any resuscitation system.

Exhaled or End-Tidal Carbon Dioxide Tension

Under normal conditions, the carbon dioxide tension in exhaled gas should be approximately equal to the carbon dioxide tension in both alveolar gas and pulmonary capillary blood. In very low flow states, if minute ventilation is fixed, the end-tidal carbon dioxide (CO₂) pressure (P_ETCO2) will vary with cardiac output. When cardiac arrest develops, blood flow to the lungs stops, so the CO₂ tension in the pulmonary capillaries, alveoli, and exhaled gas approaches zero. When effective chest compressions generate adequate pulmonary blood flow, the CO₂ delivered to the lungs and the partial pressure of CO₂ in exhaled gas should rise. ROSC is associated with a sharp and sustained rise in the P_ETCO2.

In intubated adult victims of cardiac arrest, a P_ETCO2 less than 10 mm Hg despite 20 minutes of resuscitation is associated with very low survival. Although a similar threshold has not been identified in children, if the P_ETCO2 remains less than 10 mm Hg, providers should make additional efforts to optimize CPR quality. Waveform capnography should be used for these measurements rather than intermittent colorimetric exhaled CO₂ monitors.

Estimating Coronary Perfusion Pressure

Most hospitalized children who develop cardiac arrest are in the ICU (rather than in the general care ward), and invasive monitoring catheters may already be in place. If an intra-arterial and a central venous or right atrial catheter are in place, the coronary perfusion pressure can be estimated by subtracting the right atrial or central venous pressure from the aortic relaxation pressure (it will appear as the aortic diastolic pressure). The CPR technique should be optimized to maximize the estimated coronary perfusion pressure.

Defibrillation

Because AED algorithms have been shown to be accurate in interpreting pediatric rhythms, the AHA recommends AED use for children ages 1 to 8 years in cardiac arrest, particularly for out-of-hospital arrest. Ideally, rescuers should use an AED with pediatric pads and a pediatric dose attenuator to deliver a shock dose appropriate for a small victim. However, if an AED with pediatric pads and dose attenuator is not available, rescuers should use an AED with adult pads. The pads should be placed according to the illustrations on the pads. They should not touch or overlap.

When an infant requires defibrillation, the use of a manual defibrillator is preferred to the use of an AED, because the dose can be more closely adjusted. If a manual defibrillator is not available, an AED with pediatric pads and a pediatric dose attenuator should be used. If such an AED is not available, an adult AED (one that administers an adult dose through adult pads) should be used.

When both pediatric and adult pads are available with the AED, it is important to use pediatric pads only for victims less than 8 years of age. If the pediatric pads are used for an older child, an adolescent, or an adult, they will likely attenuate the shock dose to one that is too small to eliminate VF in the larger victim.

For in-hospital manual defibrillation, rescuers should use infant pads or paddles for attempting defibrillation in infants up to 1 year of age and should use larger pads or paddles for patients 1 year of age and older. Rescuers should use the pads recommended by the defibrillator manufacturer.

Most commercially available defibrillators now use a biphasic waveform, while older defibrillators used a monophasic waveform. There are very limited data available regarding optimal manual biphasic waveform shock dose for defibrillation of children. In pediatric data from the GWTG-R in-hospital registry, an initial dose of 1 to 3 J/kg was associated with higher ROSC than a dose of more than 3 to 5 J/kg. The AHA recommends an initial shock of 2 J/kg. If the 2 J/kg initial dose fails to eliminate VF, rescuers should use a 4 J/kg dose for the second shock. For subsequent shocks, a dose of 4 J/kg is reasonable and a higher dose (up to a maximum of 10 J/kg) may be considered.

More research regarding the optimal defibrillation dose is needed. In general, weight-based defibrillation dosing may not be optimal, because the dose or current delivered to the myocardium is influenced not only by paddle or pad size and energy dose but also by transthoracic impedance, and transthoracic impedance is not linearly related to body weight. In the future, current-based defibrillation would make more sense than energy-based defibrillation.

Organization of Pediatric Resuscitation

Resuscitation is organized around 2-minute periods of uninterrupted CPR. Once a defibrillator is attached, the rhythm is checked every 2 minutes and a shock is delivered if pVT/VF is present. When CPR is interrupted for shock delivery, rescuers should have the defibrillator charged and should deliver the shock within 10 seconds or less of the last chest compression, then should resume chest compressions immediately after shock delivery. Shock effectiveness (ie, likelihood of termination of VF and ultimate ROSC) decreases with every 10 seconds that elapse between the last compression and the shock delivery. When a shock eliminates VF, the most common rhythm for 30 to 60 seconds after shock delivery is asystole or PEA, so compressions are needed immediately after shock delivery.

Advanced Airway

During resuscitation, the potential benefits of inserting an advanced airway should be weighed against the detrimental effects of the intubation attempt itself. Bag-mask ventilation for short periods may be as effective as ventilation through an advanced airway, but this form of ventilation requires training and frequent retraining. Inserting an advanced airway will enable delivery of uninterrupted chest compressions and may reduce gastric insufflation (and its attendant risks of regurgitation and aspiration), but it does require training and experience and will require interruption of chest compressions.

During resuscitation, the most experienced provider should perform intubation, with careful preparation and coordination of rescuer activities to minimize interruptions in chest compressions. Once the advanced airway is inserted, rescuers should also confirm placement using clinical examination plus a device such as a quantitative waveform capnography. If waveform capnography is not available, a qualitative colorimetric exhaled carbon dioxide detector may be used after several positive pressure breaths have been given. If the child weighs more than 20 kg, rescuers may use esophageal detector devices to confirm tracheal tube placement. These devices use suction to determine the ease with which gas can be withdrawn from the tube; the esophagus collapses easily under suction, preventing gas withdrawal, but the trachea is rigid and allows easy withdrawal of gas. Rescuers should verify correct tube position when the patient is transported and, during postcardiac arrest care, whenever the intubated child suddenly deteriorates.

Drug Therapy

Although no drug has been shown to increase survival from pediatric cardiac arrest, the use of vasoconstrictors has been shown to increase blood pressure and coronary and cerebral blood flow and return of spontaneous cardiac rhythm in animals. Drugs with beta-adrenergic effects also increase spontaneous myocardial depolarization and contractility. The AHA recommends administering a standard dose of intravenous (IV) epinephrine (0.01 mg/kg) every 3 to 5 minutes during cardiac arrest.

During resuscitation, intravenous or intraosseous drug administration is preferable to endotracheal administration (see Chapter 100). Although lipid-soluble drugs can be administered by the endotracheal route, drug absorption is poor and unpredictable, and optimal drug doses for this route of administration are unknown. In fact, the lower blood concentrations of epinephrine resulting from endotracheal administration could produce undesirable vasodilatory β-2 adrenergic effects.

Intravenous high-dose epinephrine (HDE) is no longer recommended for routine use in pediatric resuscitation, because it is associated with decreased survival and neurological outcomes. Although HDE can increase the return of spontaneous cardiac rhythm, it can also increase postcardiac arrest myocardial oxygen consumption, myocardial dysfunction, and hemodynamic instability.

When shock-refractory VF is present, either amiodarone (5 mg/kg dose, to a maximum of 3 doses) or lidocaine (1 mg/kg dose) may be given as an antiarrhythmic. Because amiodarone can produce hypotension and arrhythmias, expert consultation is advised when the drug is considered for treatment of prearrest or postcardiac arrest arrhythmias.

Magnesium administration is indicated for treating torsades de pointes and when hypomagnesemia is documented or strongly suspected. However, it is no longer used routinely during resuscitation.

Termination of Resuscitation Efforts

There are no intra-arrest factors that reliably predict poor resuscitation outcomes. In the out-of-hospital setting, age less than 1 year, unwitnessed arrest, no bystander CPR, and nonshockable rhythm on EMS arrival are associated with poor outcome. In the in-hospital setting, age greater than 1 year and longer duration of cardiac arrest have been associated with poor outcome. However, no single factor predicts outcome, and recent emphases on CPR quality, and improved postcardiac arrest care could all have a significant positive impact on CPR outcomes.

Postcardiac Arrest Care

Following ROSC, ischemic and reperfusion injuries can cause cardiorespiratory instability, perfusion abnormalities, and organ dysfunction. Reperfusion injury is characterized by calcium entry into cells, activation of inflammatory mediators, and cell death. This inflammatory response includes the development of endothelial injury, capillary leak, neutrophil activation, platelet aggregation, and increases in mediators such as tumor necrosis factor and interleukins. Increased production of free oxygen radicals and decreased production of nitric oxide can cause vasoconstriction and further ischemic injury. Hyperglycemia is common in children after an arrest, and both extreme hyperglycemia and hypoglycemia have been linked with increased mortality in children.

Skilled postcardiac arrest care in adults has more than doubled neurologically intact survival to hospital discharge. Such care requires evidence-based and protocol-driven bundled care, including targeted temperature management and support of organ system function.

Following ROSC, rescuers must begin targeted temperature management, optimize hemodynamic support, titrate inspired oxygen administration, and support of ventilation to maintain normoxemia (unless cyanotic heart disease is present) and partial pressure of carbon dioxide (PaCO₂) appropriate for that patient, and support end-organ function. The use of protocols will support consistent care that can then be evaluated and modified as needed.

Children often demonstrate a brief period of spontaneous hypothermia followed by the development of hyperthermia after cardiac arrest. While hyperthermia is associated with worse neurological outcome, the recent Therapeutic Hypothermia to Improve Survival After Cardiac Arrest in Pediatric Patients (THAPCA, see http://www.THAPCA.gov) clinical trials network failed to show any benefit of induced hypothermia when compared to normothermia after out-of-hospital cardiac arrest. Results following in-hospital cardiac arrest are anticipated soon. Thus, at present, it is appropriate to recommend controlled normothermia and avoidance of hyperthermia in this at-risk population.

Continuous Quality Improvement

Every resuscitation program or setting must develop a process of continuous quality improvement. This process requires measurement of CPR quality, monitoring and review of resuscitation outcomes and techniques, and frequent opportunities for retraining. If the resuscitation team is not measuring performance, they cannot hope to improve it.

The availability of CPR courses for lay rescuers, including “CPR in schools” programs, can increase bystander CPR for out-of-hospital cardiac arrest. However, in addition to structured training programs, training opportunities must expand beyond the classroom. Opportunities to learn about CPR and to practice compressions are now offered to the lay public through kiosks in locations such as airports and through distribution of inflatable take-home manikins with instructions provided on a DVD.

All EMS dispatchers should use protocols to enable them to rapidly identify a likely cardiac arrest and guide untrained rescuers to initiate CPR; their effectiveness must be evaluated by documenting time to cardiac arrest identification and time to CPR initiation as part of the EMS system of continuous quality improvement. Voluntary registration of bystanders willing to perform CPR has enabled a type of “crowd sourcing” that has been effective in summoning rescuers to the scene of an out-of-hospital arrest. Smartphone programs are also available to guide rescuers through the steps of CPR by pacing compressions to the correct rate and reminding rescuers of CPR sequences and steps. These and other innovations are needed to increase the likelihood and quality of bystander CPR.

It is now clear that HCPs must regularly practice the skills and teamwork of resuscitation. However, the precise training duration and retraining interval needed to achieve and maintain resuscitation skills has not been established, and likely differs among HCPs who perform resuscitation frequently and those who rarely perform resuscitation. Providers need opportunities to refresh skills, particularly “just in time” training. Training with the use of realistic simulators holds promise to improve HCP skills and performance.

Team debriefing immediately following an attempted resuscitation is essential to team improvement in subsequent resuscitations. Such debriefing requires objective and professional discussion of team performance and identification of opportunities for improvement.

Prevention of Arrest

Ideally, successful treatment of prearrest conditions will prevent arrest or at least minimize potential organ ischemia before arrest ensues. In the in-hospital setting, a pediatric cardiac arrest outside of the ICU should be a rare event, and should trigger careful evaluation of the child’s prearrest care and remediation of any shortcomings in assessment or treatment. Although rapid response or medical emergency teams may reduce the incidence of non-ICU pediatric arrests, the published success rates of these teams have varied widely.

SUGGESTED READINGS

INTRODUCTION

In many countries, pediatric medical and surgical specialty services are regionalized to either community hospitals or tertiary and quaternary pediatric centers. Thus, critically ill children in need of these services may need to be transported considerable distances from their home communities in order to receive higher levels of care.

The main goal of successful patient transport is to transfer the patient to definitive care while preventing further deterioration. In order to achieve this goal, close monitoring, continued resuscitation, and time sensitive treatments may be necessary. However, there are inherent risks related to the transport, such as exposure to weather-related hazards, transport vehicle mishaps, and difficulties with patient monitoring and assessment due to the noisy, cramped, moving environment. In order to mitigate these risks and optimize patient care, there are many measures that can be taken. First, prompt telephone or video conferencing with specialized pediatric healthcare providers can guide patient management decisions and ensure that urgent medical interventions are initiated. Second, early discussion enables mobilization of resources and decisions on the appropriate mode of transportation. These discussions should address what equipment will be required in order to adequately and efficiently prepare for transport. Third, selecting the appropriate healthcare team to accompany the patient is imperative. The transport of critically ill children and neonates by specialized pediatric and neonatal transport teams has resulted in fewer adverse events and improved outcomes than those transported by providers without this specialized training. Given that most specialized teams are located in central areas, the time it takes for retrieval teams to reach the patient must be factored into the total transport time, and thus furthers the argument that prompt referral for transport is imperative. In the following, we will discuss important considerations of stabilization and transfer, including the timing of referral, who to refer, and how to mitigate the risk of transport.

PATHOGENESIS AND EPIDEMIOLOGY

When Is Patient Transport Necessary?

Generally, patients are transported from one facility to another to receive a higher level of care based on their medical diagnosis and on their diagnostic and treatment requirements. Decisions regarding when and where to transfer a patient are based on the local and regional resource availability, as well as patient acuity. Due to the centralization of specialized services, some children also may be transported electively for routine outpatient appointments. Patients that present to remote nursing stations or hospitals without the ability to admit pediatric patients may require semi-urgent transfer to seek general pediatric expertise. Other patients may require urgent and emergent transfer to tertiary or quaternary hospitals for specialized diagnostic or surgical services, subspecialty evaluation, or admission to a neonatal intensive care unit (NICU) or pediatric intensive care unit (PICU).

Selecting the Optimal Mode of Transport

Local resources and infrastructure determine the mode of transport used to transfer children. In developing countries, patients may be transported by ambulance, as well as taxis, auto-rickshaws, bicycles, motorcycles, and even wheelbarrows. In developed countries, the discussion of patient transfer is usually limited to land and air ambulance, including fixed-wing and helicopters. Table 103-1 lists the characteristics of land and air transport. Appropriately triaging patients to a suitable mode of transport requires discussion of local resources, weather and traffic conditions, time of day, geography, and patient acuity. It is an area of ongoing research, particularly for trauma patients.

Issues Related to Altitude

If the patient’s location and/or acuity merit transport by helicopter or fixed-wing aircraft, there are a few physiologic considerations. At increased altitude, the barometric pressure decreases. Given that the partial pressure of oxygen (PO₂) is determined by the barometric pressure and fraction of inspired oxygen (FiO₂), at higher altitudes the PO₂ available at the alveolar level decreases. This causes mild desaturation in healthy individuals. In critically ill patients with significant oxygenation defects (such as those requiring 100% FiO₂), flying at high elevations is risky and efforts to improve oxygenation (eg, endotracheal intubation and effective ventilation) should be taken prior to transport.

Aeromedical transportation is subject to Boyle’s law, which describes the inverse relationship between pressure and volume, such that as pressure decreases the volume increases. Thus, as barometric pressure decreases, volume-filled spaces will naturally expand. Therefore, during flight, pneumothoraces, intracranial air, air in abscesses, or gastric air bubbles may expand with deleterious consequences. Again, this should be anticipated, and efforts to prevent complications should be considered, including interventions (such as chest tube and nasogastric tube insertion) or selecting an alternate mode of transport. These considerations are more relevant when using fixed-wing air transport because of the greater altitude.

CLINICAL MANIFESTATIONS

In order to successfully transport a patient from one facility to another, anticipation of care needs prior to catastrophic deterioration is needed. This involves some degree of clinical judgement and insight, and can sometimes be difficult to predict. However, there are some conditions that require direct admission to a PICU (such as meningococcemia or severe diabetic ketoacidosis) or evaluation at a pediatric center (such as significant pediatric trauma), and thus prompt referral should always be a priority. Regardless of the reason for patient presentation, given the logistical and geographic considerations of regional patient transfer, if a practitioner suspects the need for transfer, telephone consultation with the appropriate specialist is important to facilitate early activation of the appropriate transport team.

DIAGNOSIS

Critically ill pediatric patients who require transport present with deranged physiology due to many disease states. Commonly, patients are transported after major trauma for treatment at a pediatric trauma center, or for other medical considerations such as cardiac disease (including congenital heart disease and myocarditis), respiratory illness (eg, bronchiolitis and pneumonia), infectious causes (eg, sepsis and meningitis), and neurological conditions (including seizures and traumatic brain injury). Newborns are transported for many reasons, including prematurity, transitional conditions requiring support (eg, respiratory distress syndrome, transient tachypnea of the newborn, and meconium aspiration), workup of congenital malformations, or other illnesses (eg, sepsis and hypoxic ischemic encephalopathy). The decision regarding when and whom to transfer will depend on a case-by-case discussion between the presenting and advising/receiving centers.

TREATMENT

When an acutely ill pediatric patient presents in the community setting, the patient should be stabilized following a structured assessment and management process that focuses on airway, breathing, and circulation, as directed by the Pediatric Advanced Life Support or Advanced Trauma Life Support algorithms. Early telephone consultation with a consulting specialist is recommended. Prior arrangements and clinical needs may dictate who should be consulted and may include any of a general pediatrician, pediatric emergency physician, neonatologist, or pediatric intensivist. The consultation should include discussion of the appropriate treatment and diagnostic modalities that are required prior to transfer, and what the appropriate transport process should be. The processes by which these calls are initiated vary depending on local health system logistics and design, and all practicing clinicians should familiarize themselves with local transfer processes for their region.

The “scoop and run” philosophy (meaning achieving as rapid a turnaround as possible, often for the sickest and least stable patients) has long been seen as inadequate in the care of pediatric patients who require interfacility transport. Adequate preparation prior to transport is imperative. It has been well established that children are at risk of further physiologic deterioration during transport. In addition, system errors and equipment failure can complicate their transport and impact their care. Thus, adequate resuscitation prior to departure and anticipation of possible further deterioration is important and may necessitate aggressive, prophylactic interventions. Medical procedures are extremely challenging to perform in the transport environment where there is limited patient access, fewer medical personnel and less equipment available, environmental hazards (eg, temperature extremes), and other factors that make patient assessment and intervention challenging (eg, noise and vibration interference). Thus, decisions to perform interventions prior to transport are based on the risk of deterioration outweighing the risk of performing the intervention (including consideration of the time taken for these), even if less skilled personnel are available to perform them. These decisions should be made in conjunction with the receiving physician. For example, a patient with decreased and fluctuating level of consciousness or declining respiratory function from pneumonia may be endotracheally intubated for transport earlier than their clinical status would otherwise mandate it. Consideration of the mode of transport available should also be a factor when discussing prophylactic interventions. For example, a small pneumothorax may require a chest tube prior to aeromedical transfer given the risk of expansion en route.

Critically ill neonates and pediatric patients transported by providers with specialized training in neonatology or pediatrics, respectively, versus nonspecialized providers have repeatedly been shown to experience lower adverse event rates during transport, and have improved clinical outcomes. These specialized team members can also assist with the pretransport stabilization process and in any required interventions. Transport teams differ in provider composition, and involve a combination of nurse, paramedic, respiratory therapist, and physician providers, depending on patient acuity and local practices. Teams that have physician accompaniment on transport have not been demonstrated to be superior to those by trained non-physician transport providers with physician consultation, but studies restricted to severely ill children have not been done.

Prior to transfer, transport teams must ensure that they have adequate medication and equipment to monitor and manage a patient during transport, with reasonable provision for unexpected delays (eg, due to weather, mechanical issues, traffic). In addition, copies of all pertinent patient documentation, laboratory results, and diagnostic examinations should be prepared so that the receiving facility can continue, not reproduce, care. The use of checklists (Table 103-2) and protocols for sending facilities and transport teams can be helpful to standardize and streamline this process. Parental accompaniment on transport may improve the emotional well-being of the patient and the parent.

PREVENTION

Transferring a patient between facilities can be resource intensive and can place the patient and the transport providers at significant risk. Thus, efforts to mitigate patient risk (as discussed above) should be employed. Early resuscitation and consultation with pediatric providers, as well as the possible institution of telemedicine, may aid in diagnostic evaluation of patients and in the triaging of patients who require transfer to the appropriate transport providers.

SUGGESTED READINGS

INTRODUCTION

Respiratory support is the mainstay in caring for children in respiratory failure (RF) and an integral part of pediatric critical care. Children, especially in the first few years of life, are more prone to RF than are adults, due to anatomic and physiologic differences and, thus, it continues to be a common indication for hospitalization in this age group.

Respiratory failure is defined as the failure to maintain adequate oxygenation (hypoxemia), ventilation (hypercarbia), or both (mixed). The etiology and pathophysiology of acute respiratory dysfunction and RF are discussed in Chapter 96. In this chapter, the fundamentals of respiratory physiology will briefly be addressed in order to understand the modes of respiratory support and their application in children with RF and will be followed by discussions of modes of respiratory support in children with RF.

BASIC RESPIRATORY PHYSIOLOGY

The respiratory system must overcome elastic forces (lung and chest wall) and resistive forces (airways) to achieve gas flow. During spontaneous breathing, diaphragmatic contraction causes a drop in pleural and intrathoracic pressure resulting in air being “pulled” into the lungs, followed by passive recoil of the lungs and chest wall during exhalation. During positive-pressure ventilation (PPV), on the other hand, air is “pushed” into the lungs during inspiration, followed by passive recoil of the lungs and chest wall during expiration. Positive-pressure ventilation is, thus, by definition, nonphysiologic.

In healthy people, the work of breathing (WOB) is minimal, and the respiratory system works efficiently to meet the metabolic demands of the body. In disease states, however, the compliance and resistance of the respiratory system changes, causing increased demands on the respiratory system with resultant increased WOB. Elevated airway resistance in conditions such as bronchiolitis and asthma causes airflow limitation during expiration and hyperinflation with resultant hypercarbia due to increased pulmonary dead space; acute respiratory distress syndrome (ARDS) is associated with poor lung compliance with a tendency of alveolar units to collapse (atelectasis), contributing to ventilation-perfusion mismatch and intrapulmonary shunt, resulting in hypoxemia.

MODES OF RESPIRATORY SUPPORT

The primary aim of respiratory support is to improve gas exchange while minimizing the patient’s work of breathing. The full spectrum of commonly used respiratory support includes simple supplemental oxygen therapy; mechanical ventilation, including noninvasive positive pressure ventilation; invasive mechanical ventilation; and advanced and adjunct ventilator therapies. Extracorporeal membrane oxygenation (ECMO) is the therapy of choice when conventional respiratory support has failed (see Chapter 108) While mechanical ventilation can be achieved using negative or positive pressure, negative pressure ventilation (NPV), although appealing in concept and having inherent potential hemodynamic advantages, has failed to gain popularity as a practical mode of support in the intensive care unit (ICU), largely due to the disadvantages of limited physical access to the patient, lack of a protected airway, and challenges with secretion clearance. Thus, the discussion of MV in this chapter will focus on PPV.

Supplemental Oxygen Therapy

Nasal Cannula Oxygen

Nasal cannula oxygen is the most commonly used method to provide supplemental oxygen to patients in hypoxemic RF and is generally well tolerated. Standard nasal cannulas are limited by flow, and rates greater than 6 liters per minute (lpm) are generally not recommended due to limitations of the bubble humidification systems. The delivered fractional concentration of inspired oxygen (FiO₂) to the patient is variable due to entrainment of room air. While complications are infrequent, failure to provide humidification can result in erosion of the nasal septum and epistaxis.

Face Mask Oxygen

Face mask oxygen is an effective and simple mode of delivering oxygen to patients but is not as well tolerated by children over longer time periods as is a nasal cannula. The mask limits the ability to eat and drink while receiving supplemental oxygen and in general is used for short time periods, such as in an emergency room or acute care clinics. The ability to achieve high FiO₂ delivery is again limited due to entrainment of room air. However, a nonrebreather version of the face mask with an oxygen-rich reservoir can be used when higher FiO₂ delivery is desired, as a temporizing measure until more definitive interventions can be implemented.

High-Flow Nasal Cannula

A high-flow nasal cannula (HFNC) delivers a high flow of humidified and heated oxygen through nasal prongs. Depending on the gas delivery device and the cannula, flows up to 60 lpm can be attained for respiratory support. The HFNC was introduced as an alternative to nasal continuous positive airway pressure (CPAP) in neonates, but it is increasingly being used for patients with hypoxemic RF across all ages. Some of the postulated mechanisms of benefit include washout of anatomic dead space, subjective relief of dyspnea from meeting the inspiratory flow demand of patients, improved FiO₂ delivery by reducing the entrainment of room air as the delivered flow rates approach spontaneous demand, stenting of upper airways, and improved lung recruitment through generation of CPAP. However, HFNC should not be considered as noninvasive positive-pressure ventilation (NIPPV), and its use is generally discouraged in hypercarbic or severe hypoxemic RF.

HFNC support has been used in neonates to prevent endotracheal intubation, to support infants and children with bronchiolitis, to assist children and adults in hypoxemic RF, to provide increased oxygenation prior to intubation, and to reduce reintubation rates following extubation. HFNC support is generally well tolerated and has minimal complications. Frequent assessment of patients on HFNC is mandatory, and failure to detect clinical deterioration without timely interventions can result in poor outcomes. The decision to place infants and children on HFNC in non-ICU settings should, thus, be made carefully by the clinical care team.

Noninvasive Positive-Pressure Ventilation

Noninvasive positive-pressure ventilation is the delivery of mechanical ventilation without an endotracheal tube or tracheostomy. It includes both CPAP and bilevel ventilation. It requires a patient interface, such as a mask, and a machine to deliver the support. The aim is to provide respiratory support while avoiding the perils of invasive ventilation, such as sedation and ventilation-associated pneumonia (VAP). It can be used to support both acute and chronic RF and is helpful in both hypoxemic and hypercarbic RF. Additionally, NIPPV can provide cardiopulmonary support for patients with acute, chronic, or acute-on-chronic heart failure. It can be safely delivered at home, for either continuous or intermittent (eg, nighttime) use, but this requires careful training of caregivers and appropriate adaptation of the home environment.

Continuous Positive Airway Pressure

Continuous positive airway pressure is provided throughout the respiratory cycle while the patient is breathing spontaneously. Advantages include stenting of the upper airway and tracheobronchial tree, lung recruitment, and stenting of the lower airways in obstructive lung disease. CPAP is clinically used for patients with obstructive sleep apnea, tracheobronchomalacia, atelectasis, heart failure, and asthma with variable levels of supportive data and success. The keys to success are choosing an optimal interface based on patient tolerance and titration of pressure until the desired result is achieved.

Bilevel Positive Airway Pressure

Bilevel positive airway pressure (BiPAP) provides CPAP throughout the respiratory cycle and, in addition, augments inspiration with volume or pressure support. Delivery of inspiratory support is triggered by the patient’s respiratory effort to promote patient-ventilator synchrony. Difference between the 2 levels of pressure drives gas flow within the system. Tidal volume is determined by the driving pressure, extent of leak within the system, and compliance and resistance of the respiratory system. Bilevel positive airway pressure can be delivered using a variety of nasal prongs, or nasal or full-face masks, which interface with either regular ICU ventilators or dedicated NIPPV machines. Specialized machines tend to compensate better for system leaks and may have better patient-ventilator synchrony.

Bilevel positive airway pressure can be used effectively in a wide variety of clinical scenarios and has improved the ability to support neonates and children in RF without the need for intubation. Support with BiPAP can be effective and should be instituted early in patients failing nasal cannula or HFNC therapy. Early institution of positive pressure can prevent derecruitment and worsening hypoxemia.

Use of an oronasal or full-face mask is generally preferred in children, and heated humidification is strongly recommended. Sedation can be used with caution in appropriately monitored settings to enable tolerance of NIPPV and promote patient-ventilator synchrony. It is common to observe patients in RF with increased WOB subsequently (often rapidly) calm down and relax following initiation of BiPAP, as it improves their respiratory mechanics and unloads the respiratory muscles. Complications of BiPAP include skin breakdown and pressure ulcers due to the masks, gastric distension, emesis and aspiration, barotrauma/volutrauma, diminished venous return from positive intrathoracic pressure, and patient-ventilator asynchrony. In addition, continuous BiPAP creates challenges with verbal communication and oral feeding.

Invasive Mechanical Ventilation

Invasive mechanical ventilation can be delivered via an endotracheal tube or tracheostomy. Endotracheal intubation (orotracheal or nasotracheal) is routinely undertaken in emergency rooms, ICUs or operating rooms for children in whom NIPPV is not indicated or has failed. While use of cuffed endotracheal tubes was prohibited in the past, new low-pressure, high-volume cuffed tubes are safe and used routinely in pediatric ICUs. The key is to use appropriately sized tubes using the formula age/4 + 4 for uncuffed tubes and age/4 + 3 for cuffed tubes. Failure to pay attention to cuff pressures and/or leak and tube sizes can result in consequences to the subglottic space, including erosion, ulceration, and cartilage necrosis, with subsequent scarring and the potential need for airway reconstruction surgery and/or tracheostomy.

The intubation of children with RF is more challenging than intubation in well children in the operating room. The incidence of adverse events is higher in critically ill children and is being studied through the National Emergency Airway Registry for Children (NEAR4KIDS) initiative. Direct laryngoscopy remains the standard in most PICUs; however, use of videolaryngoscopy is gaining in popularity, as it provides the opportunity to educate learners and may improve intubation success rates in children with difficult airways.

Intermittent Positive-Pressure Ventilation

Intermittent PPV involves provision of PPV via an endotracheal tube or tracheostomy. A wide variety of modes are available without data to support superiority of 1 mode over another. A detailed discussion of all modes of ventilation is beyond the scope of this chapter, but we will discuss basic principles and common modes of ventilation.

Basic Principles of Invasive Ventilation

Modern ventilators are sophisticated machines with the ability to deliver high levels of support consistently. Despite several decades of research on this topic, no single mode of support has been noted to be superior; thus, in most cases, tailoring the mode of ventilator support to the physiologic needs of the patient, while at the same time using lung protective strategies, is the most appropriate approach.

In physics, force is measured as pressure (pressure = force/area), displacement as volume (volume = area × displacement), and the relevant rate of change as flow (average flow = Δ volume/Δ time). With respect to respiratory physiology, we are interested in the pressure necessary to achieve gas flow into the lungs. Pressure, volume, and flow are, thus, measurable variables in the respiratory cycle, and their relation is described by the equation of motion for the respiratory system:

Elastance is the inverse of compliance, and, thus, the respiratory system equation of motion can be written as P = V/C + R × F.

Modern positive-pressure ventilators typically support ventilation with the following principal variables:

• Control variable: Pressure, volume, or flow are called control variables. In modern ventilators, it is possible to switch from 1 control variable to another between, or even within, each breath.

• Phase variable: The time between breaths should be divided into 4 phases: the change from expiration to inspiration, inspiration, change from inspiration to expiration, and expiration. This convention helps to understand how a ventilator starts, what sustains and stops a breath, and what it does between inspirations. A particular variable is measured and used to start, sustain, and end each phase, and in this context, pressure, volume, flow, and time are referred to as phase variables.

• Trigger variable: While measuring 1 of the 3 variables within the equation of motion, a breath is initiated when 1 of the variables reaches a preset value. This is the trigger variable and can be pressure (inspiration starts when a drop in baseline pressure is detected), time (set the frequency to start breaths), or flow (detection of gas flow into the lungs triggers inspiration). The patient effort needed to trigger a breath is detected by the sensitivity settings, and, thus, the clinician can make it easier or harder for a patient to breathe based on the clinical situation by adjusting the trigger sensitivity.

• Limit variable: Limit refers to restricting the magnitude of a variable during inspiration. The limit variable is 1 that can reach and maintain a preset level (pressure, volume, or flow) before inspiration ends.

• Cycle variable: The variable that is measured and used to end inspiration is called the cycle variable. The cycle variable is generally flow or time but can be pressure or volume.

• Baseline variable: This variable is controlled during expiration and is measured and set relative to atmospheric pressure. While pressure, flow, and volume can serve as baseline variables, pressure is the most commonly applied baseline variable in the form of positive end-expiratory pressure (PEEP).

Conventional Modes of Ventilation

Modes of ventilation can be characterized according to a particular pattern of mandatory (machine-triggered) or spontaneous (patient-triggered) breaths that are volume controlled or pressure controlled, or have elements of both (dual controlled). Pressure-controlled modes all set a peak inspiratory pressure (PIP) that is targeted during inspiration with rapid delivery of initial gas flow to attain that target, followed by deceleration of flow. Tidal volume attained is variable and depends on the respiratory system compliance and resistance, among other factors. In volume-controlled modes, the primary gas delivery target is tidal volume, and peak flow as well as flow waveform are set with a variable PIP from breath to breath. The differences between pressure- and volume-targeted modes are summarized in Table 104-1, and the ventilator variables are described below.

• Fractional concentration of inspired oxygen (FiO₂): In general, the lowest level of inspired oxygen concentration needed to maintain an oxygen saturations > 92% should be administered in order to avoid oxygen toxicity due to free radicals and denitrogenation. An FiO₂ < 60% is generally considered by most experts to be likely to be less injurious to lung tissue for the majority of patients, although patients with past exposure to medications such as bleomycin are at greater risk from oxygen toxicity.

• Tidal volume (VT) and peak inspiratory pressure: Large VT (> 10 mL/kg) results in alveolar overdistention and ventilator-induced lung injury (VILI). The ARDS network trial published in 2000 demonstrated a survival benefit in adults who were ventilated with low (6 mL/kg IBW) versus high VT (12 mL/kg IBW). While such a study has not been replicated in children, the practice of low VT in pediatric acute respiratory distress syndrome (PARDS) was endorsed by the Pediatric Acute Lung Injury Consensus Conference (PALICC). Most clinicians try to limit the PIP to less than 30 cmH₂O in the absence of significant obstructive lung disease. Plateau pressure (P_plat) is the pressure measured during an inspiratory hold maneuver on the ventilator. It is more reflective of alveolar distending pressure and should ideally be less than 28 cmH₂O. Patients with obstructive lung disease often have a high PIP with P_plat < 28 cmH₂O, suggesting the alveolar distending pressures to be in the safe range.

• Rate, inspiratory time (Ti), and inspiratory-to-expiratory (I:E) ratio: The product of VT and rate constitute the patient’s minute ventilation. The age of the patient and the underlying cause of RF should be considered when setting the rate. Neonates and infants will need faster rates than older children and adolescents. Patients with reduced respiratory compliance (lung units with short time constants) will likely need a faster rate compared to those with obstructive lung disease (lung units with long time constants). The effect of the rate and Ti on the I:E ratio should be closely monitored. As the set rate is increased without changes in Ti, the I:E ratio may become significantly prolonged, causing air trapping and subsequent patient-ventilator dysynchrony. A longer Ti with a faster rate will result in shorter exhalation times and will cause air-trapping, hyperinflation, and the presence of intrinsic PEEP. Patients with obstructive lung disease have prolonged exhalation times and, thus, may benefit from short Ti and low rates on the ventilator, while those with restrictive lung disease may benefit from short Ti and faster rates. In general, Ti between 0.4 and 0.8 seconds is appropriate in children and should be titrated based according to the needs of the individual patient.

• Inspiratory flow: Mean inspiratory flow is a ratio of VT and Ti, and, thus, one cannot change flow without affecting the timing variables. Changing from a decelerating flow to square-wave flow pattern may shorten Ti and the time required to deliver the VT. Inspiratory flow is usually preset in volume modes of ventilation and tends to be variable in pressure modes. As clinicians, it is important to realize the combined effects of flow, timing, and volume on end-expiratory lung volume (EELV).

• Positive end-expiratory pressure: End-expiratory lung volume is critical in patients with reduced respiratory system compliance and is maintained by the application of PEEP. Allowing the lungs to inflate and deflate from optimal EELV reduces cyclic opening and closing of lung units, thus limiting VILI. Identifying optimal PEEP, however, is challenging and should be determined by clinicians at the bedside based on the response at different levels of PEEP on oxygenation, lung mechanics, and hemodynamics. Excessive PEEP can result in lung overdistention, increased intrathoracic pressure with reduced cardiac output, and carbon dioxide retention; low PEEP can cause a reduction in EELV, increased low V/Q segments, and subsequent hypoxemia. The use of PEEP in patients with obstructive lung disease is controversial given an inherent tendency of such patients to air-trap and develop auto-PEEP (PEEPi). In this setting, some patients may benefit from the addition of extrinsic PEEP to overcome PEEPi, while others may benefit from no extrinsic PEEP.

• Mean airway pressure (MAP): Mean airway pressure is the integration of pressure over time and correlates with mean lung volume, alveolar ventilation, arterial oxygenation, hemodynamic performance, and potentially barotrauma. Inspiratory flow, ventilation rate, Ti, PEEP, and PIP/VT all have an effect on MAP, although the largest effects are from PEEP and Ti. It is common to characterize the level of ventilator support based on the MAP value needed to achieve adequate oxygenation.

Synchronized Intermittent Mandatory Ventilation

Synchronized intermittent mandatory ventilation (SIMV) is an intermittent mode of ventilation available with pressure (SIMV-PC) or volume (SIMV-VC) targeting. The breaths are delivered in synchrony with the patient’s inspiratory effort. The gas supply is patient triggered during spontaneous breaths. If a patient inspiratory effort is sensed while the time window is open, a synchronized breath is delivered, and if no patient effort is detected at the time the window closes, a mandatory breath is delivered. This keeps machine inspiration or expiration in phase with the patient’s respiratory cycle and avoids patient-ventilator dysynchrony due to “breath stacking.” Pressure support is usually added to assist patient breaths. Pressure-targeted modes carry the risk of inadequate tidal volume and minute ventilation if lung compliance decreases, while volume-targeted modes maintain tidal volume but may contribute to barotrauma if rising airway pressures are not closely monitored.

Pressure-Regulated Volume Control

Pressure-regulated volume control (PRVC) provides dual control mode of ventilation with between-breath variability and combines features of volume and pressure ventilation. The target tidal volume and maximum pressure level are set. The machine attempts to achieve the volume target using a pressure-control gas delivery format using the lowest possible airway pressure. Over a series of breaths, the ventilator attempts to assess the pressure required to attain the set tidal volume with continuous adjustments between breaths based on a feedback mechanism. In patients with increased inspiratory demand, the delivered pressure may be inappropriately low, resulting in dyspnea, increased respiratory effort, and even cardiorespiratory collapse due to inadequate ventilatory support.

Airway Pressure Release Ventilation

Airway pressure release ventilation (APRV) is often considered in patients with moderate to severe hypoxemic RF and maintains 2 levels of airway pressure (high and low) applied for defined periods of time. Spontaneous breathing is allowed at both levels. Inspiratory time or T_high is much longer than expiratory time or T_low. P_high is the airway pressure maintained during T_high, and P_low is the CPAP maintained during T_low. Airway pressure release ventilation is generally well tolerated from a hemodynamic perspective. Maintenance of spontaneous breathing is essential for the success of this mode. Thus, oversedation and neuromuscular blockade should be avoided.

Pressure Support Ventilation

Pressure support ventilation (PSV) is a pressure-targeted mode in which patient inspiratory effort is supported by the ventilator at a preset level of inspiratory pressure. Inspiration is triggered by the patient and terminated when inspiratory flow falls to a prespecified level (generally 25% of peak flow). Respiratory rate, inspiratory time, and tidal volume are patient determined, while the clinician controls the inspiratory pressure level. Pressure support ventilation is often used by clinicians when weaning from ventilator support or in conditions such as asthma so patients can control their own inspiratory and expiratory times. It is an inappropriate mode for patients with central hypoventilation or apnea.

As briefly mentioned at the start of this section, no mode of ventilation has shown clear benefit over another, and, hence, no specific recommendations can be made. Ventilator support should be tailored to the underlying disease process based on sound physiologic rationale. When constant minute ventilation is desired, such as in patients with traumatic brain injury to maintain PaCO₂ between 35 and 40 mm Hg, a volume-targeted mode may be appropriate, whereas when higher MAPs are desired to optimize EELV, APRV may be more appropriate. Given the complexity of teams involved in the care of critically ill children, standardization of practice and adherence to those standards is the key to achieve success with mechanical ventilation.

Weaning of Ventilator Support and Tracheostomy

Extubation is the ultimate goal of ventilator management after recovery from RF for most patients; however, a few patients who are unable to wean undergo tracheostomies for continued long term ventilator support, if indicated. Tracheostomy offers the advantage of reduced need for sedation, lower risk of VAP, ability of the patient to mobilize and rehabilitate, and easy disconnection from ventilation for those needing only intermittent support. Timing of tracheostomy in the course of RF is a decision that should be made taking into consideration the trajectory of illness, comorbidities, age of the patient, and prognosis of the underlying disease process(es).

Intubated patients should be evaluated for readiness to extubate daily by use of a spontaneous breathing trial (SBT), using PSV, CPAP, or other techniques, if they meet readiness criteria. Those who pass an SBT should be extubated as soon as feasible to limit their time on invasive support. Several indices have been developed in an attempt to predict weaning and extubation success, but available literature suggests that these indices offer no improvement over clinical judgement. Extubation failure rates in children range from 2% to 20% with little relationship to duration of mechanical ventilation, with upper airway obstruction being the most common cause of extubation failure. Early extubation has several advantages that include limited exposure to sedatives, mitigation of ICU delirium, promotion of early mobility, and reducing ventilator-associated complications. Clinicians, however, should be wary of consequences of premature extubation, as extubation failure is associated with worse outcomes in children.

High Frequency Ventilation

High frequency ventilation (HFV) is defined as a frequency that greatly exceeds the normal respiratory rate and has a delivered VT that is at or below anatomic dead space. There are 4 types of HFV: high frequency positive pressure ventilation (HFPPV); high frequency jet ventilation (HFJV); high frequency percussive ventilation (HFPV), and high frequency oscillatory ventilation (HFOV). High frequency modes by definition generate very small VT while maintaining EELV and, thus, have generated appeal by clinicians over time for patients with severe ARDS from the perspective of lung protection. Significant experience with HFV for RF in pediatrics is limited to HFOV. Thus, further discussion about HFV in this chapter will be limited to HFOV.

HFOV is the most commonly used form of high-frequency ventilation in neonates and pediatrics, and has been 1 of the mainstays of advanced ventilation of PARDS over the past 2 decades. While it has been widely used in all age groups, HFOV is falling out of favor in adults in the light of recent negative trials. It can be considered in children with PARDS with P_Plat exceeding 28 cmH₂O in the absence of reduced chest wall compliance. High-frequency oscillatory ventilation has the advantage of maintaining lung expansion through a constant MAP, with very low tidal volumes and is thus seen as a protective form of ventilation particularly in patients with air-leak, pulmonary interstitial emphysema, and high MAP requirements. It is important to note that HFOV is the only commonly used mode in which exhalation is active.

When initiating HFOV, by convention, MAP is initially set 4 to 6 cmH₂O higher than conventional ventilator settings, and optimal lung volume is attained with stepwise increase or decrease in MAP while monitoring oxygenation and ventilation. Frequency is set based on the age and size of the patient, with neonates and infants being at higher frequencies (10–12 Hz) than older children (8–10 Hz) and adolescents (6–8 Hz). The amplitude is adjusted to achieve adequate chest “wiggle” and varies based on the extent of lung disease and size of the patient. Bias flow is typically set at 20 lpm and can be increased to help maintain MAP or to help meet spontaneous respiratory demand. Oxygenation improvement is achieved by stepwise increase in MAP and FiO₂, while ventilation is improved by increasing the amplitude or reducing the frequency. Generally, a small leak is required around the endotracheal tube to optimize ventilation, provided a stable MAP can be maintained.

Disadvantages of HFOV include the typically increased requirements for sedation or even neuromuscular blockade in children and adolescents, as well as the reduced ability to provide adequate pulmonary toilet and limited mobility due to a rigid ventilator circuit. Given that HFOV use in PARDS is associated with higher MAPs to increase EELV, hypotension from reduced venous return is often seen. Traditionally, patients on HFOV are weaned and transitioned to conventional ventilation prior to extubation. Extubation from HFOV after gradual reduction in support is possible but seldom practiced.

Adjunct Therapies

Inhaled Nitric Oxide

Inhaled nitric oxide (iNO) is not recommended for routine use in PARDS. However, its use may be considered in those with documented pulmonary hypertension or right ventricular dysfunction, or in the presence of severe hypoxemic RF as a bridge to ECMO. If used, the response to iNO should be carefully monitored, and it should be discontinued if ineffective. If effective, it should be administered at the lowest effective dose to avoid toxicities such as methemoglobinemia.

Exogenous Surfactant

While trials have suggested that a certain subset of pediatric RF patients may benefit from surfactant therapy, its routine use in patients with PARDS cannot be justified based on pediatric and adult data.

Fluid Management

Pediatric patients in RF should receive fluids to maintain adequate intravascular volume, end organ perfusion, and optimal oxygen delivery. Conservative or restricted fluid management strategy is superior to a liberal fluid strategy in ARDS. Conservative fluid strategy and appropriate use of diuretics help to reduce extravascular lung water and to improve gas exchange. Protocolized fluid management strategies are strongly encouraged to limit problems associated with fluid overload in RF.

Prone Positioning

Prone positioning improves ventilation perfusion matching by recruitment of atelectatic posterior lung segments. While previous adult trials on prone positioning in ARDS showed improvements in oxygenation without any impact on survival, 1 recent adult study showed a 28- and 90-day survival advantage from early application of prolonged prone positioning in severe ARDS. The largest pediatric trial on prone positioning in acute lung injury did not reduce ventilator-free days or improve other relevant clinical outcomes. Prone positioning is, thus, not recommended for routine use in PARDS and needs further study.

Recruitment Maneuvers

Slow incremental and decremental PEEP steps can be used as a recruitment strategy in PARDS to improve oxygenation while closely monitoring hemodynamics. Sustained inflation maneuvers may contribute to barotrauma and should be avoided. Recruitment maneuvers (RM) in PARDS needs further study.

Airway Clearance Strategies

Chest physical therapy has not been shown to be of benefit in RF. Closed suctioning systems are preferred to open ones to limit ventilator disconnection and loss of lung volumes in PARDS. Maintaining patency of the endotracheal tube and avoiding airway plugging is essential in patients with thick secretions. Thus, mucolytics may be used to enable mucus clearance in specific patient populations, although definitive data are lacking. Vest therapy, cough assist, and intermittent percussive ventilation (IPV) can help prevent the need for mechanical respiratory support or may facilitate weaning toward extubation in patients with ineffective cough.

Monitoring

The PALICC investigators have made recommendations for respiratory monitoring in patients with PARDS. These recommendations can be applied to all patients in RF and some are listed in Table 104-2. Other parameters such as ratio of dead space to tidal volume, esophageal manometry, flow-volume loops, dynamic pressure-volume curves, oxygenation index or oxygen saturation (SaO₂) index, and PaO₂/FiO₂ or SaO₂/FiO₂ can be used based on expertise of clinicians in their interpretation of the obtained data.

SUMMARY

In summary, RF is a frequent cause of pediatric ICU admission, and the level of support needed can vary significantly among patients. Children in RF should be monitored closely with timely interventions to meet their respiratory demands. The mode of support should be titrated to optimize support while practicing lung-protective ventilation strategies and prevention of VILI. Adjunctive therapies should be used appropriately to limit time on mechanical ventilation, although supportive data are quite limited in pediatrics. Children failing conventional modes of support should be considered for nonconventional modes or initiated on ECMO support, if appropriate.

SUGGESTED READINGS

INTRODUCTION

Primary etiologies of acute brain injury in children include traumatic brain injury (TBI), hypoxic-ischemic encephalopathy (HIE), stroke, cerebral hemorrhages, infections, inflammatory conditions, seizures, tumor or mass lesions, metabolic abnormalities, and toxins (Table 105-1). TBI is the leading cause of death and disability in both children older than 1 year of age and young adults. Following the primary injury, the tissues surrounding the injury and the entire brain are vulnerable to further secondary injuries. The purpose of the initial management, investigation, monitoring, and treatment of acute brain injuries is aimed at preventing this secondary injury.

Pathophysiology of Secondary Injuries

Secondary injuries evolve in the minutes to days after the primary event and include both endogenous responses to the primary injury and secondary insults that occur in the field or during the course of care, such as hypoxemia or hypotension. Mechanisms involved in endogenous secondary injuries include energy failure, excitotoxicity and apoptosis, oxidative stress, mitochondrial dysfunction, inflammation, and multiple cell death pathways.

Hypoxia-Ischemia, Energy Failure, Excitotoxicity, and Oxidative Stress

In severe injury with cessation of blood flow, as occurs in ischemic stroke or cardiac arrest, a characteristic pattern of injury ensues. Oxygen stores are depleted very rapidly (< 20 seconds), and adenosine triphosphate (ATP) stores are depleted within 5 minutes. No further ATP can be generated to fuel energy-dependent cellular processes, and cell membranes depolarize, resulting in influx of sodium and calcium, efflux of potassium, cellular swelling, oxygen free radical production, and release of excitatory neurotransmitters such as glutamate from astrocytes and neurons. The release of glutamate triggers further depolarization of adjacent cells, and this potent combination of increased cellular metabolism and ischemia accelerates hypoxic injury. The same cellular dysfunction also occurs in situations of altered blood flow and is demonstrated in TBI, status epilepticus, and meningitis. In adult studies, cerebral blood flow (CBF) of less than 55 mL/100 g/min of brain tissue (at normothermia) led to inhibition of protein synthesis, key to the regeneration of injured tissues. CBF of less than 20 mL/100 g/min resulted in anoxic depolarization. These thresholds are likely higher in injured brain tissue, rendered vulnerable by excitotoxicity, seizures, and impaired oxygen utilization from mitochondrial dysfunction. Either globally or more focally, disturbance to the neurovascular unit that regulates CBF occurs in all forms of severe acute brain injury as a consequence of direct tissue disruption, edema, intracranial hypertension, vasospasm and loss of autoregulation. Much of the secondary injury in severe TBI and meningitis is hypoxic-ischemic in nature.

Necrosis and Programed Cell Death

Energy failure, excitotoxicity, and oxidative stress are the principle mechanisms leading to cell death, which can take the form of necrosis or apoptosis. The age of the child (ie, maturity), the severity and duration of the primary and secondary insults, and the vulnerability of the brain region contribute differently to determining the vulnerability or the resilience of the tissue. In circumstances of severe injury, the most vulnerable brain regions are the watershed areas of the intervascular boundary zones and areas with the highest metabolic rate, such as the thalami, basal ganglia, and sensorimotor cortex.

Cerebral edema in acute brain injury peaks at 24 to 72 hours after the initial insult and can result from three mechanisms: astrocyte and neuronal swelling, vasogenic edema, and osmolar swelling. Vasogenic edema results from inflammation and disruption of the blood–brain barrier (BBB) with endothelial dysfunction, such as in meningoencephalitis. Osmolar swelling occurs when intracellular macromolecules degrade, thus increasing intracellular osmolarity and drawing in of water.

Complications

The complications of acute brain injury can be focal and/or global. Focal neuroanatomical deficits are revealed on clinical examination with functional assessments (eg, Pediatric NIH Stroke Scale) or more globally with coma, with or without associated intracranial hypertension, seizures, or autonomic dysfunction. All of these factors can further exacerbate secondary injuries.

Intracranial Compliance

According to the Monro-Kellie doctrine, intracranial pressure (ICP) relates to the intracranial volumes of three components: brain parenchyma, blood, and cerebrospinal fluid (CSF). Volume inside the developed cranium is fixed, therefore an increase in volume of one of the compartments or from a mass lesion will lead to an increase in pressure unless there is a corresponding decrease in one of the other compartments’ volumes. Initially, the volume taken up by edema or a mass lesion such as an intracranial hemorrhage will be compensated for by an increase in CSF reabsorption with minimal increase in ICP. However, when the intracranial compliance is impaired as the lesion increases in size, and the capacity for further CSF displacement decreases, small increases in volume lead to significant increases in ICP. Intracranial hypertension will compromise CBF, which is proportional to the cerebral perfusion pressure (CPP; defined as the difference between the mean arterial pressure and the mean ICP or mean central venous pressure). This decrease in CBF impairs oxygen and substrate delivery, causing further ischemia and leading to a vicious cycle of secondary injury, edema, and intracranial hypertension. Ultimately, severe intracranial hypertension can lead to herniation of brain tissue through the foramen magnum, brain-stem compression, and death. Importantly, infants with unfused sutures, despite an increased capacity to accommodate changes in intracranial volume, are also at risk of increased ICP and herniation.

INITIAL EVALUATION AND MANAGEMENT

Acute brain injury can present with altered level of consciousness, seizures, or focal neurologic deficits. The purpose of the initial evaluation is to identify and treat the acute injuries, and to preserve viable brain tissue by preventing secondary insults. The initial evaluation of acute neurologic dysfunction is discussed in detail in Chapter 98. Here, we will focus mainly on priorities in patient management.

The immediate management priorities in patients with acute brain injury are to maintain a patent airway, deliver adequate oxygenation, maintain a normal blood pressure, immobilize the cervical spine in patients with suspected trauma, avoid hypercarbia, and complete a focused neurologic assessment and rapid patient survey. Additionally, all investigations that will help dictate appropriate management (eg, measuring the blood glucose level and treating hypoglycemia promptly; Fig. 105-1) should be expedited.

Stabilization of Airway, Breathing, and Circulation

Appropriate respiratory and cardiovascular support in the initial management of patients with severe head injuries is likely to profoundly impact their outcome. In a study of children with severe TBI from the National Pediatric Trauma Registry, those without hypotensive or hypoxic episodes fared best. The presence of hypotension tripled mortality, and the combination of both hypotension and hypoxia was associated with the worst outcomes. Indications for endotracheal intubation include upper airway obstruction or loss of upper airway reflexes, apnea, inadequate oxygenation or ventilation, coma with a Glasgow Coma Scale (GCS) score < 9, or signs of intracranial hypertension or impending herniation such as posturing or worsening coma scale score. Continuous cardiopulmonary monitoring should be commenced as early as possible. Rapid sequence intubation should be performed with manual axial in-line stabilization and neuroprotective measures. Nasotracheal intubation is relatively contraindicated in patients with suspected trauma until a base-of-skull fracture has been excluded on imaging. In intubated patients without signs of herniation, oxygenation and ventilation should target peripheral capillary oxygen saturation (SpO₂) ≥ 95% and normocapnia with end-tidal CO₂ monitoring used to titrate ventilation (partial pressure of CO₂ [PaCO₂] 35–45 mm Hg and pH 7.35–7.45). As a word of caution, rapid changes in PaCO₂ should be avoided unless there is an urgent need to treat intracranial hypertension. The reason for this is that CBF is greatly influenced by the pH of the CSF, which changes rapidly in response to sudden changes in PaCO₂. Sedation and analgesia should be given to prevent agitation and coughing, but agents particularly associated with hypotension should be avoided. In patients with signs of herniation, such as a fixed and dilated pupil or posturing, hyperventilation can be used as a bridge to prompt neuroimaging (computed tomography [CT]) and a decision regarding neurosurgical intervention. Hyperventilation is otherwise contraindicated because hypocapnea leads to cerebral vasospasm and a decrease in cerebral blood volume and CBF, increasing the risk of secondary ischemic injury, especially in the first 24 hours after injury when CBF is often at its lowest (Fig. 105-2).

Intravascular access should be obtained in patients with acute brain injury. Blood pressure is maintained with isotonic saline and vasoactive medications as needed. Glucose should be checked, and normoglycemia maintained. The patient should be exposed and log-rolled to avoid exacerbating any spinal injury. The temperature should be kept normal. While there is no evidence to support the routine use of deliberate hypothermia as a neuroprotective intervention, there is good evidence that hyperthermia is detrimental through increases in cerebral metabolic rate. Fever or temperature swings are relatively common in the presence of brain injury and control of body temperature using a cooling blanket targeting normothermia can be useful in the early management of brain-injured patients.

Initial Investigations

The purposes of initial investigations (see Fig. 105-1) are threefold:

1. Identify and correct metabolic disturbances.

2. Identify and treat intracranial and spinal lesions that require emergent neurosurgical or medical intervention, such as a mass lesion, unstable cervical spine fracture, or stroke.

3. Identify and manage complications of acute brain injury, such as seizures and intracranial hypertension.

Initial Neuroimaging

Indications for neuroimaging are persistent altered level of consciousness, new focal neurologic deficit, severe headache, vomiting, amnesia, signs of skull fracture or penetrating skull injury, and seizures with atypical or concerning features. Several decision-making rules for neuroimaging in TBI have been validated in the PECARN (Pediatric Emergency Care Applied Research Network), CATCH (Canadian Assessment of Tomography for Childhood Head injury), and CHALICE (Children’s Head Injury Algorithm for the Prediction of Important Clinical Events) guidelines. Imaging of the spine should be obtained in patients with suspected trauma and tenderness over the spine, focal sensory or motor deficits, altered level of consciousness, or distracting injuries.

Cranial CT can rapidly be performed in the unstable patient and is generally the initial neuroimaging modality of choice. CT is adequate for identification of intracranial lesions amenable to surgical intervention, such as hemorrhage, mass lesions, and obstructive hydrocephalus, and is ideal for the identification of skull and cervical spine fractures. The major drawback of CT is radiation exposure and the theoretical risk of radiation-associated cancers; however, dose-reduction scanning protocols have become standard for children. Caution should be taken when interpreting a CT obtained in the first 24 hours, as some abnormalities may not be visible (eg, strokes or demyelination or global ischemia).

Magnetic resonance imaging (MRI) has advantages over CT but has the drawbacks of long image acquisition times that require immobility and limited access to the patient during the scan. MRI is the preferred modality for examining the spinal cord, and the posterior fossa, due to artifact from the dense temporal petrous bones on CT. It also can better demonstrate demyelination, infection, and inflammation. Diffusion-weighted MR imaging can demonstrate stroke (from ischemia to infarction) within hours, much earlier and better than can CT. It provides finer anatomic detail than CT, which may be required, for instance, to guide resection of a tumor. Therefore, although CT alone is generally adequate for initial neuroimaging to exclude intracranial blood, fractures, and hydrocephalus, early MRI may be required to diagnose and expedite treatment of stroke or demyelination, or to delineate a tumor for surgical planning. Due to the increased risk of upper cervical spine injury in young children and the lack of sensitivity of CT for cord and ligamentous injury, spinal MRI is performed in patients in whom the spine cannot be cleared on clinical grounds, even if the CT or plain radiographs are normal.

Lumbar Puncture

A lumbar puncture to collect spinal fluid and to measure CSF pressure with manometry can assist in diagnosing infection, inflammatory and metabolic conditions, and subarachnoid hemorrhage, but should not be performed without excluding intracranial hypertension due to the risk of precipitating brain-stem herniation. A normal head CT does not exclude elevated ICP, and there are several documented cases of children with GCS 15 and raised ICP.

After initial stabilization, patients with severe brain injury should be transferred to a tertiary care facility with expertise in pediatric neurocritical care, neurosurgery, and neurology.

MONITORING

Patients with moderate to severe brain injury should be managed in the intensive care unit. Close observation with continuous cardiopulmonary monitoring and constant reassessment with neurological examinations (eg, GCS score or Pediatric NIH Stroke Scale score) are key to track a patient’s progress and monitor for complications. For critically ill children, multiple neuromonitoring modalities are available, with the greatest emphasis on ICP monitoring.

Intracranial Pressure Monitoring

The major aim of ICP monitoring is to prevent secondary hypoxic-ischemic injury from low CPP. CPP can be derived from the mean arterial blood pressure (ABP) and mean ICP and can be displayed with other continuous variables. All patients with severe TBI should be considered for ICP monitoring. Its use has not been shown to be beneficial in raised ICP from other forms of brain injury, such as asphyxia or infection. ICP is generally monitored either via a catheter placed in the lateral ventricle (an external ventricular drain) or an intraparenchymal microtransducer. The major advantages of an external ventricular drain are better accuracy and the ability to drain CSF to lower ICP. Microtransducers have lesser risks of infection and hemorrhage but cannot be recalibrated to 0 and tend to noticeably drift over the course of several days.

Pressure Reactivity and Pressure–Volume Compensatory Reserve

In health, the integrity of the neurovascular unit ensures that when arterial blood pressure (ABP) increases, pressure regulation leads to vasoconstriction in order to maintain a constant CBF, which in turn leads to a decrease in cerebral blood volume and a decrease in ICP (Fig. 105-3). Autoregulation normally takes place over a wide range of blood pressures but cerebrovascular reactivity can be deranged in the injured brain and can fail below a certain threshold CPP. With disturbed cerebrovascular reactivity or loss of cerebrovascular reserve, changes in ABP are directly transmitted to ICP.

Indices of cerebrovascular integrity can be derived from waveform and trend analysis of the ICP, ABP, and CPP. Examples include the pressure reactivity index (PRx), which captures the effect of slow changes in ABP on ICP when ICP is measured with a closed intracranial system (ie, no CSF drainage), and the pressure–volume compensation index (RAP), which is derived from small changes in intracranial volume throughout the cardiac cycle. Both the PRx and RAP indices have been shown to correlate with outcome in a limited number of adult and pediatric studies. Further studies are required to determine whether applying such measures in treatment algorithms can influence outcome.

Continuous Electroencephalogram

The use of continuous electroencephalogram (EEG) is increasing in the PICU. While the primary purpose is to detect and treat seizures, continuous EEG can also track the degree of encephalopathy over time, assist in prognostication, and monitor the achievement of burst suppression during the use of barbiturate therapy in severe TBI with refractory intracranial hypertension or refractory status epilepticus. Most commonly, a 20-lead EEG is used along with digital video recording, clinical event input from the bedside nurse, electromyography leads, and cardiopulmonary monitoring. Software detection algorithms are used to flag events suspicious for electrographic seizures but the output still requires review and interpretation by a neurologist or neurophysiologist, making this a resource-intensive form of monitoring. Recent studies of critically ill children report the prevalence of electrographic seizures to be around 30% in severe TBI, 35% in HIE secondary to cardiac arrest, 45% in CNS infection, and 60% in stroke. A shorter 1-hour EEG fails to identify half of children with seizures, one-third of which are exclusively subclinical. The yield increases to around 90% with 24 hours of monitoring.

The provision of a continuous EEG service is a resource-heavy undertaking and must be available on nights and weekends. While it is clear that continuous EEG adds information, more evidence is needed to determine whether treatment guided by this information will improve outcomes.

Transcranial Doppler

Transcranial Doppler (TCD) allows for the measurement of the velocity of blood in intracranial arteries, most commonly of the middle cerebral artery, and provides a repeatable surrogate for CBF. TCD-derived measures can also be used to estimate autoregulatory capacity or vasospasm.

Near-Infrared Spectroscopy

Near-infrared spectroscopy (NIRS) uses diode light sources to emit multiple wavelengths of near-infrared light (700–1000 nm) and measure tissue oxyhemoglobin and deoxyhemoglobin. Approximately 75% of the blood in the cerebral cortex is postcapillary, and NIRS provides a regional venous-weighted estimate of hemoglobin oxygen saturation. Cerebral NIRS has been increasingly applied in the monitoring of congenital heart surgery. While low cerebral NIRS saturations have been shown to predict adverse neurodevelopmental outcomes in some groups, prospective studies of algorithms incorporating NIRS in the intensive care unit are, in general, lacking.

Regional Brain Tissue Oxygenation (PbtO₂) or Microdialysis

For research purposes, catheters can be placed into brain parenchyma to measure regional oxygen tension, pH, and temperature. Microdialysis catheters are also available and can measure changes in glucose, lactate, pyruvate, neurotransmitters, and various amino acids. These catheters may be more sensitive to conditions of ischemia and their inclusion in management protocols has been associated with neurological outcomes, but pediatric experience is limited. A major drawback is that the measurements only reflect the metabolic conditions of a single brain region. Blood flow is known to be heterogeneous even in global injuries such as HIE after cardiac arrest, which begs the question of how best to use the information.

TREATMENT

The management of severe acute brain injury requires input from multiple teams, including critical care, neurosurgery, neurology, and allied health, remembering that such injury often means a very long hospital course and subsequent rehabilitation for some patients. The general principles are prevention of secondary insults and treatment of complications, such as raised ICP. Specific treatments related to status epilepticus, stroke, and central nervous system infection are covered elsewhere in this book. We will focus on the treatment of raised ICP and the use of therapeutic targeted temperature management and hypothermia therapy.

Management of Raised Intracranial Pressure

The second edition of Guidelines for the Acute Medical Management of Severe Traumatic Brain Injury in Infants, Children, and Adolescents synthesizes the best evidence for the management of increased ICP in pediatric patients. Although there was insufficient evidence to support a specific treatment algorithm, we present a sensible approach to the management of raised ICP in intubated patients with ICP monitoring, the specifics and order of which may vary between centers. Although most of the evidence relates to severe TBI, the same approach can reasonably be applied to other forms of brain injury with cerebral edema and intracranial hypertension.

Measured ICP should be maintained below 20 mm Hg. CPP should be maintained above 40 mm Hg in children 0 to 5 years, above 50 mm Hg in children 6 to 17 years, and above 50 to 60 mm Hg in adults.

First-Tier Therapies

Raised ICP in intubated patients should initially be treated with adequate sedation, analgesia, and neuromuscular blockade to prevent discomfort, agitation, coughing, and ventilator-patient dyssynchrony. The head of the bed should be elevated to 30 degrees with the head midline and no compression of the neck veins. Temperature should be kept normal. Peripheral capillary oxygen saturation ≥ 95% should be maintained, and blood gas targets include PaCO₂ 35 to 45 mm Hg and pH 7.35 to 7.45. Electrolytes should be monitored carefully, with special attention to sodium, which is often targeted at high-normal levels in the early phase of management.

In patients with a ventriculostomy catheter (external ventricular drain), CSF can be drained. This is often accomplished by setting the catheter to continuously drain at pressures above 15 mm Hg or by intermittent drainage to maintain ICP < 20 mm Hg.

If ICP remains elevated, hyperosmolar therapy, in the form of 3% saline 5 to 10 mL/kg or mannitol 0.25 to 1 g/kg can be given. Hypertonic saline is often preferred, as it avoids the osmotic diuresis and risk of hypotension induced by mannitol. However, sodium equilibrates across the BBB in a few hours and thus loses its effectiveness unless the serum sodium level is continuously increased. Furthermore, hypernatremia is associated with renal failure and ARDS in children treated for intracranial hypertension. Therefore, the use of hypertonic saline should be limited to the minimum amount required to maintain ICP < 20 mm Hg rather than used to target a certain serum sodium level.

Second-Tier Therapies

In patients with persistently raised ICP despite optimization of these first tier therapies, mechanical problems or seizures should be ruled out. Due to the prevalence of subclinical seizures in acute brain injury, as previously described, seizure prophylaxis with phenytoin or levetiracetam should be considered in patients with severe acute brain injury. A repeat cranial CT should be considered to assess for surgical lesions. The following second-tier therapies are effective at reducing ICP but may be associated with adverse effects and have not been shown to improve outcomes and may cause harm in some.

Decompressive craniectomy releases ICP by removal of part of the skull with or without the dura. Despite effectively reducing ICP, its effect on patient outcomes is less clear. One small pediatric study found that it may improve functional outcomes. However, a large adult trial demonstrated no survival benefit and concerningly found a greater risk of unfavorable neurological outcome in those who underwent decompressive craniectomy. Results from another large trial are forthcoming.

High-dose barbiturate therapy can be titrated to produce electrophysiological burst suppression, which reduces cerebral metabolism, thereby reducing CBF and ICP. Barbiturate therapy has not been shown to improve outcome and is associated with an increased risk of infection.

Therapeutic Hypothermia

Fever after any form of severe injury is associated with worse outcomes because of its impact on metabolism and cell death pathways. Mild hypothermia (32–34°C) does lower the ICP and has been used as a rescue therapy for raised ICP or as a first-line therapy for 24 to 72 hours to prevent secondary injury in severe TBI and HIE after cardiac arrest. Children are most often cooled using surface cooling. There have been multiple trials showing that the use of hypothermia therapy in severe TBI does not improve outcomes and may cause harm, including in pediatric patients. These trials have been criticized on the grounds that cooling was only achieved by 8 to 12 hours after injury, whereas animal models suggest that cooling within 4 to 6 hours is required for benefit. Cooled patients often need muscle relaxants to prevent shivering and should be observed for coagulopathy, electrolyte abnormalities, dysrhythmias, and altered drug metabolism. The ideal rate of rewarming is not known but should not exceed 0.5^°C per hour and may need to be as slow as 0.7^°C per 4 hours or may need to be titrated to systemic and cerebral hemodynamic endpoints. While therapeutic hypothermia has limited, if any, role after TBI in children, there is little doubt that fever is detrimental in this setting and that targeted normothermia should be part of all neuroprotective protocols for the first 48 to 72 hours after significant injury.

Mild hypothermia has been extensively studied in the management of HIE after cardiac arrest and has been shown to be beneficial in adults with ventricular tachycardia (VT) or ventricular fibrillation (VF) cardiac arrest. The 2 adult studies that showed benefit with therapeutic hypothermia both had ice-cold saline infused intravenously into the patients by paramedics in the ambulance, on the way to the hospital. However, targeted temperature normothermia has been recently shown to be equivalent to therapeutic hypothermia in adults after cardiac arrest. Cardiac arrest (mainly asystole compared to adults with VF) is different in children, with 80% to 90% being asphyxial in nature, and evidence for therapeutic hypothermia is lacking. A recent trial of therapeutic hypothermia in pediatric out-of-hospital cardiac arrest found that hypothermia did not confer any benefit in terms of survival when compared to controlled normothermia, and results from an in-hospital cardiac arrest cohort are forthcoming. By contrast, cooling in neonatal HIE is likely beneficial in terms of overall survival and functional outcome in survivors.

OUTCOME PREDICTION

Prognostication can be helpful to discriminate between children who will recover to enjoy a reasonable quality of life and those who will certainly die or be dependent on others for all aspects of daily life. Prognostication in the ICU remains challenging and can be confounded by therapies such as sedation and neuromuscular blockade. Decisions are further complicated by differences in what families perceive to be a good or favorable outcome. High specificity and positive predictive value of tests used in prognostication are key to avoid forming an impression that leads to withdrawal of life supportive therapy in a child who would otherwise have enjoyed a good quality of life.

Etiology and Circumstances

In children with return of spontaneous circulation (ROSC) after cardiac arrest, out-of-hospital location, younger age, longer duration of CPR, and asystole or pulseless electrical activity versus VT/VF as the initial rhythm are associated with poor outcome. In severe TBI, lower GCS score, particularly the motor score, after resuscitation; cerebral edema with unreactive pupils to light; and nonaccidental mechanism of injury portend a poorer prognosis.

Physical Examination

The predictive value of clinical signs is better outside of 72 hours following injury and should be interpreted or delayed depending upon the level of sedation, neuromuscular blockade, and hypothermia. Absent pupillary responses, absent corneal reflexes and absent or extensor motor response after 24 hours carries a high predictive value for poor outcome in HIE after cardiac arrest. Interpretation requires more caution in severe TBI because of possible isolated injury to reflex arcs.

Electrophysiology

Bilaterally absent somatosensory evoked potentials (SSEP) carry a greater than 90% specificity for poor outcome if performed outside of 24 hours from injury. Particularly in severe TBI, one must exclude structural abnormalities that can interfere with the test such as cord injury or subdural collection.

Burst suppression or a nonreactive EEG in the absence of sedation and metabolic abnormalities are strongly predictive of a poor outcome. Refractory status epilepticus on continuous EEG monitoring is associated with greatly increased odds of death or decline in functional status, but it lacks specificity.

Neuroimaging

Normal head CTs are not helpful in predicting longer term outcomes, especially when done in the first 24 hours. MRI 3 to 5 days after injury is a better predictor, with the most useful sequences being diffusion-weighted imaging (DWI), diffusion tensor imaging (DTI), susceptibility-weighted imaging (SWI), and MR spectroscopy.

Biomarkers

Several biomarkers have been studied in relation to outcomes after brain injury. These are ideally measured in the blood or urine, specific to brain tissue and elevated in proportion to the degree of brain damage. The most commonly studied biomarkers are neuron-specific enolase, s100b, and myelin basic protein. Some studies have reported impressive test characteristics but results between studies are inconsistent and no useful cutoff thresholds have emerged. At present, brain biomarkers cannot be recommended to guide clinical decisions, and should be considered in the context of research.

SUGGESTED READINGS

INTRODUCTION

Sepsis has been defined in adults and children in terms of changes in vital signs (heart rate, respiratory rate, temperature) and white cell count in the context of infection (Table 106-1). The systemic inflammatory response syndrome previously described the same response in the absence of confirmed infection (eg, after trauma or major surgery). Over the past decade, the Surviving Sepsis Campaign has made important efforts to develop consensus statements defining sepsis, sepsis syndrome, septic shock, and the SIRS in adults. The first set of guidelines were introduced in 2008 and were updated in 2012 and again in 2016 (Sepsis-3). These consensus statements included updated definitions for sepsis, sepsis syndrome, septic shock, and the systemic inflammatory response syndrome, as well as evidence-based guidelines and algorithms for the management of these conditions in adults. In this framework, sepsis is described as “a life-threatening organ dysfunction caused by a dysregulated host response to infection.” Although new consensus definitions have not yet been developed for the pediatric population, this description can be applied to patients of all ages.

Regardless of how we define it, the early recognition of sepsis, resuscitation from shock, and control of the causative agent all have dramatic effects on survival, and delay in any of these steps is associated with an exponential increase in mortality. This is because the vast majority of deaths from sepsis in children occur in the first hours after diagnosis (Fig. 106-1). Recognition of the fact that the prehospital and emergency room management of patients with sepsis can determine their prognosis has been key in defining resuscitation priorities and timing of interventions for all those involved in their care.

PATHOGENESIS AND EPIDEMIOLOGY

The pathogenesis of sepsis is not fully understood and there is no accepted gold standard to define the dysregulated response. However there exist 3 distinct (though sometimes overlapping) “phenotypes” of sepsis. The first phenotype is an overexuberant proinflammatory response, which at its extreme (eg, in the setting of meningococcal sepsis with purpura fulminans) includes widespread endothelial activation, disseminated intravascular coagulopathy, myocardial depression, and microvascular dysfunction. The second clinical manifestation is with multiple organ-system failure. Finally, for a subgroup of patients with sepsis that fail to respond adequately to an invading microorganism, a state called “acquired immunoparalysis” or “monocyte deactivation” can be present. Increasingly there is a view that some of the more extreme phenotypes (eg, multiple organ failure with predominant hepatobiliary dysfunction and coagulopathy) may be considered as a form of hemophagocytic lymphohistiocytosis (HLH). This is an area of active investigation, and will undoubtedly be better characterized over coming years. At the time of writing, for most intensivists, care for sepsis includes treating the cause and providing organ support while awaiting recovery, rather than interventions specific to the phenotype.

The key pathological result of the inflammatory response that is relevant to early resuscitation is the net myocardial depressant effect. Importantly, however, the effect to which this myocardial depression can be compensated by other responses (eg, tachycardia, ventricular dilatation) and volume status varies widely between cases and results in a wide variety of hemodynamic presentations of septic shock that have only recently been defined in children (Fig. 106-2).

Sepsis rates are increasing in developed countries, perhaps in part reflecting a growing population of children living with chronic illness who are at greater risk, or reflecting more aggressive therapeutic immunosuppression for malignancy. A recent study estimated that around 1 in 20 pediatric intensive care unit (PICU) admissions are for sepsis or septic shock, though many more follow other serious infections. Newborns and infants have the highest rates of sepsis and septic shock and the worst outcomes. Serious comorbidities (including prematurity and congenital malformations), a history of complex surgical interventions, relatively immature immune systems, and greater difficulties in diagnosis all contribute to the vulnerability of this age group.

PRESENTATION AND DIAGNOSIS

A single definition, tool, or biomarker to identify whether an individual patient has sepsis, or is at risk of developing septic shock, does not currently exist. However, pediatric resuscitation courses (Pediatric Advanced Life Support [PALS] and Advanced Pediatric Life Support [APLS]) teach clinicians how to recognize a seriously ill child, and emphasize the importance of simultaneous assessment and resuscitation: “treating things that kill first.” Additionally, more customized sepsis bundle guidance is now recommended, with a number of care steps to be delivered to achieve compliance (eg, the UK Sepsis Six bundle). If initial resuscitation does not reverse sepsis-associated shock, then more specific guidance, continuing into the ICU admission, advocates not only the reversal of the clinical features of shock, but also the importance of monitoring hemodynamic parameters and markers of oxygen delivery such as cardiac index, systemic vascular resistance index, and central venous oxygen saturation, taken from the superior vena cava. All such guidelines advocate steps within a specific timeframe and further action if initial steps are not effective.

Algorithms that identify high-risk features and mandate reassessment in case of uncertainty can be very effective. A recent review document by the National Institute for Clinical Excellence in the United Kingdom produced valuable algorithms for children and adults in various settings. The pediatric algorithm provides a pragmatic approach to the categorization of the risk (high, moderate, or low) of sepsis according to objective clinical features that include heart rate, respiratory rate, and level of consciousness.

The other key point to note is that because sepsis is a dynamic, exponential process, it is only rarely possible to exclude early sepsis, and the hallmark of sepsis is a rapid deterioration in the absence of appropriate treatment. Hence, a clinical assessment that concludes that sepsis is not currently present should include safety-net advice about what to do if things change. In the face of uncertainly, a planned reassessment is recommended.

CLINICAL MANIFESTATIONS

There is no single symptom or sign that distinguishes sepsis from other causes of critical illness, nor is there a gold standard definition for sepsis. The threshold for a suspicion of sepsis must therefore consider the environment in which the assessment is taking place. In other words, a lower threshold would be appropriate for a bone marrow transplant recipient or others with immune dysfunction, or for younger children.

For the first responders, the overwhelming clinical priorities for a septic patient that has been resuscitated are continued observation for cardiorespiratory system failure (remembering that hypotension is a late sign), new organ system dysfunction, and source control (Fig. 106-3). Clinical practice surveys have consistently shown that the clinical care of children with sepsis is neither as prompt nor as reliable as staff or lay people would presume to be the case. This is of particular importance since the adherence to algorithms for assessment and intervention such as Surviving Sepsis or the Sepsis Six (Fig. 106-4), have been shown to improve both processes of care and outcomes. Most clinicians would plan to transfer a child with sepsis to the ICU prior to, or at the time when, fluid refractory shock is suspected, and certainly when vasoactive therapy is being considered.

EARLY INTERVENTIONS FOR SUSPECTED SEPSIS

Antibiotics

Without prompt and vigilant attention to source control, the outcomes of sepsis will inevitably be compromised. The recommendation from most sepsis care bundles is that appropriate antibiotics should be given within 60 minutes. The choice of antibiotics should be determined according to local policies with knowledge of the relevant pathogens and likely antibiotic resistances or sensitivities, as well as additional factors such as severe immunosuppression, chronic disease, or a semipermanent catheter for long-term parenteral nutrition.

Fluid Resuscitation

The Starling relationship of increased stroke volume with increased end-diastolic filling is the basis for rapid intravenous volume expansion in septic shock. Current recommendations are for rapid administration of 20 mL/kg of crystalloid or colloid in the absence of signs of fluid overload (eg, hepatomegaly, lung crackles), repeated as necessary. While high-level evidence for this intervention is weak and subject to extensive debate, in an environment with ready access to positive pressure ventilation, early aggressive fluid resuscitation has been associated with substantially improved outcomes. We recommend this approach but wish to highlight that this is an observed trial of a fluid bolus, which might be curtailed if signs of fluid overload (eg, breathlessness, crepitations on chest auscultation, or increased liver size) occur, or repeated if it leads to clinical improvement but not reversal of shock. This is a dynamic situation needing the clinician to remain by the bedside to assess the physiological response to therapy.

The above principles of rapid fluid expansion for children with septic shock may not apply in all settings, particularly in developing nations. The Fluid Expansion as Supportive Therapy (FEAST) study, a large randomized study of early fluid resuscitation in children with sepsis presenting in a resource-poor area, demonstrated that not all children with sepsis respond similarly to early fluid administration. The study found that rapid and generous fluid resuscitation according to standard algorithms was associated with increased harm, and this led the World Health Organization (WHO) to adapt a different standard in resource-poor areas. The WHO has produced guidelines known as Emergency Triage Assessment and Treatment (ETAT) that specifically considers resource-limited environments and approaches to fluid resuscitation in children with malnutrition. The main differences for malnourished children being the cautious and more limited early fluid intravenous bolus administration, early blood transfusion in patients with severe anemia, and early introduction of nasogastric fluids rather than continued intravenous fluid therapy.

Vasoactive Drug Infusions

If a child fails to respond to initial fluid and early inotrope resuscitation, then he or she is considered to be in a state of fluid-resistant septic shock. The early use of inotropes is recommended in patients with septic shock that do not respond promptly to fluid according to published consensus guidelines. Shock in children has been shown to have different manifestations compared to adults in whom vasoplegic high output or warm shock is usual. Children initially may present with this adult pattern—which some researchers have found more prevalent in hospital-acquired sepsis—but many present with cold shock, in which low cardiac output and high systemic vascular resistance predominate, often in community-acquired sepsis. Specific vasoactive medication to normalize this hemodynamic pattern is recommended: epinephrine for cold shock and norepinephrine for warm shock. However, this pattern changes over the period of resuscitation and stabilization. A child presenting in cold shock may transform to warm shock after a few hours, so frequent measurements of hemodynamic parameters and tailoring of vasoactive medication to achieve the hemodynamic parameter guideline goals is required. If this does not work, more aggressive treatment for catecholamine-resistance shock is mandated. Depending on shock morphology and response, aggressive treatment might include further inotrope titration, addition of inodilators, or extracorporeal membrane oxygenation (ECMO) support.

Early Goal-Directed Therapy

Early goal-directed therapy refers to the resuscitation of patients according to specific hemodynamic targets beyond simple vital signs; examples being central venous oxygen saturation or the clearance of serum lactate. There is limited evidence to support this approach in populations with high predicted case fatality rate (≥ 40%). However, with the increasing adoption of aggressive early resuscitation bundles, the utility of this approach has been further questioned.

Mechanical Ventilation

Early consideration of mechanical ventilation is recommended in septic shock. However, great care must be taken not to induce further myocardial depression with induction agents. Inhalational agents, propofol, barbiturates, and high dose benzodiazepines should be avoided in this setting. In addition, etomidate should not be used, given the increased risk of adrenal suppression that has been associated with its administration in patients with sepsis. Once ventilation is established, recognition of the risk of acute lung injury or acute respiratory distress syndrome and adoption of a lung-protective strategy is recommended.

Extracorporeal Membrane Oxygenation

ECMO has been used successfully in refractory shock and severe acute respiratory distress syndrome in pediatric sepsis, with an overall survival to discharge of at least 50%. While there are no randomized trials of ECMO for septic shock, the typical physiology of shock, the often very elevated (supranormal) oxygen demands, and the need for high flows have led to the consideration of central cannulation (direct cannulation of the aorta and right atrium) as the approach of choice. While this approach does raise additional concerns for bleeding in patients who may already be coagulopathic, it provides the advantage of almost unlimited flows to meet the elevated metabolic demands of the hypoperfused tissues. When used early, ECMO runs are typically short (< 5 days) without long-term sequelae.

PREVENTION

The value of immunization in preventing community-acquired infection is undisputed and is probably medicine’s greatest influence on global health. Infection control measures in the hospital setting (eg, hand-washing, central venous line care, ventilator-associated pneumonia prevention, selective digestive decontamination) have been shown to have important benefits. The use of specific prophylaxis in immunosuppressed children also reduces the incidence of severe opportunistic infections.

SUMMARY

Sepsis is a clinical syndrome that is vitally important to suspect, recognize, and treat early (even if the diagnosis is initially uncertain), with continued escalation if signs of shock persist. The application of sepsis bundles to guide clinical care is recommended, with the proviso that pediatric presentations, particularly in younger children, can differ from those of adults. However, early intervention and resuscitation, and source control remain the mainstay of the early management of sepsis and the most important determinants of outcome.

SUGGESTED READINGS

INTRODUCTION

The clinical syndrome of shock has great potential for significant morbidity and mortality and is one of the most challenging conditions to treat. In accordance with Barcroft’s original construct for conceptualizing shock that was published nearly 100 years ago, shock results from 1 or more of the following mechanisms: impaired oxygenation or hypoxic hypoxia, reduced oxygen carrying capacity or anemic hypoxia, limited cardiac output or stagnant hypoxia, or impaired oxygen utilization or histotoxic hypoxia (eg, cyanide toxicity). Shock results from inadequate oxygen delivery relative to oxygen demand, and if the body’s intrinsic compensatory mechanisms of increased cardiac output and oxygen extraction are insufficient, or left inadequately treated, this can rapidly result in organ dysfunction or failure. The etiologies of shock are broad, as are its manifestations, and in this review we will discuss the pathophysiology, assessment, and treatment of cardiogenic shock.

PATHOPHYSIOLOGY

Shock results from inadequate oxygen delivery relative to oxygen demand. Shock is not necessarily a problem of blood volume, cardiac output, or blood pressure, but it is always a problem of inadequate tissue oxygen delivery:

where DO₂ is oxygen delivery, CaO₂ is arterial oxygen content, and CO is cardiac output.

Cardiogenic shock is due to low cardiac output, the causes of which can be classified according to the determinants of cardiac output: inadequate preload, excessive afterload, intrinsic muscle failure, or dysrhythmia.

ETIOLOGY

Cardiogenic shock can result from a number of etiologies. Although not an exhaustive list, the most common categories and causes are listed below.

• Undiagnosed (or untreated) critical congenital heart disease: left heart obstructive lesions (critical coarctation, hypoplastic left heart syndrome, aortic stenosis); right heart obstructive lesions (critical pulmonary stenosis or pulmonary atresia); transposition of the great arteries

• Postoperative congenital heart disease: early low output syndrome; residual cardiac defects; late postoperative decompensation due to right, left, or bi-ventricular systolic or diastolic dysfunction; or pulmonary hypertension

• Primary myocardial disease: dilated, restrictive, hypertrophic cardiomyopathies

• Acute myocarditis

• Pericardial disease: acute tamponade or chronic pericardial constriction

• Arrhythmias: chronic untreated or undertreated arrhythmia resulting in cardiomyopathy, also pacemaker-induced cardiomyopathy can result from long term cardiac pacing

• Acute coronary syndromes

• Other etiologies of shock can manifest as cardiogenic shock

CLINICAL RECOGNITION

The timely recognition of cardiogenic shock requires a high index of suspicion, an appreciation for high-risk groups (eg, patients with known underlying cardiac disease) and a comprehensive consideration of all available information, including the medical history, physical examination, laboratory tests, and hemodynamic parameters.

HISTORY AND PHYSICAL EXAMINATION

Early markers of cardiogenic shock, from both the history and examination, may be subtle and easily missed or confused with other noncardiac causes of acute illness or shock leading to inappropriate interventions that may worsen the child’s clinical condition. The most common misdiagnoses are hypovolemia as the cause for tachycardia (resulting in fluid administration that can lead to further decompensation), or asthma, pneumonia, or a combination leading to the common findings of respiratory distress, wheeze, rales, and clinical signs of lobar collapse.

Cardiogenic shock most typically presents with a history of lethargy, nausea, vomiting, and anorexia in a child, and decreased intake and tiring or sweatiness with feeds in an infant. Caregivers commonly report a history of cough and wheeze and other signs of respiratory distress (eg, tachypnea, labored breathing, grunting respirations, and intercostal and subcostal retractions). There may be a history of weight loss, though fluid retention may in extreme cases falsely counter this observation. A history of wheeze, also resulting from pulmonary edema, is very common in children with cardiogenic shock, and this finding may misdirect the unsuspecting clinician to an algorithm that involves the management of primary respiratory rather than cardiac disease. Examination may reveal a sternotomy or thoracotomy from previous cardiac surgery. On auscultation, there may be also be widespread crackles resulting from cardiogenic pulmonary edema, or reduced air entry to the left lung resulting from collapse of the lower lobe due to cardiomegaly. Tachycardia and a gallop may also be present, and there may be absent or diminished lower limb pulses raising the possibility of a coarctation of the aorta. Common abdominal findings include hepatomegaly and abdominal tenderness resulting from hepatic congestion. Decreased mentation, irritability, or altered mental status are late and ominous findings in patients with cardiogenic shock, suggestive of cerebral hypoperfusion. The perfusion exam may be abnormal and consist of decreased peripheral pulses, prolonged capillary refill and cool extremities. While the perfusion exam is an integral part of the assessment of any patient, it does have significant limitations, including a lack of correlation with measured hemodynamic indices such as cardiac output and systemic vascular resistance. Nonetheless, with this limitation in mind, it is essential to establish a baseline assessment of perfusion and to perform serial examinations in order to assess the response to interventions and for determining the clinical trajectory.

HEMODYNAMIC MONITORING

Tachycardia is very common in patients with cardiogenic shock, and while this is most often a compensatory sinus tachycardia, it may also be an arrhythmia (supraventricular tachycardia, atrial flutter, or ectopic atrial tachycardia being most common, though ventricular tachycardia should also be excluded). Electrocardiogram (ECG) monitoring including a 12-lead should be routinely performed as soon as these are available. Patients may present with hypotension, so-called decompensated shock; however, it is not uncommon for the patient with cardiogenic shock to present with a normal if not elevated blood pressure due to a compensatory increase in systemic vascular resistance. Venous access should be established early in patients with cardiogenic shock, and when possible, central venous access should be obtained to assist with resuscitation as well as measurement of the central venous pressure (CVP) and central or mixed venous oxygen saturations (SVO₂). The CVP measures the right ventricular filling pressure and gives some indication of intravascular volume status and cardiac function. The CVP is often elevated in patients presenting with cardiogenic shock. The SVO₂ (normally 75% or greater) is an indirect measure of adequacy of cardiac output and status of tissue oxygen extraction, and a lower SvO₂ indicates higher oxygen extraction and therefore more precarious circulatory status.

LABORATORY EVALUATION

All critically ill patients should have a comprehensive chemistry panel. Patients presenting with cardiogenic shock and acute kidney injury may initially have a normal serum creatinine, as changes in serum creatinine lag behind the acute insult and actual decrement in glomerular filtration rate. Indicators of hepatocellular injury, such as elevated transaminases, while not specific for hepatic injury may be elevated in a shock state. Similarly, in shock, the endothelium is activated, resulting in a prothrombotic phenotype and a consumptive coagulopathy. Hypocalcemia may be the cause of or contribute to the development of myocardial dysfunction. Other electrolyte abnormalities that may contribute to cardiac dysfunction include disturbances in potassium, magnesium, and phosphorous homeostasis. An arterial blood gas is indicated in the management of critically ill patients, allowing for the assessment of gas exchange, acid–base balance and serum lactate. An elevated anion gap may be an indication of a lactic acidosis and shock state, whereas a nonanion gap acidosis is generally due to a loss of bicarbonate or excessive chloride administration (due to diarrhea or renal tubular acidosis, or following volume expansion with normal saline). The anion gap is calculated as shown below:

It may seem evident, but it cannot be overemphasized that a patient with a normal lactate level may be in a state of impending shock. The use of venous oximetry and near infrared spectroscopy (NIRS) to assess the extent of oxygen extraction can help in providing a more timely assessment of impending and existing shock than a lactate level alone.

In the intubated patient, capnography (end-tidal carbon dioxide [CO₂] monitoring) should be used. In the absence of a significant leak, a normal arterial-to-end-tidal CO₂ gradient is consistent with hypoventilation (ie, gradient < 3–5 mm Hg), and an elevated gradient is due to wasted (or dead space) ventilation. A large gradient, indicative of a very low or zero end-tidal CO₂, is seen in patients with limited pulmonary perfusion or in cardiac arrest, and end-tidal CO₂ monitoring may be used to assess the effectiveness of chest compressions during cardiopulmonary resuscitation and the return of spontaneous circulation, which are indicated by an increase in end-tidal CO₂ and a narrowing arterial-to-end-tidal difference.

Serum troponin levels may provide some indication of the cause and severity of myocardial injury, and serial levels will indicate time course of an underlying insult, particularly if it is ischemic. Certain subtypes of troponin (I and T) are very sensitive and are specific biomarkers of the cardiomyocyte, and may be elevated due to hypoxic–ischemic injury, myocarditis/inflammatory processes, and trauma, and following cardiac surgery. Elevations in troponin are also seen in severe sepsis and pericarditis. Finally, the B-type natriuretic peptide (BNP), which is indicative of cardiac distension, is a useful biomarker in the presence of cardiogenic shock. While this cannot distinguish among the multitudes of underlying pathophysiologies, it can provide a surrogate for the progression of the clinical course.

Ancillary Investigations

Chest Radiograph

A chest radiograph typically shows cardiomegaly and may also provide additional information, such as congenital heart disease, that may contribute to the diagnosis. Pulmonary edema may be present and is typically limited to the interstitium, and thus oxygenation is minimally if at all impacted. In more severe cases, alveolar edema is present, in which case oxygenation is impaired. Importantly, the chest radiograph will help to distinguish primary respiratory from cardiac disease. It is important to appreciate that, in longstanding heart failure, chronic pulmonary venous hypertension leads to functional and structural adaptive changes within the lung, and interstitial edema may be absent or minimal despite the presence of significant heart failure. The accumulation of interstitial edema in the bronchovascular sheath may compress small airways, leading to airway obstruction, hyperinflation, and flattened diaphragms.

Echocardiography

An echocardiogram should be obtained in any child presenting with cardiogenic shock, or in a situation in which it may be in the differential diagnosis. This will be key to distinguish between systolic and diastolic disease, assess for pericardial fluid and tamponade, and in the diagnosis of any structural heart disease or residual defects in children with operated congenital heart disease. Additionally, echocardiography can be used to assess the degree of dysfunction and also may help assess response to therapy. It is important to note, however, that symptomatic improvement typically predates objective echocardiographic improvement, and indeed the latter may never normalize despite the child returning to baseline after appropriate interventions.

PRINCIPLES OF MEDICAL MANAGEMENT: OPTIMIZING THE RELATIONSHIP BETWEEN OXYGEN DELIVERY AND CONSUMPTION

For patients at risk for or in a state of shock, all strategies that optimize the relationship between oxygen delivery and demand should be considered. These include adequate hemoglobin (to maximize the oxygen carrying capacity and delivery) as well as other measures that target intravascular volume, contractility, and loading conditions to optimize cardiac output. Strategies that minimize oxygen consumption should also be considered and include mechanical ventilation, elimination of anxiety with sedation, and the avoidance of hyperthermia, as elevated body temperature increases the metabolic rate and therefore global oxygen demand.

The Role of Mechanical Ventilation

While under normal conditions, the respiratory pump consumes less than 3% of total body oxygen consumption and receives less than 5% of total CO, in patients with cardiogenic shock, the diaphragmatic oxygen demands alone may increase to 50% of total cardiac output. Thus, one of the basic tenets of managing a critically ill patient in an impending or existing shock state is to mechanically unload the respiratory pump, allowing for a redistribution of a limited CO to other vital organs, including the brain. This may be accomplished with noninvasive continuous or bi-level positive airway pressure (CPAP or BiPAP, respectively) or invasively with endotracheal positive pressure ventilation. An additional benefit of positive airway pressure and the resulting increase in intrathoracic pressure is that it “mechanically” unloads the left ventricle by creating a pressure gradient between intra- and extrathoracic arterial vessels. Furthermore, mechanical ventilation eliminates labored breathing, and in doing so leads to a partial withdrawal of adrenergic activity and the associated increase in systemic vascular resistance, contributing to a reduction in left ventricular afterload, which is particularly beneficial to patients with left ventricular systolic failure.

Determinants of Preload and Afterload and Their Pharmacological Manipulation

The optimal management of the patient in impending or existing cardiogenic shock requires a thorough understanding of circulatory physiology and determinants of cardiac output, the pathophysiology of cardiogenic shock, an assessment and determination of each patient’s hemodynamic profile, and the careful selection of vasoactive agents that are best suited for each patient.

Systemic venous return, which is a determinant of right heart filling or preload, results from the pressure gradient between the systemic venous reservoirs (the mean systemic pressure) and the right atrium. Increased intravascular volume increases the mean systemic venous pressure, systemic venous return, right atrial pressure, right ventricular diastolic pressure (ie, filling pressure), and ultimately right ventricular diastolic volume (ie, preload). Venoconstrictors increase systemic venous return by decreasing venous capacitance and increasing the mean systemic pressure, whereas venodilators decrease the effective intravascular volume by increasing venous capacitance and sequestering intravascular volume within the venous reservoirs. A diuretic such as furosemide decreases the mean systemic pressure acutely by inducing venodilation and over time by decreasing intravascular volume.

Impaired systolic ventricular function leads to increased sensitivity to changes in afterload. Ventricular afterload may be reduced pharmacologically by arterial vasodilators, or mechanically (eg, positive pressure ventilation). Arterial vasodilators decrease vascular impedance, increasing stroke volume and cardiac output and reducing end-systolic volume but have no appreciable effect on preload or ventricular filling pressures. Inotropes increase stroke volume and cardiac output by augmenting contractility without changing filling pressures, and almost all inotropes produce a dose-dependent increase the heart rate that may actually contribute to an increase in cardiac output. However, tachycardia can also increase myocardial oxygen consumption and may compromise coronary filling, as well as increasing the risk of tachyarrhythmias, all of which can further exacerbate the depressed stroke volume and cardiac output.

Vasodilators, inotropes, inodilators, and diuretics are in an important minority. Prostaglandins are the mainstay of pharmacological management of cardiogenic shock. While deciding the optimal approach to vasoactive therapy, it is essential to have a clear understanding of the underlying etiology and pathophysiology, and of the desirable and unwanted effects of potential agents. The vasoactive management of a ductal-dependent lesion presenting with cardiogenic shock would be completely different to that of a child presenting with acute myocarditis. Another example would be the almost contradictory management approaches for acute or end-stage systolic vs diastolic heart failure. The mechanisms of action of specific pharmacological agents are discussed in detail in Chapter 111.

Intravascular Volume

In patients with cardiac dysfunction/failure, an understanding of the nature of the dysfunction (ie, systolic vs diastolic) as well as a determination of the optimal ventricular filling pressure are indicated because ventricular compliance (the relationship between pressure and volume for a distensible structure) is highly variable and what generates stroke volume is not the diastolic pressure per se but rather the extent to which the myocardium is stretched (ie, ventricular end diastolic volume). Further confounding the interpretation of the CVP is the fact that in the presence of cardiac and or pulmonary disease there may be only limited correlation between the CVP/right ventricular diastolic pressure and left ventricular diastolic pressure. By using a CVP, or in some postoperative patients a left atrial pressure, the optimal ventricular filling pressure can be determined by administering volume and monitoring the hemodynamic response. A favorable response is objectively reflected in a prompt rise in arterial blood pressure and fall in heart rate (ie, preload reserve is present). With a favorable response. it can be inferred that the ventricle is positioned on the ascending portion of its pressure–stroke volume curve (Fig. 107-1). A lack of response is consistent with the ventricle residing on the flat portion, or even on a potential downturn of its pressure–stroke volume curve (ie, preload reserve has been exhausted), in which case inotropy and/or afterload-reducing therapies are indicated to improve cardiac output. Diuretics and/or venodilators may be prudent in order to decrease systemic venous return and ventricular filling pressures (ie, preload), ameliorating pulmonary and systemic venous congestion.

Unlike in patients with systolic dysfunction, volume expansion may be indicated in patients with ventricular diastolic dysfunction/decreased ventricular compliance where an elevated ventricular filling pressure is needed in order to adequately generate sufficient diastolic filling to produce an adequate stroke volume. Significant diastolic dysfunction/decreased ventricular compliance is seen in patients with ventricular hypertrophy (eg, resulting from an obstructive lesion, such as semilunar valve stenosis or systemic hypertension, and hypertrophic cardiomyopathy) and restrictive cardiomyopathy, and in some patients after right heart surgery or following the Fontan procedure. The extent to which a pericardial effusion impacts ventricular compliance and filling depends on the amount and rate of fluid accumulation. With time, the compliance of the pericardium increases, allowing for a greater amount of fluid accumulation without necessarily a commensurate increase in pressure. In any case, once stroke volume and CO begin to wane, the situation becomes urgent, as the ventricle is now operating on the steep portion of its pressure–volume (compliance) curve. Volume is indicated as a temporizing measure until the fluid can be removed.

Mechanical Circulatory Support

In patients with cardiogenic shock in whom conventional medical management has failed, those who are rapidly decompensating, and those in whom definitive surgical management is not part of the early treatment algorithm, mechanical cardiac support—either extracorporeal membrane oxygenation (ECMO) or a ventricular assist device (VAD)—should be considered (see Chapter 108). Both ECMO and VAD provide temporary circulatory support, ensuring adequate systemic oxygen delivery and end-organ perfusion while minimizing myocardial oxygen consumption. Both ECMO and VAD support allow the myocardium a period of rest and recovery, and in some cases provide a window of stability for further diagnostic assessment.

Special Considerations for Infants

When managing cardiogenic shock, it is important to remember that infants have less cardiopulmonary reserve than do children and adults, which is an important consideration that cannot be overemphasized. The cardiomyocyte of the newborn/infant possesses relatively few, poorly organized contractile proteins, and the immature sarcoplasmic reticulum leads to inefficient calcium handling. As a result, the myocardium is less responsive to fluid and inotropes, and is more susceptible to increases in afterload. The mechanics of the lungs, chest wall, and diaphragm are also immature, with more chest wall compliance and reduced diaphragmatic contractile reserve, resulting in higher oxygen consumption by the respiratory unit in health. This is significantly exaggerated in the presence of cardiogenic shock which is characterized by tachypnea and increased respiratory effort.

TARGETED MANAGEMENT OF CARDIOGENIC SHOCK

Systolic Heart Failure

The optimal first-line strategy for treating systolic heart failure is to alter ventricular loading conditions, resulting in the desired hemodynamic effects of an increase in stroke volume and cardiac output without any unwanted effects on myocardial oxygen delivery and demand. This is typically achieved with a combination of vasodilators or inodilators and diuretics, and on occasions, the careful use of inotropes, in order both to reduce load and to improve contractility. When selecting the right agent or agents, it is important to focus on the desirable and unwanted effects that are often dose dependent, as well as the onset and duration of action and half-life of agents. A note of caution should also be introduced for drugs with a longer half-life, such as milrinone, which is commonly used to support patients in established end-stage heart failure. Milrinone is an ideal drug in terms of its hemodynamic effects that combine vasodilation with improved diastolic performance and some improvement in contractility, but its vasodilator effects may manifest with hypotension in an already marginal patient with acute cardiogenic shock. Unlike many other vasoactive agents, milrinone has a relatively long half-life, further prolonging its effects in the presence of renal dysfunction, which often coexists in patients with cardiogenic shock. As an alternative, in the acute phase where hypotension is present or may be manifested with vasodilation, epinephrine in low doses provides some afterload reduction while providing significantly greater inotropic support than dobutamine, dopamine, or milrinone. It may be prudent to initiate therapy with epinephrine in order to establish the desired clinical trajectory with the subsequent addition of or transition to vasodilators or inodilators.

Diastolic Heart Failure

Diastolic heart failure is characterized by a low stroke volume and cardiac output that is due to inadequate ventricular filling despite an elevated ventricular end diastolic pressure. As discussed, determining the optimal filling pressure (adequate preload) is essential. In these patients, diuretics and vasodilators will reduce the systemic venous return and decrease pulmonary venous pressure and the formation of pulmonary edema, but may compromise stroke volume and cardiac output by reducing filling. In the absence of systolic dysfunction, inotropes are generally deleterious, as they typically cause an increase in heart rate, which may compromise ventricular filling while increasing myocardial oxygen consumption and the propensity for developing tachyarrhythmias. In general, the primary approach in this setting is careful support of intravascular volume and control of heart rate.

Postoperative Cardiac Dysfunction

There are several intraoperative factors that may contribute to postoperative cardiovascular dysfunction after cardiopulmonary bypass. The systemic inflammatory response to cardiopulmonary bypass and myocardial ischemia reperfusion injury can lead to elevated systemic vascular resistance and left ventricular afterload as well as myocardial systolic and diastolic dysfunction. Furthermore, pulmonary ischemia reperfusion injury leads to pulmonary endothelial dysfunction and heightened pulmonary vascular reactivity and elevated right ventricular afterload. Similarly, neurohormonal activation is often present postoperatively, contributing to increases in left ventricular afterload, which may compromise cardiac output and systemic oxygen delivery if left ventricular systolic function is impaired.

Congenital Heart Disease

Cardiogenic shock may be the presenting feature of several types of unoperated congenital heart diseases, regardless of ventricular function. Important examples that merit discussion are ductal-dependent lesions with obstruction to systemic or pulmonary blood flow. In any critically ill–appearing newborn with or without cyanosis, a ductal-dependent lesion should be considered and prostaglandin E₁ (PGE₁) should be initiated without delay if it is in the differential diagnosis. Maintaining patency of the ductus is essential for systemic perfusion in newborns with left-sided obstructive lesions, or for maintaining pulmonary perfusion in newborns with right-sided obstructive lesions. Prostaglandin E₁ may also be indicated in transposition of the great arteries to provide adequate intercirculatory mixing. Important dose-dependent side effects include apnea, irritability, fever, and hypotension. Additional vasoactive agents may be indicated to provide further afterload reduction or inotropic support during the resuscitative and perioperative periods.

In newborns with single ventricle physiology, the systemic and pulmonary circulations are in parallel, with the respective vascular resistances determining the distribution of cardiac output to the systemic and pulmonary circulations. These infants are at significant risk of cardiogenic shock due to systemic hypoperfusion (and often, pulmonary overcirculation). Optimization of systemic perfusion in these precarious patients can be achieved by avoiding factors that excessively dilate the pulmonary vasculature (eg, hyperoxia, alkalosis, and hypocarbia) as well as careful control of afterload on the systemic ventricle. Achieving this critical balance requires meticulous attention to ventilation, with very careful use of systemic vasodilators, inodilators, or very low doses of epinephrine.

Critical Aortic Stenosis

Critical aortic stenosis is characterized by afterload-induced heart failure, with neurohormonal activation, and typically presents in the newborn period as cardiogenic shock with pallor, globally diminished pulses, end-organ dysfunction, and acidosis. For these patients, the therapy of choice is to urgently address the obstruction through the surgical or transcatheter approach. The clinical condition can be temporarily ameliorated through mechanical ventilation (to reduce afterload as well as the work of breathing) and the very careful use of short-acting vasodilators or inodilators, or very-low-dose inotropes and diuretics. Additionally, it is worthwhile considering maintaining ductal patency with PGE₁. In this situation, the addition of PGE₁ provides increased systemic perfusion (with deoxygenated blood) through a right-to-left ductal shunt.

Severe Aortic and Mitral Regurgitation

Heart failure due to mitral and/or aortic regurgitation is associated with marked activation of neurohormonal pathways and elevations in systemic vascular resistance. In these situations, the careful use of a combined veno- and arterial vasodilator, such as nitroprusside, has been shown to provide significant acute hemodynamic benefit by increasing the forward stroke volume while reducing ventricular filling pressures. Aortic regurgitation causes diastolic hypotension that falls further with afterload reduction, which may compromise coronary perfusion. Nonetheless, the fall in diastolic pressure appears to be well tolerated, as the relationship between myocardial oxygen delivery and demand improves substantially as a result of a significant increase in forward stroke volume and a reduction in ventricular diastolic and systolic loads. Inotropes with afterload-reducing properties may also provide significant hemodynamic benefit.

CONCLUSION

An understanding of the pathophysiology of cardiogenic shock, circulatory physiology, and therapeutic strategies that improve the relationship between global and myocardial oxygen supply and demand are necessary for optimizing the outcomes of patients at risk for or in a state of shock. Essential to this exercise is the timely and accurate assessment of cardiac function, cardiac output, and tissue oxygenation.

SUGGESTED READINGS

BACKGROUND

Extracorporeal membrane oxygenation (ECMO) is used to provide temporary mechanical support for severe cardiac, respiratory, or cardiorespiratory failure that is unresponsive to conventional therapies. While the principles of ECMO and cardiopulmonary bypass are similar in that they both support the cardiorespiratory system, ECMO is a bedside technology in the ICU that is used for much longer periods of support, and ideally should be a less complex and invasive procedure in terms of cannulation and perfusion techniques, durability of equipment, patient monitoring, and troubleshooting. ECMO cannulation is ideally extrathoracic (although for cardiac ECMO, direct cardiac cannulation is often used), and the miniaturized circuitry efficiently pumps blood through the extracorporeal circuit and provides excellent gas exchange through the membrane oxygenator, while minimizing complications of red cell trauma, embolization, clot, and mechanical failure.

The first reported successful outcome of ECMO support in the ICU was in 1971, in an adult trauma victim with severe acute respiratory distress syndrome who was supported for 75 hours on a custom-built ECMO circuit. Perhaps the most famous “first” in the history of ECMO was in 1975, when a newborn girl named Esperanza, who was dying of meconium aspiration syndrome, was rescued with the assistance of a research ECMO circuit. Esperanza (meaning hope in Spanish) became the symbol of the promise of this new therapy.

During the 1970s, there were a number of reports of the anecdotal use of ECMO in adults, in individual patients or small cohorts with variable outcomes, and then in 1979, a single randomized clinical trial from Zapol and colleagues demonstrated no benefit from ECMO over conventional treatment. ECMO was largely abandoned in adults for the next 2 decades.

The early pediatric and neonatal experience was different, and as a result of small but encouraging trials in neonates and children, pediatricians and pediatric surgical specialists continued to develop ECMO. The landmark UK Collaborative ECMO Trial in the early 1990s demonstrated improved survival in newborns with hypoxemic respiratory failure randomized for ECMO support compared to conventional therapy. The worldwide use of ECMO for neonatal acute hypoxemic respiratory failure continued to increase until the mid-1990s, with a decline over the past decade. The use of cardiac ECMO however began to take off in the late 1990s, and continues. The pandemic H1N1 influenza outbreak of 2009 resulted in a rapid surge in the international experience with ECMO to support respiratory failure in adults with encouraging outcomes reported at experienced ECMO centers in Australia, New Zealand, and the United Kingdom. Since then, the numbers of adults cannulated for ECMO has overtaken the annual number of pediatric (non-neonatal) cases worldwide. To date, since its inception in the mid-1980s, the Extracorporeal Life Support Organization (ELSO) registry (a voluntary international registry for pediatric and adult ECMO centers) reports more than 78 000 ECMO cases, with annual adult cardiac and respiratory ECMO numbers now exceeding their pediatric counterparts.

OVERALL GOALS OF ECMO

ECMO is a supportive, not curative, therapy, and offers a period of rest and recovery of the failing heart and/or lungs, and an opportunity for underlying diseases to be treated. Additionally, the initiation of ECMO should prompt a thorough reappraisal of the patient’s illness, aggressive diagnostic clarification, and implementation of a comprehensive treatment plan that will (1) help reverse the disease, and (2) maximize the patient’s chances of weaning from ECMO. Often before cannulation, ECMO patients are so unstable and are treated empirically without a full diagnostic workup. ECMO provides a relative physiologic stability that enables further evaluation using imaging techniques (eg, computed tomography [CT], echocardiography [echo], angiography), lung biopsy, or other interventions as needed, which then help inform prognosis and guide therapy.

The physiologic goals of ECMO are straightforward. ECMO must provide oxygenation and CO₂ removal and must adequately support the cardiac output and satisfy systemic oxygen delivery as needed. Current generation ECMO oxygenators are highly efficient at achieving gas exchange, as long as blood flow is adequate, and most ECMO pumps can provide supranormal blood flow that is fast enough to satisfy even the highest metabolic demands, such as in the setting of septic shock.

Most patients are cannulated for ECMO because they have a reversible condition and for these patients, ECMO serves as a “bridge to recovery.” Patients with acute hypoxemic respiratory failure with ARDS, or those with post-cardiopulmonary bypass myocardial stun are examples of common indications for ECMO in whom a period of organ rest is expected to be followed by recovery. Other patients are cannulated with the goal of supporting them until a transplantable organ (eg, lung or heart) or long term mechanical device (eg, ventricular assist device) can be placed. The therapeutic goal of ECMO for these patients is often referred to as “bridge to transplant,” “bridge to device,” or “bridge to decision.”

THE PHYSIOLOGY OF ECMO SUPPORT

ECMO, regardless of the mode, involves drainage of systemic venous blood to an external oxygenator that performs CO₂ removal and oxygenation, with subsequent delivery of oxygenated blood to the patient. While providing cardiorespiratory support, the introduction of the ECMO circuit inevitably disrupts the body’s natural equilibrium. This happens through a number of mechanisms. First, there is a need for active anticoagulation to prevent thrombosis caused by blood exposure to artificial surfaces. Second, the blood must be warmed in the extracorporeal circuit. Third, the contact of blood with exogenous surfaces and circuitry leads to an inflammatory response akin to the systemic inflammatory response. Last, in some cases, ECMO provides a portal for other physiology-altering therapies, such as continuous renal replacement therapy or plasmapheresis.

Carbon Dioxide Clearance

Carbon dioxide exchange depends on the gradient between the partial pressures of CO₂ (PCO₂) on either side of the membrane: the blood (venous) phase (PvCO2) versus gas phase (PgpCO₂), as well as the overall membrane surface area, the material and geometry of the membrane, and the gas flow rate through it. The gas that flows through the oxygenator, known as the sweep gas, has a PCO₂ of 0 mmHg. Since the oxygenator surface area, material, and geometry are fixed, the sweep gas controls CO₂ elimination, and the higher the sweep is, the greater the CO₂ removal is.

Oxygenation

Oxygen exchange depends on the gradient of O₂ partial pressures between the gas (PgpO₂) and blood (PvO₂) phases in the oxygenator, the surface area, and the O₂ permeability of the membrane, as well as the O₂ diffusion capacity in blood. The sweep fraction of inspired oxygen (FiO₂) determines the PO₂ gradient in the oxygenator, and sweep flow rate has little impact on blood oxygenation (except at the lowest sweep flow rates). Oxygenation determines the total O₂ content of blood (CaO₂), which itself depends on red blood cell (RBC) hemoglobin concentration and O₂ saturation.

Oxygen Delivery

Tissue oxygen delivery (DO₂) is a function of both blood O₂ content (CaO₂) and cardiac output (CO). The extracorporeal circuit generates continuous nonpulsatile flow, but the net circulatory support is very different depending on the mode of ECMO. The circulatory contribution of venovenous (VV) ECMO is in general secondary to the improvements in oxygenation of pulmonary and systemic blood, cardiopulmonary interactions due to reduced ventilatory settings, and right ventricular afterload on cardiac function. In contrast, venoarterial (VA) ECMO can support some or all of the cardiac output, or in some cases can provide even supranormal circulatory support depending on the flows and underlying condition. At higher flow rates, more blood is directed through the extracorporeal circuit and diverted from the native heart resulting in less pulsatility, as evidenced on the pulse contour of the arterial blood pressure trace.

MODES OF ECMO SUPPORT

The site of post-oxygenator blood return (vein, atrium, artery) determines the ECMO mode: thus for VV ECMO, oxygenated blood is returned to the systemic venous system (central vein or right atrium), and for VA ECMO, oxygenated blood is returned to the systemic arterial system (aorta, carotid artery, or femoral artery). Both modes involve drainage of systemic venous blood from large central veins or the right atrium. VA ECMO provides full or partial cardiopulmonary support for cardiac, respiratory, or cardiorespiratory failure, whereas VV ECMO primarily provides respiratory support with some secondary beneficial circulatory effects for patients with severe hypoxemic and/or hypercarbic respiratory failure.

Venovenous ECMO support

Indications

In general, VV ECMO is indicated for single organ (respiratory) failure, for pulmonary hypertension in the newborn, or in multisystem diseases with associated respiratory failure where cardiac function is relatively preserved. The list of conditions that can be supported with VV ECMO is extensive and continues to grow (Table 108-1), and this is due to many factors, including increased familiarity with the therapy, potential for rehabilitation on ECMO, and a readiness to offer longer ECMO runs (often exceeding 3–4 weeks) with a low complication rate in order to offer a realistic chance of lung recovery. Over the past decade there has been a philosophical shift in considering ECMO as a more standard therapy, which may be preferable to aggressive and potentially destructive ventilator strategies, rather than one of last resort.

Cannulation

Cannulation location and type are determined by patient factors as well as by institutional practice and preference for cannula type. Anatomic considerations include apparent vessel sizes and courses (assessed by precannulation imaging), presence of adjacent stomata or injuries that might contaminate a cannulation site, and likelihood of the need for patient mobility while on ECMO support. VV ECMO support can be achieved using 2 separate cannulae: 1 for drainage that is typically introduced to the femoral vein with the tip lying in the inferior vena cava, and a separate return cannula typically introduced to the jugular vein with the tip in the right atrium. Alternatively, a double-lumen cannula can be used that has the advantage of only using a single site, typically the internal jugular or subclavian vein. While more convenient, double-lumen VV cannulae are more challenging to position optimally, with not only the cannula tip but the orientation being important determinants of successful support. They can also be flow limited, particularly by venous drainage at higher flow rates.

Management Goals

VV ECMO management revolves around ensuring adequate tissue O₂ delivery and adequate and appropriate respiratory support, while of course providing meticulous general intensive care with careful attention to fluid balance, nutrition, sedation toxicity, infection control, and patient positioning, and providing physical rehabilitation as appropriate. Determining adequate tissue oxygenation in patients is an inexact science, but most intensivists have developed a sense of sufficiency using clinical and laboratory markers. Since tissue O₂ delivery depends on adequate hemoglobin concentration, arterial O₂ saturation, and cardiac output, these must be constantly assessed. For most VV ECMO patients, hemoglobin concentrations of 10 to 12 g/dL with SaO₂ in the 75% to 80% range are adequate. The reason for targeting (or accepting) lower saturations is that they are generally adequate to meet the body’s metabolic demands, and given that the oxygenated blood is returned to the right (not the left) side of the heart and that lung injury is severe in the early phase, higher saturations are often unachievable.

Without direct invasive measurement, the assessment of cardiac output is imprecise, but a combination of poor pulses, cool distal extremities, low mixed venous saturations, rising lactate concentrations, and declining near-infrared spectroscopy saturation trends may point to inadequate tissue oxygenation and inadequate cardiac output. The causes may be related to the adequacy of the ECMO support, and if such a situation arises acutely, it is essential to look for other contributory factors including sepsis, bleeding, tachycardia, or cardiac dysfunction, as well as mechanical reasons. In order to better support the cardiac output on VV ECMO, assuming that any acute mechanical issues have been ruled out, the options are to introduce inotropes and/or vasodilators to support the heart, or to convert to VA ECMO.

A factor that may have contributed to the dismal results of early ECMO trials in adults was that aggressive ventilation was often continued after cannulation. This practice gave way to the notion of rest settings, which are typically much lower than precannulation settings but aim to maintain lung inflation with a positive mean airway pressure, thus avoiding complete atelectasis. Some adult and pediatric centers have explored extubating VV ECMO patients, which also provides the potential advantage of avoidance of ongoing sedation, thus enabling some rehabilitation and spontaneous airway clearance while on ECMO.

Duration of Support

The anticipated duration of VV ECMO support will be related to the underlying disease. Typically, the duration is 1 to 3 days for pediatric asthma, less than 7 days for newborns with acute hypoxemic respiratory failure and pulmonary hypertension (eg, meconium aspiration), 10 to 14 days for congenital diaphragmatic hernia, and significantly longer—up to several weeks—for other pediatric respiratory indications (eg, infective, parenchymal, or interstitial).

Weaning

The actual practice of weaning a patient from VV ECMO is relatively straightforward, assuming that it is being done in a controlled, planned manner. More challenging is the timing of the decision of when to wean, and what are the respiratory goals for the transition off circuit. In some cases (eg, when a patient is having uncontrolled bleeding), the respiratory goal will have to be any ventilator settings that can keep a patient alive, including advanced strategies and the use of nitric oxide. Ideally however, intensivists will time the weaning based upon clinical and radiological signs of pulmonary improvement (eg, chest x-ray, tidal volumes, improving oxygen saturations, and good CO₂ clearance on reduced or no sweep gas flow). Most patients will transition from ECMO to low levels of positive pressure ventilation if still intubated on ECMO, and those that have been extubated on ECMO would generally be decannulated on noninvasive respiratory support.

An advantage of VV over VA ECMO is that a clinical trial off ECMO can be done with the circuit flowing and cannulae in position for any period of time, by disconnecting the sweep gas and capping the oxygenator, allowing the blood and gas phases inside to equilibrate over the course of 30 to 60 minutes, with ventilator or respiratory support being titrated according to the patient’s blood gases. If the support required with the oxygenator capped is on target, generally after a trial of at least 2 to 4 hours, the ECMO cannulae are removed. If the patient’s ventilator support requirements exceed the desired target, decannulation is postponed until a successful trial off is achieved.

Outcomes

Survival following VV ECMO varies according to the underlying pathology. For Neonatal respiratory ECMO, according to the ELSO registry 2015 International Summary, cumulative survival-to-discharge rates range from 75% to 84%, being highest among those patients cannulated for meconium aspiration syndrome (90%) and respiratory distress syndrome (84%) and lowest among those with congenital diaphragmatic hernia (51%). Beyond the neonatal period, survival is lower but closely grouped among all VV ECMO categories, between 64% and 69%. Highest survival rates occurred in patients cannulated for aspiration pneumonia (68%) and viral pneumonias (65%) and lowest for Pneumocystis pneumonia (51%). Since these numbers represent cumulative survival, they are slow to reflect recent improvements in VV ECMO technology and management, which routinely exceed 80% in many high-volume centers.

Venoarterial ECMO

Indications

VA ECMO provides pulmonary and circulatory support (Fig. 108-1). While it can be used in isolated respiratory disease (and historically was the mainstay for respiratory support), the need to cannulate a large artery that will either need to be ligated or repaired at the end of the ECMO course, and the higher complication rate of VA ECMO, make it a less desirable mode than VV ECMO for this application. VA ECMO is most appropriately deployed for either isolated severe cardiovascular disease (eg, dysrhythmia, severe contractile dysfunction post-cardiotomy, or cardiogenic shock), after cardiac arrest (eg, extracorporeal cardiopulmonary resuscitation, or ECPR), or in patients with combined respiratory and cardiovascular disease poorly responsive to inotropic or vasopressor support (Table 108-2). One important note of caution warrants mentioning: ECMO support in general is ineffective in the presence of aortic regurgitation, as a significant proportion of the postoxygenator arterial blood flows back to the left ventricle, resulting in inadequate cardiac output, coronary flow, and volume overload of the left ventricle.

Cannulation Strategies

Generally, the patient’s diagnosis and flow requirements drive the cannulation approach in VA ECMO. Patients with normal metabolic requirements need normal or mildly increased ECMO flows—typically 75% to 110% of their estimated cardiac output (eg, patients after cardiac surgery with a biventricular circulation or patients with a dysrhythmia)—whereas patients with increased metabolic needs may need flows that exceed 150% of normal cardiac output (eg, patients with single ventricle physiology or high-output septic shock). Patients requiring high-flow ECMO support, and a high proportion of those placed on ECMO after cardiac surgery, will typically undergo central cannulation in which short, wide-bore cannulae are placed via a sternotomy directly in the right atrium and ascending aorta to enable higher flow rates. Patients not requiring high flows, those without existing sternotomies, or those in the setting of cardiac arrest will typically be placed on VA ECMO through peripheral cannulation via the right internal jugular vein or a femoral vein, with blood returning to the right carotid artery or femoral artery. Femoral access is more commonly used in older children, adolescents, and adults.

Management Goals

VA ECMO aims to optimize tissue DO₂ to meet local metabolic demands and decrease myocardial work so the heart can rest and recover from the antecedent insult. Alternatively, in patients with end-stage cardiac failure, VA ECMO may provide temporary cardiac support as a bridge to a long-term cardiac ventricular assist device. VA ECMO helps with myocardial recovery by “unloading” the heart. This is accomplished by decreasing ventricular end-diastolic pressures, which optimizes coronary perfusion pressure, decreases wall stress, and decreases myocardial work. Cardiac output provided by VA ECMO usually approximates 80% of the total cardiac output, and there is always some blood return to the left ventricle, from the bronchial vessels for example, and at least minimal ejection through the aortic valve. This residual endogenous cardiac output helps prevent blood stasis and inappropriate cardiac loading. Sometimes, to achieve aortic valve opening, inotropic agents and/or vasodilators to decrease systemic vascular resistance (SVR) may be required. In situations with inadequate oxygen delivery, increasing ECMO flow, optimizing intravascular volume augmenting oxygen-carrying capacity by increasing hemoglobin concentrations to > 10 g/dL, and controlling the SVR may all help to improve tissue DO₂.

Achieving proper oxygenation and ventilation is generally straightforward with VA support, as the lungs (whether or healthy or diseased) are mostly bypassed by the extracorporeal circuit and as a consequence, there is less functional right-to-left shunting than in VV ECMO. As in VV ECMO, low ventilator settings (ie, rest settings) are typically used for most patients on VA ECMO, with the proviso that in patients on femoral ECMO support with cardiac ejection of pulmonary venous returning blood, the upper body oxygen saturation should be monitored to ensure adequate cerebral oxygenation. As in VV ECMO, the sweep gas flow controls CO₂ clearance, with higher flow increasing CO₂ removal. Oxygenation is managed by titrating the sweep gas FiO₂ and blood flow rate. High blood flow rates, however, may result in transit times through the oxygenator that exceed the system’s ability to maximally oxygenate the blood, especially if the venous blood is severely hypoxemic. Placement of a second oxygenator in parallel to the first cuts the transit time in half and doubles the amount of time the system has to oxygenate the blood prior to returning it to the patient.

Duration of VA ECMO Support

The anticipated duration of VA ECMO support depends on the underlying disease. If the indication is a low cardiac output or difficulty separating from cardiopulmonary bypass after cardiac surgery, then the ECMO run would generally be less than 1 week, with survival being much reduced beyond 10 days of support. Inability to achieve any weaning after 3 to 5 days should prompt the use of echo, CT, or cardiac catheterization to search for an additional cause (eg, residual cardiac lesion). In the setting of septic shock, assuming that high flows are achieved early in the disease course, a relatively short course (less than 5 days) would be anticipated. If VA ECMO support is a bridge to a long-term ventricular assist device, then this conversion is generally performed once the full assessment of transplant candidacy, and other organ system recovery, has occurred.

Weaning

Weaning patients from VA ECMO requires there to be improved cardiovascular function as evidenced by return of heart rate variability, control of dysrhythmias, decreased vasoactive medication requirements, and improving pulse pressure, suggesting ejection. It is important to prepare the patient for a successful wean. To that end, it is essential to optimize ventilator support, institute appropriate inotropic therapy before weaning to give the vasoactive medication time to work, optimize hematocrit level (> 30%), and minimize excess oxygen consumption with adequate sedation and analgesia, and neuromuscular blockade if required.

A preliminary assessment of weaning readiness may involve decreasing ECMO flows to 30% to 50% of predicted cardiac output, in a sufficiently loaded heart, potentially with a modest degree of inotropic support. If a patient becomes tachycardic or hypotensive and/or develops lactic acidosis, then he/she may not be ready to wean from ECMO. Contractile function can also be assessed by echocardiogram on decreased ECMO flows to help predict how the heart will perform off mechanical support (Table 108-3).

Outcomes

According to the latest ELSO International Summary, overall survival in VA ECMO for cardiac disease ranges between 41% and 51% in neonatal and pediatric patients, respectively, and is 41% for ECPR. These results represent an average of survival rates from centers with different expertise and resources. Centers with lower patient volumes and less experience likely experience lower overall survival (and likely higher morbidity) than higher-volume ECMO centers, whose outcomes are more favorable.

COMPLICATIONS

ECMO is a low-volume, high-risk therapy that includes complex circuitry, multiple connections and access points, and systemic anticoagulation, and therefore represents a rich source of potential complications. Key players include bleeding and clotting (see below), sepsis, additional organ failure, or mechanical complications. Indeed, mortality related to complications is as important a consideration as mortality related to the underlying disease or lack of recovery from ECMO. The ELSO registry tracks 9 classes of complications (Table 108-4).

ANTICOAGULATION

Finding the proper balance between blocking the natural tendency for blood exposed to thrombogenic surfaces to clot and preventing uncontrolled bleeding within the body is critical to a successful ECMO run. An ideal anticoagulation regimen would achieve these dual goals and would also be easy to monitor and titrate. Not surprisingly, neither the ideal anticoagulant nor the ideal monitoring system exists. Unfractionated heparin remains the preferred drug for anticoagulation while on ECMO and is given as a bolus during cannulation, and then as an infusion throughout the period of support. Most ECLS programs monitor the anticoagulation effect using the activated clotting time (ACT) as the primary endpoint, with therapeutic ranges varying from 150 to 220 seconds, depending on the specific assay and the type of heparin (porcine or bovine). Other tests of heparin’s anticoagulation activity include partial thromboplastin times, heparin or anti-factor Xa levels, and other standard coagulation tests to monitor the blood’s innate ability to clot. Some units monitor the whole blood clotting tendencies using the thromboelastogram or similar technologies. Collaboration between the ECMO and coagulation teams is an indispensable method to optimize the anticoagulation regimen for individual patients, speed the diagnosis of the frequent complications of systemic heparinization (eg, bleeding, heparin resistance, thrombophilia), and prompt their rapid and specific correction.

FUTURE DIRECTIONS

Over the past four decades, ECMO has evolved from a therapy of last resort with an almost unacceptably high mortality rate to a more routine consideration in the management of severe cardiopulmonary disease. While inroads against the major complications associated with ECMO have been made, more work needs to be done. Nonetheless, there is cause for optimism, and as ECMO approaches its 50th anniversary, the prospects for routine integration into critical care have never been brighter. The field has started to solve some of the major problems that have resulted in lethal complications for patients supported for more than a few days, and technology has become both highly sophisticated and less beset with risk, while at the same time becoming simpler to manage. Over time, it is likely that ECMO will continue to become a more routine therapy that replaces more hazardous frontline treatments for severe cardiopulmonary diseases.

SUGGESTED READINGS

INTRODUCTION

The term perioperative care refers to all the clinical activities that take place around a surgical intervention, from the preoperative assessment to the postoperative discharge. In modern institutions, perioperative care is performed in a defined physical area that usually accommodates scheduling, admission, preoperative preparation, the procedure itself (the operating room or a procedural suite), recovery from the procedure (phase I and II), and discharge. Support services—including sterile processing, materials management, the pharmacy, the laboratory, pathology, and diagnostic imaging—are frequently adjacent. The success and subsequent demand for new procedures such as cardiac catheterization or endoscopy, which are not strictly surgical but require expertise and equipment similar to surgery, have created a need to duplicate existing perioperative resources in other areas. The same principles that are detailed here for surgical procedures apply to these procedural areas.

PREOPERATIVE ASSESSMENT AND CARE

Nearly all parents and children experience anxiety at the prospect of surgery. Preparing children for their surgical and operating room experience should begin before the patient makes contact with their anesthesiologist. The surgeon or primary care provider should be the first ones to ensure that the child receives the best perioperative care possible and to assure the child’s family that this will be the case, answering their questions or facilitating access to someone who can answer them.

Many institutions arrange orientation and a preoperative tour days before surgery to instruct and educate the child and parent on what to expect. As part of the preoperative program, most of these institutions have developed child-friendly environments, with coloring books, tours, videos, and the invaluable help of a child-life specialist who can help reduce the level of anxiety. Allowing children to take a favorite toy or blanket into the operating room provides great comfort to many children.

The primary goal of preoperative assessment and care is preparing a patient for surgery, and ensuring that a full assessment of risk has been performed. This involves the following points: (1) screening for conditions that may require consultation workup or treatment, (2) optimizing any pre-existing medical conditions, and (3) counseling patients and parents about the expected course of anesthesia and surgery. The anesthesiologists should make every effort to see the child and the family to understand their needs and expectations and to perform a complete preoperative evaluation. The anesthetic, postoperative pain management, and discharge plans can all be explained to the child, if appropriate, and to the family at this time. This is also a good opportunity to obtain informed consent.

The preoperative evaluation usually consists of a careful history and an targeted physical examination aimed at establishing surgical risk and designing anesthetic planning and postoperative care. The physical examination should be sufficiently complete to detect any major problems and should be informed by the medical history (Table 109-1).

Laboratory Evaluation

Because of low predictive value and cost-effectiveness, many centers no longer perform routine laboratory screening tests in otherwise healthy children scheduled for minor elective surgery. However, important baseline values (eg, hemoglobin level determination) in anticipation of major blood loss or specific diagnostic tests to rule out major organ-specific pathology are still performed. At present, there are no universal guidelines concerning preoperative laboratory testing; many anesthesiologists follow institutional protocols or their own personal preferences.

Mild anemia (hemoglobin [Hb] ≤ 9.5 g/dL) has been reported in up to 2% of pediatric patients undergoing elective surgery. No studies to date have shown justification to modify the perioperative management in otherwise healthy children who have a borderline low hemoglobin concentration, and therefore routine hemoglobin determinations are usually not necessary. However, it is reasonable to obtain preoperative hemoglobin values in (1) former preterm infants (postoperative apnea), (2) patients with chronic illness, (3) children with sickle cell disease (including sickle cell anemia, sickle Hb C disease, and the sickle thalassemias), and (4) those instances where having a baseline hemoglobin value may be useful (along with blood typing and cross-matching) in anticipation of significant surgical loss. In patients with sickle cell disease, the optimal level of sickle Hb and the role of transfusion to avoid perioperative complications in different categories of surgery and anesthesia (regional vs general) is still unknown. Perioperative evaluation and management of patients with this disease is best performed in conjunction with the patient’s hematologist and surgeon.

Routine coagulation screening also has a weak predictive value for assessing the risk of surgical bleeding in patients with no history of abnormal bleeding. Screening is only advisable in patients in whom a hemostatic defect is suspected either by history or physical examination, patients in whom minimal postoperative bleeding could be critical (eg, neurosurgical interventions), or when the surgical procedure is likely to cause perioperative coagulation dysfunction (eg, cardiopulmonary bypass).

The value of routine preoperative pregnancy testing in women of childbearing age, particularly in adolescent females, is controversial. There is no scientific evidence that short-term exposure to anesthetics can induce spontaneous abortions or congenital malformations. However, surgery involves physical and psychological stresses, and exposure to x-rays and nonanesthetic medications may result in finite fetal risk.

Arguments against preoperative pregnancy testing are based on lack of demonstrated cost-effectiveness and on confidentiality and consent issues.

Perioperative Surgical Home

Efforts to improve quality, efficiency and safety of perioperative care have led to the development of the perioperative surgical home. This concept involves more than a preoperative or preanesthesia assessment clinic; it seeks to integrate the preoperative evaluation of the patient and preparation of the family and patient for the entire course of perioperative care. This includes coordination of medical and surgical specialists and developing a plan for the preparation and intra-hospital care that allows the patient to safely transition into the surgical home and back into the care of their primary medical specialist. “Roadmaps” not only establish expectations for the family and perioperative team, but they allow for the development of procedural dashboards that can assist in monitoring the patient’s progress throughout the course of care. In pediatrics, the use of roadmaps has been proposed as a method of safely managing medically complex medical patients outside of their medical home. The focus of this approach is family- and patient-centered care and utilizes the electronic medical record to improve communication between providers to deliver safe, high-quality, and efficient care.