Physical Examination

Electrocardiography

In pediatric anesthesia, ECG is most useful for tracking the heart rate and diagnosing intraoperative rate-related arrhythmias, of which the two most common are bradycardia and supraventricular tachycardia. ECG is much less prone to movement-related artifact than the original pulse oximeter, although new pulse oximeter devices have eliminated most motion artifacts. In small infants, hypoxemia-related bradycardia may occur before the pulse oximeter reveals oxyhemoglobin desaturation. Conversely, resolution of hypoxemia is heralded by the transition from bradycardia to normal sinus rhythm. Premature ventricular contractions are commonly observed when halothane is used as the general anesthetic agent, especially during periods of hypercapnia or catecholamine release. The precordial stethoscope as a single monitor provides a much better indication of cardiac contractility and thus the overall hemodynamic status.

Electrolyte abnormalities may also be uncovered through the use of ECG. Hyperkalemia produces the characteristically prominent T waves. Hypocalcemia, which may occur during rapid administration of citrated blood products, prolongs the QT interval. Because ischemic changes in normal pediatric patients are rare, and because lead II provides a good view of atrial activity for arrhythmia diagnosis, the latter is recommended for the routine intraoperative electrocardiographic monitoring of pediatric patients.

In children, the normal heart rate varies with age (Table 11-1). The normal heart rate of the newborn ranges from 120 to 160 beats per minute, although lower rates (e.g., 70) are frequently observed during sleep, and higher rates (>200) are common during anxiety or pain. Heart rates tend to decrease with age and in parallel with decreases in oxygen consumption. In addition, many children have a noticeable variation in heart rate with respiration (i.e., sinus arrhythmia).

TABLE 11-1 Normal Resting Heart Rates of Infants and Children

Modified from Liebman J, Plonsey R, Gillette PC, editors: Pediatric electrocardiography, Baltimore, Md, 1982, Williams & Wilkins.

Systemic Arterial Pressure

Noninvasive Measurement

Blood pressure is easily measured noninvasively in children and small infants using oscillotonometry. In children, oscillometric measurements of systolic arterial pressure (Bruner et al., 1981; Friesen and Lichtor, 1981) and mean arterial pressure (Kimble et al., 1981) usually correlate well with the Riva Rocci mercury column method, as well as with direct arterial pressure measurement, but oscillometric measurements tend to underestimate the diastolic component. During routine uncomplicated cases, measurement of blood pressure should be performed every 3 to 5 minutes while the child is anesthetized—determinations that are too frequent can result in limb ischemia. The blood pressure cuff is most commonly placed on the upper arm but can be placed on the forearm, thigh, or calf. There is inconsistent correlation of measurements obtained between the upper and lower limbs.

The width of the blood pressure cuff should cover approximately two thirds of the total length of the upper arm (or other extremity portion to which it is applied). A cuff that is too small or too narrow incompletely occludes the artery, resulting in the premature return of detectable flow and hence falsely increasing the pressure measurement (Park et al., 1976; Kimble et al., 1981). The error can be as great as 30 mm Hg. A cuff that is too wide can dampen the arterial wave and result in a falsely low pressure, but the magnitude of this error is small (Kimble et al., 1981). Blood pressure increases gradually throughout childhood (Figs. 11-1 and 11-2) and depends on the height of the child: taller children demonstrate a higher blood pressure (Table 11-2). Blood pressure ranges in premature infants have been defined (Table 11-3) and vary depending on the health status of the infant and mother.

FIGURE 11-1 Age-specific percentiles of blood pressure measurements in boys, from birth to 12 months of age. Values for girls are slightly lower.

(From National Heart, Lung, and Blood Institute: Report of the Second Task Force on Blood Pressure Control in Children, Bethesda, Md, 1987, The Institute. Reproduced by permission of Pediatrics 79:1, copyright 1987.)

FIGURE 11-2 Age-specific percentiles for blood pressure measurements in boys, 1 to 13 years of age. Values for girls are slightly lower.

(From National Heart, Lung, and Blood Institute: Report of the Second Task Force on Blood Pressure Control in Children, Bethesda, Md, 1987, The Institute. Reproduced by permission of Pediatrics 79:1, copyright 1987.)

TABLE 11-2 Systemic Arterial Pressure

TABLE 11-3 Blood Pressure Ranges in Healthy Premature Infants*

Systolic Pressure Variation

Systolic pressure variation is a noninvasive way to determine volume status and fluid responsiveness. It is defined as the difference between the maximal and minimal values of systolic blood pressure during a positive pressure breath. Initially during a positive pressure breath, there is a transient increase in systolic blood pressure (delta up) followed within four or five beats by a decrease in systolic blood pressure (delta down). Increases in intrathoracic pressure during positive pressure ventilation cause a decrease in systolic blood pressure because of decreased preload to the right ventricle, increased afterload to the right ventricle, and decreased afterload to the left ventricle. This decrease is greater during hypovolemia. Systolic pressure changes in response to respiratory variation have been used to determine hypovolemia (Greilich and Johnston, 2007).

The difference between the maximal systolic blood pressure and minimal systolic blood pressure during a respiratory cycle can help predict volume status. In fact, when this value is divided by the mean of the two systolic pressure values, it provides a percentage of respiratory change in arterial pulse pressure. The equation for this calculation is as follows:

where ΔPP% is respiratory change in pulse pressure (mm Hg).

Michard and others (1999) demonstrated a strong relationship in adult ventilated patients between pulse pressure changes and cardiac output. Patients with pulse pressure changes (ΔPP) that are greater than 10% may be fluid responsive and benefit from the administration of intravenous fluids.

Changes in the pulse oximetry waveform (plethysmographic waveform amplitude) have been shown to predict fluid responsiveness (Pizov et al., 2010). Bedside use of this variable is challenging. The “pleth variability index” (Masimo Corp., Irvine, CA) automatically calculates the waveform amplitude variation and may predict fluid responsiveness noninvasively (Cannesson et al., 2008a, 2008b).

Central Venous Pressure

There are four relative indications for central venous catheterization: inadequate peripheral venous access, central venous pressure monitoring, infusion of hyperosmolar or sclerosing substances, and a planned operative procedure with a high risk for hemodynamically significant venous air embolism. There is no absolute indication for central venous pressure monitoring in pediatrics. Unlike direct systemic arterial pressures, central venous pressure itself rarely provides the sole basis for therapeutic action. It does, however, provide useful information that, taken together with other data, help to form a management plan. The procedures for which this monitoring deserves consideration include large estimated blood loss or fluid shifts (>50% EBV), deliberate hypotension, cardiac surgery with CPB, situations in which the usual signs of hypovolemia are likely to be misleading (e.g., renal failure, congestive heart failure), and procedures with expected moderate blood loss or fluid shifts. The normal values for central venous pressure in children are similar to those in adults (mean, 2 to 6 mm Hg).

Every insertion site that has been used in adults can be used in children. Access to the central circulation can be achieved from the internal and external jugular, subclavian, basilar, umbilical, and femoral veins. The site selected depends on the experience of the operator and the indication for the catheter. If venous access is the only requirement, one might elect to use visible veins (e.g., basilar, external jugular) or those with a lower risk for complications (e.g., femoral). Situations that require true intrathoracic central venous placement also require placement of the catheter into the internal jugular vein or subclavian vein. The umbilical vein can be used in neonates for volume resuscitation, but the high frequency with which these catheters enter the branch portal veins introduces a significant risk for permanent liver injury if sclerosing or hyperosmolar solutions are infused. Because a catheter tip can erode through the wall of the right atrium, care must be taken to avoid intracardiac tip placement. The catheter should be advanced only until the orifice lies in the intrathoracic great vessels, and its position should be confirmed radiographically.

Catheters of various sizes (2.5 to 10 French), lengths, and composition are available for pediatric applications (Cook Critical Care, Bloomington, IN, and other companies). Selection is based on the size of the patient (Andropoulos et al., 2001) and the purpose of the catheter. The composition of the catheter depends on its intended use. Teflon is fairly resistant to thrombus formation, but concerns about perforation by catheters have prompted the development of softer catheter materials, especially for long-term use (e.g., Silastic and polyurethane). The catheters are generally inserted via the Seldinger technique, using landmarks that are similar to those used in adults.

There are no absolute contraindications to placing a central venous catheter, but each site has potential risks. All sites share the common complications of infection (site cellulitis, bacteremia), venous thrombosis with potential emboli, air embolism, catheter malfunction (occlusion, dislodgment, or fractures), dysrhythmias (when the catheter tip is in the heart), and bleeding. Universal precautions and sterile technique should be used when placing a central venous catheter. The risks involved in cannulating the internal jugular vein include carotid artery puncture, Horner’s syndrome, pneumothorax, and injury to the thoracic duct when the left internal jugular vein is cannulated. The high approach to the internal jugular vein, at the midpoint of the sternocleidomastoid muscle, results in comparable success with fewer complications than lower approaches (Coté et al., 1979). Two-dimensional ultrasound scanning improves localization of the internal jugular vein and increases the success rate of central venous cannulation in adults and children (Verghese et al., 2002; Hind et al., 2003). Using this device, Alderson and others (1993) reported an 18% prevalence of anatomic variations in children younger than 6 years that would preclude or significantly hinder the successful cannulation of the internal jugular vein using anatomic landmarks alone. In addition, Hong and colleagues (2010) reported that rotating the head away from the neutral position increases the degree of carotid artery and internal jugular vein overlap, and decreases the incidence of lateral positioning of the internal jugular vein to the carotid artery.

Mixed Venous Oxygenation and Monitoring

Mixed venous oxygenation is defined as the oxygenation saturation or content of venous blood from the pulmonary artery. The collection of a mixed venous sample requires a pulmonary artery catheter. The Fick equation for mixed venous oxygen saturation (Svo₂) depends on four variables: oxygen consumption (), cardiac output, hemoglobin, and arterial oxygen saturation.

Normal values for mixed venous saturation range from 65% to 75%. A decrease in Svo₂ occurs because of increased oxygen consumption (stress, pain, hyperthermia, shivering) or decreased oxygen delivery (anemia, decreased cardiac output, decreased Pao₂, decreased Sao₂). An increase in Svo₂ occurs because of a decrease in (hypothermia, anesthesia) or an increase in oxygen delivery (increased hemoglobin, increased cardiac output, increased Pao₂, increased Sao₂). Mixed venous oxygenation is used to assess the balance between oxygen delivery and oxygen consumption for patients in the operating room and intensive care unit. It is a global index of tissue oxygenation.

A significant limitation of mixed venous saturation is that it requires the collection of blood from the pulmonary artery. This is a particular limitation in the smaller pediatric patients. Blood from the superior vena cava (central venous saturation [Scvo₂]) and right atrium have been investigated as possible surrogate markers for mixed venous oxygen saturation, with mixed results.

Dueck and coworkers (2005) found a significant variation between central venous saturations and mixed venous saturations in the same patient, so individual values of Scvo₂ cannot be substituted for Svo₂ values. However, although the absolute values do not correlate, there is a correlation between the trends in Scvo₂ values and in Svo₂ values. Perez and colleagues (2009), in a retrospective pediatric study, identified a correlation between right atrial and mixed venous oxygen saturations. Scvo₂ is used clinically in pediatric patients. Continuous monitoring of Scvo₂ can be performed in neonates, infants, and children.

In infants, continuous monitoring of Scvo₂ can be achieved with a fiberoptic probe. One type of fiberoptic probe is designed as a percutaneous catheter from CeVOX (Pulsion Medical Systems AG, Munich, Germany). This 2-F probe is 31 cm in length and measures central venous oxygen saturations using spectrophotometry. Muller and coworkers (2007) described placing the catheter percutaneously through a 16-gauge single-lumen catheter in the femoral or subclavian vein. There were only three patients in this study, so accurate correlation with central venous blood samples cannot be determined.

The Pediasat system (Edwards Life Sciences) has been described in infants and children having orthopedic, craniofacial, and cardiac surgery. It comes in four sizes (4.5 F, 5 cm; 4.5 F, 8 cm; 5.5 F, 8 cm; and 5.5 F, 15 cm) and provides continuous readings of central venous oxygen saturation. Liakapolous and colleagues (2007) and Ranucci and colleagues (2008) demonstrated good correlation between Scvo₂ values from the Pediasat system when compared with co-oximetry values obtained from blood samples drawn from the distal port.

Artifact may limit the clinical usefulness of Scvo₂ catheters for intraoperative care. Manipulation of the catheter by the surgeon when the chest is open or direct light from the surgeon’s headlight may alter the readings of the catheter.

Pulmonary Artery Catheters

Since its introduction in 1970, indications for the use of the flow-directed balloon-tipped pulmonary artery (Swan-Ganz) catheter in pediatric patients have been slow to evolve. Although the validity and value of the data that these catheters generate remain controversial in pediatrics, the technical difficulties and complications associated with their use are significant. Pulmonary artery pressure measurement can help guide therapy in children with elevated or volatile pulmonary vascular resistance, but the interpretation of the flow data they generate is hindered by several factors. First, the desired cardiac output varies according to age, disease state, and other elements of management that alter metabolic demand in complex ways, thereby introducing significant uncertainty in assigning a target value. Second, the prevalence of intracardiac communications that permit shunting of blood causes discrepancies in pulmonary and systemic blood flow that may vary continuously and are difficult to quantify. Finally, despite several studies demonstrating reasonable accuracy when thermodilution is compared with other methods of flow determination, such as the Fick equation (Freed and Keane, 1978) and dye dilution (Colgan and Stewart, 1977), the precision of these determinations in small infants is low and has a 25% intersample variability. In patients with congenital heart malformations, for example, measurement errors are introduced by shunting and complex anatomy, and the risks of improper placement of the flow-directed pulmonary artery catheter are increased. Alternatively, directly placed pulmonary artery catheters can provide the necessary information regarding pulmonary vascular resistance and residual left-to-right shunts, and left atrial catheters reflect filling and diastolic function of the left ventricle after cardiac surgery.

In some situations, pulmonary artery catheters provide useful information. In children who have severe coexisting pulmonary and circulatory failure, pulmonary artery catheters can help to quantify the hemodynamic impact of extreme respiratory support measures and guide complex fluid and pharmacologic regimens. They may also be useful in patients with underlying pulmonary hypertension or poorly compensated left ventricular dysfunction who undergo acute surgical stress (e.g., arteriovenous malformation clipping or aortic cross-clamping). Given the uncertainty about optimal systemic flow in a given child, mixed venous oxygen saturation may serve as a better indication of global perfusion. In the absence of left-to-right shunts, this sample is best obtained from the pulmonary artery.

Pulmonary artery catheters can be difficult to insert, especially in infants or in children with low cardiac output. They may be placed in any vein used for access to the central venous system, but the most reliable veins are the right internal jugular and the femoral. In infants and children smaller than 15 kg, it is technically difficult to place an introducer sheath in the neck vessels; the femoral veins are preferable. Multilumen catheters capable of thermodilution are available in two sizes, 5 and 7 F, with four options for the right atrium–to–pulmonary artery interluminal distance. Catheter recommendations are based on age (Table 11-4). The proper placement of these catheters can take a long time, and thus the assistance of fluoroscopy is recommended for infants and children less than 30 kg and for larger children who have a low cardiac output.

TABLE 11-4 Guidelines for Multilumen Pulmonary Artery Catheters for Infants and Children

The risks of balloon-tipped pulmonary artery catheters are numerous and include the risks of central venous catheter placement discussed previously, as well as the complications seen in adult patients with pulmonary artery catheters: infection, air emboli, thrombus, pulmonary artery rupture, acute right bundle branch block, and intracardiac knots. Other complications are more common with children: misleading information, paradoxical systemic emboli, disruption of an intracardiac repair, and high-grade right ventricular outflow tract obstruction because of the relatively large balloon diameter. The presence of intracardiac and extracardiac malformations may result in an aberrant catheter course, leading to incorrect data and an increased risk for systemic emboli.

Cardiac output can be estimated in children through indicator dilution (e.g., thermodilution or dye dilution) and noninvasive techniques. Doppler determinations of aortic blood velocity can be used to quantify systemic flow if the angle of the incident ultrasound beam and the cross-sectional area of the aorta are reliably determined (Alverson et al., 1982). Transthoracic and transesophageal evaluations of Doppler cardiac output in children have proved to be less promising (Notterman et al., 1989; Muhiudeen et al., 1991). Thoracic bioimpedance, a method that estimates stroke volume on the basis of changes in thoracic impedance, has been applied to children as small as 3.6 kg. Although some correlation exists between bioimpedance and indicator dilution methods, reproducibility is poor (O’Connell et al., 1991). Further details and the complexities encountered in the measurement of cardiac output in children are beyond the scope of this chapter but have been reviewed previously (Tibby and Murdoch, 2002).

A noninvasive cardiac output monitor has been developed that determines cardiac output via the Fick principle for rebreathed CO₂ (Respironics; Novametrix Medical Systems Inc., Wallingford, CT) (Capek and Roy, 1988). The noninvasive cardiac output monitor has been clinically validated in adults and is approved by the U.S. Food and Drug Administration (FDA) for use, but it requires tidal volumes of 200 mL or greater (Guzzi et al., 2003; Watt et al., 2004).

Transesophageal Echocardiography

The value of transesophageal echocardiography (TEE) for monitoring hemodynamics and to evaluate preoperative and postoperative cardiac anatomy has been appreciated from the time of its introduction to the operating room. In the infancy of TEE in the late 1970s, an M-mode transducer was passed into the esophagus, plotting the distance of structures from the transducer on the y-axis and time on the x-axis (Frazin et al., 1976). Although they provided valuable information, M-mode images were too limited alone. In 1982, Schlüter and colleagues (1982) described their experience using a transducer capable of two-dimensional images mounted on a gastroscope. The usefulness of the images obtained was readily apparent, and since that time technology has catapulted the field of TEE to the forefront of cardiovascular monitoring. Now, multiple companies (Phillips, Acuson/Siemens, General Electric) manufacture advanced TEE-specific probes capable of two- and three-dimensional imaging on top of the original M-mode. Doppler has also been incorporated, providing the examiner the ability to extrapolate vast amounts of information from their patients.

The probes use ultrasound waves generated by the vibration of piezoelectric crystals in the tip of the transducer. Theses waves have frequencies between 2.5 and 7.5 MHz. The wave’s ability to travel through tissues depends on the specific density of the tissues. Differences in impedance cause some of the energy generated by the crystals to be reflected and allow some to continue on to the next tissue plane until all the energy has dissipated. If the impedance is too great, as is the case with bone, all of the energy will be reflected. Images are then generated on the basis of the time from the impulse and the remaining energy found in the waves as they return to the transducer.

The quality of the image depends on a number of factors. The closer structures are to the transducer, the greater is the intensity of the image. Transducers with higher frequencies have less penetration than lower-frequency transducers; thus, if one is trying to visualize structures far away, a lower frequency provides better imaging. Images can also be adjusted with gain, so that unwanted echoes in the near field may be dampened and echoes in the far field enhanced.

Doppler technology uses low-intensity ultrasound reflected from moving columns of blood. The frequency of the returning echo is analyzed and estimations of direction and velocity are made. Accuracy of the estimation depends on the direction of flow relative to the direction of the beam. Parallel orientation provides the most accurate values, and increases in the angle of measurement falsely lower the predicted velocities. Measurements taken at angles greater than 20 degrees have a significant amount of error. Both pulsed-wave Doppler, if one is interested in determining velocity at a specific location, and continuous wave, to identify the highest velocity across an entire line of sight, are available for use during a TEE examination.

A modification of the Bernoulli equation allows an examiner to estimate pressure gradients using the measured velocities. Simplified, the change in pressure between two points is equal to four times the maximum velocity squared (Holen et al., 1977; Hatle et al., 1978). Using this principle, the stenosis of valves, severity of aortic coarctation, obstruction caused by muscle bundles or membranes, and a multitude of other clinical questions can be answered.

TEE has played a vital role in improving outcomes in pediatric patients with congenital heart disease. Typically, examinations are performed preoperatively to confirm anatomy, and postoperatively to assess repairs and evaluate function. One study looking at 865 consecutive examinations demonstrated alterations to surgical plans based on preoperative TEE examinations in 2% of patients. The same study found that 12% of patients had post-bypass examinations that led to surgical interventions. Many of these interventions saved patients from unnecessary revision operations and their associated morbidity. In addition to surgical interventions, medical management with drugs and fluids was affected in 20% of the examined patients (Bettex et al., 2003).

An earlier report found that pediatric patients leaving the operating room without residual defects seen on the echocardiographic examination had a risk for reoperation of 3%, versus 42% for those who were found to have residual defects (Ungerleider et al., 1989). This ability to detect problems early provides a significantly improved outcome and subsequently reduces costs, as the need for repeat operations is reduced (Randolph et al., 2002).

As evidence supporting the value of TEE in cardiac surgery has increased, recommendations for proper performance of the examination have matured. The American Society of Echocardiographers and the Society of Cardiovascular Anesthesiologists have developed guidelines for adult intraoperative TEE examinations to encourage complete examinations with standardized views and nomenclature so as to improve communication between various care providers (Shanewise et al., 1999). Discussion of the complete examinations is beyond the scope of this chapter, but some of the probe locations and angles are useful for obtaining basic information regarding a patient’s condition (Table 11-5). Similar guidelines have also been published to assist clinicians with the pediatric examination (Lai et al., 2006).

TABLE 11-5 Transesophageal Echocardiography Cross Sections

Modified from Shanewise JS, Cheung AT, Aronson S et al: ASE/SCA guidelines for performing a comprehensive intraoperative multiplane transesophageal echocardiography examination: recommendations of the American Society of Echocardiography Council for Intraoperative Echocardiography and the Society of Cardiovascular Anesthesiologists Task Force for Certification in Perioperative Transesophageal Echocardiography, J Am Soc Echocardiogr 12:884, 1999.

Although the benefits have been demonstrated, providers must also be aware of the complications associated with the TEE examination. Problems that may be encountered include damage to the oral cavity, esophagus, or stomach; compromised ventilation; inadvertent extubation; right main stem advancement; vascular compression; and arrhythmias. In light of these serious but rare events, caution should precede placement of the transesophageal probe (Stevenson, 1999). Contraindications to the TEE examination include an unrepaired tracheoesophageal fistula, esophageal web, and recent esophageal or gastric surgery. Failure to ascertain a history of such events may lead to significant damage or even perforation of the esophagus.

Another issue encountered in the pediatric population is the size of the patient. Adult-sized probes are generally appropriate for patients older than 8 years or heavier than 20 kg. Pediatric probes are available for patients less than 20 kg. Most clinicians refrain from placing transesophageal probes in patients weighing less than 3 kg. If echocardiographic images are important to a particular surgery and TEE is contraindicated or impossible, an epicardial probe can be passed to the operative field via a sterile sleeve to obtain necessary imaging.

Temperature

Temperature monitoring is vital during pediatric anesthesia, as children may exhibit hypothermia or hyperthermia, both of which can have profound physiologic consequences (see Chapter 6, Thermoregulation).

Urine Output

Urine output often reflects intravascular volume status and cardiac output. Proper assessment of urine output requires recognition of the physiologic mechanisms that exert an affect on urine flow in children. During the first week of life, the glomerular filtration rate and renal plasma flow are only 25% of normal adult values (Arant, 1978). The neonatal kidney is limited in its ability to concentrate the urine (Simpson and Stephenson, 1993). By the end of the first week of life, the kidneys begin to reach absorption thresholds for sodium and glucose that approach adult levels.

Normal newborns produce between 0.5 and 4 mL urine/kg per hour in the first 3 hours of life (Strauss et al., 1981). Urine flow, which initially ranges from 15 to 60 mL/kg per day, reaches as much as 120 mL/kg per day by the end of the first week of life, with 90% of neonates producing 0.5 to 5 mL/kg per hour (Douglas, 1972; Guignard, 1982). In the neonate who is less than 1 week old, urine flow alone is not a sensitive index of changes in cardiac output or intravascular volume. The limited capacity of the neonatal kidneys to compensate for diminished or excessive intravascular volume demands more precise management of blood and fluid replacement in these infants. Beyond the neonatal period, a urine flow of 0.5 to 1 mL/kg per hour usually indicates adequate renal perfusion and function.

Intraoperative monitoring of urine output is indicated in procedures in which large shifts in fluid, blood, or hemodynamics are anticipated, including blood loss greater than 20% of the EBV, third-space replacement exceeding 50% of the EBV, CPB, neurosurgery, deliberate hypotension, planned use of diuretics, or planned hemodilution. Silastic Foley catheters are available in sizes small enough (6 F) for full-term neonates. Alternatively, a small feeding tube can be used in premature infants and in those with a small urethra. In infants, urinary bladder catheters should be connected to a urinometer capable of measuring small volumes, or to a vented 10- to 20-mL syringe.

Noninvasive Respiratory Gas Monitoring

Carbon Dioxide

Capnometry is the instantaneous measurement of CO₂ in the breathing circuit; it depicts this information in a continuous graphic display in which both the quality and the quantity of ventilation can be evaluated (Figs. 11-3 to 11-6).

FIGURE 11-3 Common capnographic diagnoses: rebreathing. Top: Normal capnographs. The graph on the right is compressed over a longer time Note the plateau, suggesting valid end-tidal CO₂ data and the return to baseline between breaths; there is no rebreathing. Bottom: Rebreathing, as there is no return to baseline. This can occur with inadequate fresh gas flow or floating unidirectional valves. The small initial deflection before exhalation (the preexhalation hump) can occasionally be seen. It represents the inhalation late in the inspiratory phase of more concentrated exhaled gas from the previous exhalation.

FIGURE 11-4 Common capnographic diagnoses: poor sampling. Left: Poor sampling as evidenced by absence of a plateau phase. This diagnosis could not be made with a capnometer that was incapable of real-time graphics. This is typical of small neonates whose small exhaled volumes are washed out by fresh gas flow. Right: The speckled curve projects the full exhaled breath that would be seen if fresh gas flow was diverted. Note that the plateau would be higher than the actual curve by an unpredictable amount. The digital information derived from a curve like the one on the left is useless.

FIGURE 11-5 Common capnographic diagnoses: reduction in end-tidal carbon dioxide tension (Petco₂). There are many reasons for sudden reduction in Petco₂, some of which result in characteristic capnographic patterns. A, Abrupt reduction to zero (or nearly zero) typically indicates mechanical disruption, disconnection, accidental extubation, or plugged sampling line. B, Sudden reduction to a lower Petco₂ while preserving plateau and characteristics of a good trace indicate sudden increase in dead space ventilation as occurs with a pulmonary embolus (either thrombus or air). C, Exponential reduction to zero (CO₂ washout curve) is characteristic of no pulmonary blood flow and thus either massive embolus or cardiac arrest.

FIGURE 11-6 Common capnographic diagnoses: irregular tracings. Irregularities in the curve are common, especially at the end of exhalation when exhaled gas flow is lowest. A, Diaphragmatic activity indicating spontaneous respiratory effort, usually the result of dissipating neuromuscular blockade. B, Cardiac oscillations: fluctuations in intrathoracic gas volume as a result of cardiac activity, usually a benign finding.

Before 1998, capnography was considered a standard monitor by the ASA for confirming the initial placement and continuous presence of an endotracheal tube. This section of the ASA monitoring standards was updated in 1998 and states that capnography should be used to confirm adequate ventilation during general anesthesia with or without an endotracheal tube (during laryngeal mask airway, face mask, or natural-airway anesthesia). Specifically, these guidelines state, “Continual monitoring for the presence of expired carbon dioxide shall be performed unless invalidated by the nature of the patient, procedure or equipment…. Continual end-tidal carbon dioxide analysis, in use from the time of endotracheal tube/laryngeal mask placement, until extubation/removal or initiating transfer to a postoperative care location, shall be performed using a quantitative method such as capnography, capnometry or mass spectroscopy” (ASA, 2003).

Most capnometers use the principle of infrared light absorption by sampling circuit gas in either a mainstream or a sidestream fashion. Sidestream analyzers aspirate a sample from the circuit and transport it via a long, narrow-bore tube to a distant analyzing chamber. Advantages include a lightweight airway adapter and the remote location of the delicate components of the analyzing chamber. Disadvantages of sidestream systems include potential occlusion of the sampling tube, distortion or dilution of the exhaled gas wave during aspiration and transport to the analyzing chamber, and the delay necessary to transport and analyze the sample. Innovations in capnography technology have allowed a sampling rate as low as 30 mL/min (i.e., microstream technology).

Mainstream analyzers use a sample chamber placed directly into the circuit. They have the advantage of providing virtually instantaneous analysis by avoiding transport of the sample. Such a system necessitates the addition of a delicate and bulky sensor to the proximal airway connection, where it might easily serve as a fixation point to dislodge a small tracheal tube. Solid-state innovations have dramatically reduced the weight of the mainstream sensors, but they remain significantly more hazardous when added to the circuits of neonates and small infants. Although early mainstream sample chambers added as much as 17 mL of dead space to the circuit, currently available models reduce this volume to 2 mL or less.

Capnography in pediatric anesthesia is used to confirm placement of an endotracheal tube in the correct tracheal position and to continuously assess the adequacy of ventilation. Capnography also provides information about the respiratory rate, breathing pattern, endotracheal tube patency, and, indirectly, degree of neuromuscular blockade. Capnography can assist with the diagnosis of metabolic and cardiovascular events and can provide an early warning of a faulty anesthesia delivery system. In pediatric patients, an abnormal increase in end-tidal carbon dioxide tension (Petco₂) most commonly signifies hypoventilation, but, rarely, it also indicates the presence of increased CO₂ production, as occurs with temperature elevation or as an early sign of malignant hyperthermia. On the other hand, an abnormally low Petco₂ may indicate an increase in dead space or suggest a state of low pulmonary perfusion. Sudden absence of the capnographic tracing indicates a breathing circuit disconnection, and the abnormal presence of inspired CO₂ signifies the presence of a faulty unidirectional valve, an exhausted CO₂ absorber, or, when a semiopen circuit is being used, rebreathing secondary to an insufficient fresh gas flow.

The capnographic tracing of small infants is often characterized by the lack of an apparent alveolar plateau. This is usually a result of a higher respiratory rate, an excessively high sampling flow for the volume of CO₂ produced, excessive dead space in the breathing circuit, or an excessive leak around an uncuffed endotracheal tube.

The degree to which Petco₂ reflects Paco₂ is subject to many variables, some technical and others physiologic. The technical issues of primary importance in the accurate measurement of mean alveolar CO₂ tension include the volume and flow rate of exhaled gas, the aspirating flow rate (for sidestream analyzers), the fresh gas flow rate, the type of breathing circuit, and the circuit location of the sampling chamber (mainstream analyzers) or lumen of the aspirating tubing (sidestream analyzers). These variables are of particular importance in the small neonate whose small exhaled volumes at low flow rates are often diluted by high fresh gas flows or aspirating gas flows (Badgwell et al., 1987, 1993; Rich et al., 1990; Spahr-Schopfer et al., 1993). Badgwell and others (1987) demonstrated exponential increases in the discrepancy between Petco₂ and Paco₂ values with progressive reduction in patients who weighed less than 12 kg (Fig. 11-7). The coaxial distal sampling tube that they advocated dramatically improved the correlation.

FIGURE 11-7 Gradient between end-tidal carbon dioxide (Etco₂) determinations made in the proximal and distal ends of a tracheal tube. Exponential increases in the gradient suggest substantial potential inaccuracy in proximal Etco₂ determinations for children weighing less than 12 kg.

(From Badgwell JM, McLeod ME, Lerman J, et al: End-tidal Pco2 measurements sampled at the distal and proximal ends of the endotracheal tube in infants and children, Anesth Analg 66:959, 1987.)

The physiologic variable that introduces the most significant error in Petco₂ is dead space ventilation (Swedlow, 1986). Apart from children with severe pulmonary pathology or acute events such as pulmonary embolus, the most prevalent pediatric population in whom substantial dead space ventilation occurs are those with cyanotic congenital heart disease, particularly right-to-left shunts (Burrows, 1989; Fletcher, 1991).

Pulse Oximetry

Pulse oximetry, invented by Takuo Aoyagi in 1974 and developed as a clinical monitor a decade later, has markedly improved patient safety, especially during the perioperative period (Yelderman and New, 1983; Aoyagi, 2003; Severinghaus, 2007). It provides an estimate of the oxyhemoglobin saturation in arterial blood with a probe on a fingertip (New, 1985). The pulse oximeter uses plethysmography to determine the systolic portion of the cardiac cycle. During systole, there is a greater volume of blood in a pulsatile arterial vascular bed. This vascular bed is positioned between a sensor that contains a two-wavelength (red and infrared, 660 and 940 nm) light-emitting diode and a photodiode receptor (the wavelength varies depending on the manufacturer). Less light is transmitted in systole than in diastole because of the increased volume of the arterial bed. Sophisticated algorithms based on the amount of light absorbed differentiate systole from diastole. Oxygenated and deoxygenated blood absorb different quantities of light, proportional to their concentrations or the percentage of saturation according to the Beer-Lambert law. Once systole is identified through plethysmography, the arterial saturation during this period is determined by using the ratio of light absorption at the two different wavelengths through this vascular bed. The ratio is matched to data acquired over a range of experimentally determined saturations and ratios of light absorption stored in the instrument’s memory to determine the arterial saturation. Using the ratio of light absorption at two wavelengths makes unnecessary the calibration and zeroing to adjust to the size of the patient or to skin pigment (Aoyagi, 2003). In the 80% to 100% saturation range, arterial saturation values determined by this method correlated well with in vitro measurements (New, 1985). Based on the algorithm used to determine systole and the time-averaging process used over several cardiac cycles, different pulse oximeters have slightly different responses to a variety of clinical situations. Because the device must identify the pulse-added absorption, it may confuse motion of the extremity to which the sensor is attached with pulsating motion and abort the display of saturation data or, worse, display inaccurate data (see later).

Pulse oximetry was introduced into pediatric practice in the United States in the late 1980s (Salyer, 2003). It serves as an early warning signal of impending or actual hypoxemia, often before the onset of cyanosis, frequently reminding anesthesiologists of the alarming rapidity with which infants develop hypoxemia as compared with toddlers and older children (Xue et al., 1996). Continuous-use pulse oximetry is included in the basic monitoring standards of the ASA (2003). Anesthesiologist-blinded studies have demonstrated that the use of pulse oximetry facilitates earlier recognition and fewer episodes of hypoxemia (Coté et al., 1988, 1991). Pulse oximetry has rapidly evolved into a standard monitor during pediatric anesthesia and, as a result, has never been subjected to rigorous outcome studies with a true control group (an anesthetic without a pulse oximeter) (Cohen et al., 1988). There are no outcome studies that demonstrate proven benefit from the use of pulse oximetry (Moller et al., 1993a, 1993b; Pedersen et al., 2001).

A number of well-described limitations of pulse oximetry affect the accuracy of commercially available pulse oximeters. Among other factors, ambient lighting conditions, motion, peripheral circulation to the extremity, body temperature, and abnormal hemoglobins can interfere with accuracy and consistency of pulse oximeter performance (Schramm et al., 1997; Trivedi et al., 1997). Of all these factors, motion artifact is clinically the most common problem, resulting in inaccurate readings, data loss, missed alarms, and false alarms. Patient motion (e.g., shivering and seizures in adults and children, as well as kicking, stretching, and crying in infants) causes movement of venous blood, as well as other normally nonpulsatile body fluid components (e.g., tissue fluid in edematous patients), along with the arterial blood. The pulsatile components of the body fluids other than arterial blood may lead to falsely low saturation readings. If motion is combined with low perfusion at the sensor site, then the venous blood makes a more significant contribution to the pulsatile component and drives the Sao₂ as measured with a pulse oximeter (or SpO₂) even lower. Various algorithm-based motion rejection methods used by manufacturers to minimize the impact of motion artifact and reduce false alarms include averaging saturation data over longer periods of time, freezing values until uncorrupted signals are present, and implementing alarm delays (Petterson et al., 2007). Other technologies use parallel processing of multiple sophisticated algorithms to isolate the biological signal in the presence of noise. When compared with conventional and other motion-tolerant pulse oximeters, Signal Extraction Technology (Masimo Corp., Irvine, Calif) has been reported to have a better performance in a variety of pediatric clinical settings, including newborn and pediatric intensive care units, the PACU, and the pediatric sleep lab (Ahlborn et al., 2000; Bohnhorst et al., 2000; Malviya et al., 2000; Barker, 2002; Brouillette et al., 2002; Trang et al., 2004; Workie et al., 2005).

Although the pulse oximeter is a continuous monitor, it does not instantaneously reflect the arterial saturation or the degree of desaturation. Most commercially available pulse oximeters display the average SpO₂ numbers of the most recent 5 to 15 pulses, depending on models and manufacturers, causing a delay in detecting desaturation. When the patient is breathing high concentrations of oxygen and the blood is fully saturated, a substantial decrease in Pao₂ can occur without a change in Sao₂. Reynolds and associates (1993) detected desaturation in children 30 seconds earlier in probes placed centrally (forehead) than in those placed on a finger. By the time the value indicated by a peripheral sensor had decreased 5%, the value indicated on a central sensor was 30% to 40% lower (Agashe et al., 2006). The precise mechanism for this discrepancy remains unknown, although Severinghaus and Naifeh (1987) postulated that it reflects peripheral blood transit, capillary composition, and oxygen utilization. Centrally placed reflective forehead pulse oximeter sensors are reported to respond faster and to be less likely to be compromised by vasoconstriction than transmittance finger probes (Choi et al., 2010).

Pulse oximeters are designed to warn practitioners when the arterial saturation decreases to below normal, not to serve as quantitative devices in hypoxemic patients. Compared with measured arterial saturation in children with cyanotic congenital heart disease, most pulse oximeters exhibit a progressively positive bias, typically reaching 5% to 15% at an Sao₂ of 60%, in addition to a significant reduction in precision (±8% to 10%) (Gidding, 1992; Schmitt et al., 1993) (Fig. 11-8). More recently, a pulse oximetry sensor designed specifically for patients who are chronically hypoxemic, such as children with congenital heart disease, has become available. This specialized sensor, called the Blue sensor, has reported precision of less than ± 4% between saturations of 60% and 80%. When compared with a standard pulse oximetry sensor on postoperative palliative cardiac surgery patients, the Blue sensor had a bias of 0.21 and an accuracy root-mean-square (RMS) error of 3.6% compared with Sao₂ values, whereas the standard sensor had a bias of 0.86 and an accuracy RMS error of 6.4% compared with Sao₂ values (Cannesson et al., 2008c). The Blue sensor demonstrates considerable improvement in the accuracy of pulse oximetry monitoring of cyanotic children.

FIGURE 11-8 Accuracy and precision of pulse oximetry (Nellcor N-100) in chronically hypoxemic children with congenital heart malformations. Data comparing pulse oximetry (SpO₂) with CO-Oximetry (Sao₂) reflect a 5.8% mean positive bias for SpO₂, with wide discrepancies between the two techniques (±2 SD = –3.8% to +15.4%).

(From Schmitt HJ, Schuetz WH, Proeschel PA, et al: Accuracy of pulse oximetry in children with cyanotic congenital heart disease, J Cardiothorac Vasc Anesth 7:61, 1993.)

Interference with the expected spectrophotometric absorption pattern also causes errors in measurement. Low hemoglobin (<5 g/dL) and abnormal hemoglobin species (e.g., methemoglobin, carboxyhemoglobin) cause inaccurate saturation estimates by pulse oximeters, which become progressively inaccurate as Sao₂ decreases (New, 1985; Barker and Tremper, 1987; Barker et al., 1989; Watcha et al., 1989). In contrast, unusual or abnormal hemoglobin molecules, such as fetal hemoglobin and sickle cell hemoglobin, apparently have little effect on the saturation measurement (Jennis and Peabody, 1987; Ortiz et al., 1999). Intravenous dyes such as methylene blue and indocyanine green, affect the expected light absorption and produce spurious information (Ralston et al., 1991). Aberrant radiation (e.g., electromagnetic energy from the electrocautery, infrared heat lamps, operating room lights) and infrared light pulses from lasers and neuronavigation systems also cause incorrect saturation determinations (Brooks et al., 1984; Costarino et al., 1987; Hanowell et al., 1987; Mathes et al., 2008; Zhang et al., 2009).

Multiple-Wavelength Pulse Oximetry

Pulse oximeters that use two wavelengths of light are subject to interference from other hemoglobin species that absorb light across similar spectra. Dyshemoglobins such as carboxyhemoglobin and methemoglobin cannot be detected by pulse oximetry, but their presence in arterial blood can bias the oximeter’s estimate of Sao₂. Multiple-wavelength pulse oximetry (Pulse CO-Oximetry; Masimo Corp., Irvine, CA) has been developed to detect and measure both carboxyhemoglobin and methemoglobin levels noninvasively and continuously, although the accuracy of Sao₂ estimations from these devices is still compromised (Barker et al., 2006). Pulse CO-Oximetry devices that measure carboxyhemoglobin have been shown to be useful in the emergency room for detection of occult carbon monoxide poisoning, and for monitoring and guiding emergency treatment of CO exposure (Suner et al., 2008; Bledsoe et al., 2009; Piatkowski et al., 2009). Methemoglobinemia can be caused by a wide variety of oxidizing drugs and their metabolites and by exposure to industrial chemicals such as aniline dyes, chlorates, and bromates. The clinical usefulness of noninvasive methemoglobin detection and monitoring by Pulse CO-Oximetry has been demonstrated for the diagnosis and treatment of severe methemoglobinemia in the hospital setting (Annabi and Barker, 2009).

Plethysmographic Waveform Analysis

For many years, clinicians have used the shape and amplitude of the plethysmographic waveform from pulse oximetry as a crude and imprecise method to assess blood volume and cardiac output (Partridge, 1987). Respiratory variations in pulse oximetry plethysmographic waveform amplitude can predict fluid responsiveness noninvasively in mechanically ventilated patients, but this method of assessment is technically challenging—neither automated nor continuous, and not available in clinical practice (Cannesson et al., 2007). An index that provides an automatic and continuous calculation of the respiratory variations in the waveform amplitude is now available with some pulse oximeters. The pleth variability index has been shown to predict fluid responsiveness during cardiac surgery (Cannesson et al., 2008b) and to predict changes in cardiac index with changes in positive end-expiratory pressure after cardiac surgery in mechanically ventilated, sedated patients (Desebbe et al. 2010). This index may allow the optimization of fluid loading noninvasively during and after surgery, but further validation is required before these results can be extrapolated to the general population.

Near-Infrared Spectroscopy

Near-infrared spectroscopy (NIRS) is another tool available to the anesthesiologist that provides information regarding tissue oxygenation. Its applications range from cerebral oximetry during circulatory arrest, to monitoring for compartment syndrome in the setting of trauma (Tobias and Hoernschemeyer, 2007; Bernal et al., 2010). Although NIRS continues to be studied for its impact on patient outcomes, it is increasing popular as a noninvasive monitor of vital data.

NIRS relies on principles similar to those of pulse oximetry, exploiting the variable absorption of light by color-containing compounds in the body. Infrared light emitted by the device enters the skin and scatters through the tissues beneath, some energy being absorbed and some reflected. The portion of the light that returns to the skin is analyzed and conclusions are drawn about the content of the tissues it passed through (Cohn, 2007).

The basis for analysis of the reflected light is found in the Beer-Lambert law. According to this law, reduction in the intensity of light passing through a substance depends on the absorbance of the substance and its thickness. These principles are used to quantify concentrations of oxygenated hemoglobin, which absorbs light at 805 nm, and deoxygenated hemoglobin, which absorbs at 730 nm, and to determine the regional oxygen saturation (rSo₂) of the tissue examined (Tobias, 2006).

Unlike pulse oximetry, NIRS does not require pulsatile flow, because it looks at all the blood in a given tissue including arterial, venous, and capillary blood. This means that NIRS is particularly valuable when pulsatility is lost (e.g., in CPB or circulatory arrest).

NIRS technology is commercially available from only a few companies. The INVOS Cerebral/Somatic Oximeter (Somanetics Corp., Troy, MI), which has been on the market for a long time, provides real-time monitoring capability for the rSo₂ of blood in tissues beneath the probe. To target specific tissues such as the brain or kidneys and not the skin and soft tissues between the probe and these organs, the device has been designed to eliminate data from the superficial tissues. The probe incorporates a light-emitting diode and two sensors located at fixed distances from the light. Infrared light passing through the superficial tissue is picked up by the closest sensor, and infrared light passing through the deeper tissues is picked up by the further sensor. Information from the superficial tissues is subtracted from the data returning from deeper tissues to provide accurate estimates of the deep tissue oxygenation (Tobias, 2006).

The Fore-Sight (Cass Medical Systems, Branford, CT) uses laser light at four wavelengths. Based on the same principles as the Sonametics device, it also accounts for the spurious information collected from the superficial tissues to focus on the deeper areas of interest (Chakravarti et al., 2008). Both devices have approved uses in pediatrics, and probes of different sizes fit this population.

Clinicians presented with data on tissue oxygenation must know how to respond to it, and much has been written about the clinical usefulness of NIRS. Cerebral oximetry has been the most widely studied use of NIRS, and here NIRS has proved to be both a sensitive and specific indicator of cerebral tissue oxygenation (Al-Rawi et al., 2001). Kurth and coworkers (2002) noted thresholds for normal and abnormal values in a piglet model. They found baseline cerebral rSo₂ values of 68%, and they found that lactate began rising when the rSo₂ dropped to 44%, electroencephalographic changes started at around 42%, and ATP loss occurred at 33%. This information, along with corroborating findings in adult studies, has led to a weak consensus that an rSo₂ reduction of 20% from baseline, or an absolute value of 50%, serves as an indication of potential hypoxic injury and warrants intervention.

To establish the value of cerebral oximetry in the pediatric population, Hoffman and colleagues (2008) studied patients treated for hypoplastic left heart syndrome, and they demonstrated improved neurologic scores at age 4 to 5 years in patients who were monitored with cerebral oximetry at the time of their surgery and treated for abnormal values (rSo₂ < 55%). This is encouraging, but most agree that there is still insufficient evidence to support clinical decisions based on NIRS values alone (Hirsch et al., 2009).

NIRS has generated interest in additional applications, including shock, compartment syndrome, plastics, and regional perfusion. Adult literature demonstrates NIRS ability to detect compartment syndrome as well as improvement in tissue oxygenation after fasciotomy (Giannotti et al., 2000). Studies of free flaps found that NIRS identified early evidence of flap failure prior to routine clinical evidence (Holzle et al., 2006).

NIRS provides information regarding tissue oxygenation for a broad spectrum of clinical scenarios. As studies continue, clinicians will better understand how to interpret and respond to the data so as to improve the outcomes of their patients.

Bispectral Index

In 1996, the FDA approved the use of the bispectral index (BIS) monitor (Aspect Medical Systems, Nattick, MA), a device based on electroencephalography (EEG) and used to predict the relative level of hypnosis, or unconsciousness, in anesthetized patients (Rosow and Manberg, 1998). Using a patch affixed to the patient’s forehead, the BIS monitor integrates various EEG descriptors into a single, dimensionless, empirically calibrated number ranging from 0 to 100, where 0 represents electrical silence and 100 represents full wakefulness. A state of unconsciousness consistent with BIS values less than 60 usually ensures a lack of intraoperative recall (Glass et al., 1997). In adults, titration of anesthetics to a targeted BIS value between 40 and 60 results in the administration of relatively lower doses of anesthetics and earlier awakening (Gan et al., 1997). Studies in adults suggest that routine BIS monitoring is associated with reduced intraoperative awareness during high-risk surgical procedures (e.g., microlaryngeal surgery, cesarean section, cardiac bypass) (Myles et al., 2003).

In anesthetized children on mechanical ventilation, as well as in those spontaneously breathing through a face mask, BIS values are inversely proportional to the end-tidal concentration of halothane, sevoflurane, and desflurane (Denman et al., 2000; Davidson et al., 2001; Degoute et al., 2001; Tirel et al., 2006; Kern et al., 2007). Similar relationships between the depth of anesthesia (minimum alveolar concentration [MAC], plasma concentration) and BIS values have been studied during total intravenous anesthesia (TIVA), with or without additional opioids, in children with target-controlled infusion of propofol (Jeleazcov et al., 2007; Tirel et al., 2008; Rigouzzo et al., 2008; Malherbe et al., 2010). BIS values during sevoflurane anesthesia appear to be proportionately less in children with quadriplegic cerebral palsy and those who are mentally compromised (Choudhry and Brenn, 2002; Valkenburg et al., 2009).

In children, at a given depth of anesthesia (i.e., MAC), there are significant nonlinear inverse correlations between BIS values and age: BIS values are higher in toddlers than in older children in all studies, including TIVA with propofol. Furthermore, in addition to age-related variations in BIS values, interindividual variabilities in BIS index are much higher in infants and children than in adults (Tirel et al.; 2006, Kern et al., 2007; Rigouzzo et al., 2008). These characteristics of BIS in children may be associated with age-related differences in brain maturation and synapse formation throughout childhood (Watcha, 2001). The reliability of the BIS index is further diminished in infants younger than 1 year (Tirel et al., 2006; Jeleazcov et al., 2007; Kern et al., 2007; Rigouzzo et al., 2008).

The clinical benefits of measuring BIS and maintaining appropriate levels of anesthesia, such as reduced risk for intraoperative awareness and improved recovery time, may still be valid in pediatric patients, with the realization of certain characteristics of and differences in children. In adolescents undergoing scoliosis surgery, BIS can predict voluntary patient movement in response to commands during the intraoperative wake-up test (McCann et al., 2002). BIS monitoring was also used successfully for the intraoperative wake-up test in a neonate undergoing neurosurgery for the repair of myelomeningocele (Govindarajan et al., 2006). BIS monitoring in children aged 3 to 18 years who are undergoing tonsillectomy and adenoidectomy is associated with reduced recovery times (Bannister et al., 2001). However, in the same study, BIS monitoring did not affect recovery times in children younger than 3 years who were undergoing hernia repair.

More recently, interest in the efficacy of BIS in monitoring the depth of sedation has increased in other specialties, including critical care, emergency medicine, dentistry, and general pediatrics (Kerssens and Sebel, 2006; Malviya et al., 2006, 2007; Sadhasivam et al., 2006; Froom et al., 2008; Baygin et al., 2010). Also, BIS monitoring has been used in documenting positive sedative effects of regional and spinal blockade with local anesthetics in children, with or without general anesthesia (Davidson et al., 2006; Hermanns et al., 2006). Future studies are expected to further delineate the use of BIS in the pediatric population.

Neurophysiologic Monitoring

It has long been appreciated that the patient’s physiologic status is dynamic, and that rapid and life-threatening changes may occur during surgery. The comparative abilities to evaluate the functional status of the nervous system by clinical means and by commonly used physiologic monitoring tools that are available to anesthesiologists is limited. Routine anesthesia monitoring may reflect stress on the central nervous system (CNS). For example, changes in heart rate related to both brainstem and vagal stimulation provide invaluable information when correlated with surgical activity. However, intraoperative neurophysiologic monitoring (IOM) adds yet another dimension, as well as specificity, to assessment of the status of the patient during surgery and anesthesia.

Neurophysiologic techniques provide important and reliable alternative tools for assessment of function of the pediatric CNS (Sclabassi and Krieger, 1995). These techniques provide objective measures of the functioning of the CNS and can serve to localize, warn of, and document deterioration in neuronal function. Intraoperatively, continuous monitoring of the area of the CNS that is at risk from surgical and anesthetic manipulation provides immediate insight into the effects of these manipulations. Rather than waiting to evaluate the neurologic examination of a child in the postoperative period, continuous IOM provides an immediate view of the integrity of the CNS, permitting changes in operative and anesthetic technique to minimize or correct the deleterious effects of intraoperative manipulations. Advantages of these methods are that the results are objective and quantifiable, the site of the lesion can be identified, and clinically latent and evolving lesions can frequently be demonstrated. These techniques have roles to play both in the diagnostic investigation of pediatric CNS function and in the field of intraoperative assessment of CNS function.

The measures used must be both specific to the neural tissue being manipulated and sensitive to changes in the functioning of the neural tissue produced by the surgical manipulations. Monitoring of the electrical activity dependent on the functioning of the brainstem (brainstem auditory evoked potential [BAEP] and brainstem somatosensory evoked potential [BSEP]), the cortex (EEG, somatosensory evoked potential [SEP], and visual evoked potential [VEP]), the spinal cord (SEP, motor evoked potential [MEP], and electromyogram [EMG]), the various cranial nerves (EMG), and peripheral nerves (compound action potential [CAP] and EMG) provides a multidimensional assessment of the integrity of the neural structures at risk. In addition, many of these measures provide information not only about function itself but also about variables that directly or indirectly affect function, such as blood flow, hypoxia, and hypotension. The goal of IOM is to provide information to the surgeon and anesthesiologist that allows them to modify their operative strategy before inducing additional deficits in the functioning of the CNS.

Neurophysiologic Measures

Neurophysiologic measures that are routinely used provide a functional map of much of the entire neuroaxis. These include the EEG, an unstimulated measure of cortical function suitable for providing the following information:

Maturational Effects

The functional assessment of the pediatric CNS presents difficult and unique problems. The pediatric CNS differs from the adult CNS in that it is maturing over the first several years of life; that is, the neural tissue, the myelin coating of the axonal processes, and the vascular supply to the CNS all show significant changes. These developmental anatomic changes are reflected in maturational functional changes as measured by ascending SEP and by descending MEP activity, VEPs, and BAEPs (Starr et al., 1977; Cracco et al., 1979; Guthkelch et al., 1982). A number of factors contribute to the maturational changes of evoked potentials, and the use of age- and size-matched normal controls is essential. For intraoperative monitoring purposes, infants act as their own controls. Central and peripheral myelination is believed to be completed by 5 years of age, and from then until maturity, the dominant factor affecting SEP latency is height (Yakovlev and Lecours, 1967; Gilmore et al., 1985) (Fig. 11-9).

FIGURE 11-9 Maturational changes in cortical somatosensory median nerve evoked potentials (MSPs) (A), flash visual evoked potentials (VEPs) (B), and brainstem auditory evoked potentials (BAEPs) (C), in early infancy. Note the decreasing latency (the time measured from the initiation of the stimulus to the point of maximum amplitude of the evoked potential) and enhancing morphology for the identified waves in all three modalities.

Electroencephalography

The functioning of the cerebral cortex is extremely sensitive to changes in arterial oxygenation and insufficient cerebral blood flow or an inadequate partial pressure of oxygen; this sensitivity is rapidly reflected in the EEG (Meyer and Marx, 1972). Oxidative metabolism supplies the energy for maintenance of the membrane potential of nerve cells, and the EEG depends directly on the transmembrane potentials of neurons, reflecting disturbances of cerebral metabolism such as hypoxia. Some factors that may contribute to ischemic events in surgical patients are decreased oxygen-carrying capacity resulting from hypovolemia, and decreased cerebral perfusion pressure resulting from factors associated with decreased systemic arterial pressure, increased intracranial pressure, and mechanical obstruction of cerebral vessels (Freye, 1990). It is thought that having two channels of continuous EEG monitoring is adequate, because the problems are related more to global or hemispheric effects than to precise focality. The EEG can be observed both as the ongoing unprocessed signal and in a Fourier-transformed representation. The electroencephalographic appearances of any ischemic or hypoxic events are similar, and differentiation between the various putative causative factors is made by being particularly attentive to the clinical situation. For example, blood pressure, ECG, oxygen saturation, administered drugs, and surgical manipulations may all have an observable effect. Other concurrent factors that may alter the EEG are changes in the depth of anesthesia, temperature changes, and changes in CO₂ content. These factors can be recognized by their relatively slow onset, lasting for several minutes, in contrast to the changes of ischemia, which generally occur within seconds. In some situations, the EEG may be acutely depressed on injection of an anesthetic that rapidly passes the blood-brain barrier. Such situations may be found with the use of high-dosage opioid anesthesia, in which fentanyl induces an immediate and marked reduction in fast-frequency activity in the EEG, with an increase in low-frequency, high-amplitude activity in the delta range (Freye, 1990).

A simplified but useful summary of possible changes includes the following:

Somatosensory and Motor Evoked Potentials

The neurophysiologic measures of value in assessing the spinal cord consist of SEPs, produced by stimulating various peripheral nerves, and MEPs, which may be observed as either compound muscle or nerve action potentials and which may be produced by either transcranial or spinal cord stimulation. It is advantageous to think of the SEPs as characterizing the ascending activity in the spinal cord and the MEPs as characterizing the descending activity in the spinal cord. This distinction, although not particularly important with respect to the sensory activity, is potentially extremely important with respect to the descending activity, because important questions remain as to what pathways are being stimulated (Rose, 1994), these distinctions are not as absolute as we would like to think.

Somatosensory Evoked Potentials (Ascending Activity)

SEPs depend on the stimulation of the large afferent fibers of peripheral nerves. After stimulation of peripheral nerves in the arms or the legs, SEPs can be reproducibly recorded over the spine and scalp. In the spinal cord, the SEPs are conducted primarily through the dorsal columns.

SEPs are a sequence of potentials generated in the peripheral nerves, dorsal horn nuclei, dorsal column pathways, and dorsal column nuclei of the spinal cord; the medial lemniscal pathways of the brainstem; and the thalamus and thalamocortical and parietal regions of the brain after the application of a transient electrical stimulus to a peripheral nerve (Sclabassi et al., 1993b). When recorded from electrodes on the surface of the body, the potentials of interest are very small, ranging in size from 2 to 5 mcV, and occur in approximately the first 100 msec after the application of a stimulus (often referred to as early and middle latency potentials). Evoked potentials are described in terms of latency and amplitude. Latency is the time measured from the application of a stimulus to the point of maximum amplitude of the evoked potential. Some types of SEP have more than one peak, and the time between peaks is the interpeak latency. The amplitude is the voltage difference between two peaks of opposite polarity, or a reference potential. Measurements of latencies, amplitudes, and interpeak latencies characterize SEP recordings. Changes in these measurements during a surgical procedure may represent injury to the neural tissue between the stimulus generator and the recording electrode.

In all cases, the stimuli are electrical impulses applied transcutaneously, at a rate of 0.7 to 5.3 Hz, depending on the robustness of the response, which is typically a function of the patient’s age and pathology. Typically, responses to 128 stimuli are averaged; in many cases, as few as 12 responses may be averaged, providing near real-time updating of the responses.

All types of SEPs are used for intraoperative monitoring, primarily during spinal, cortical, and posterior fossa surgery. Potentials can be recorded after stimulation of the median nerve at the wrist, the common peroneal nerve, the posterior tibial nerve, and the dorsal nerve of the penis and the clitoris. Multiple types of responses from different stimuli and different sources are often simultaneously recorded, allowing the entire neuroaxis to be monitored. Monitoring multiple upper and lower extremity responses simultaneously during spinal surgery allows cord injury to be distinguished from global problems.

Median and Ulnar Nerve Evoked Potentials

The median (MSPs) and ulnar (USPs) nerve evoked potentials are all useful in assessing the brachial plexus, upper spinal cord, brainstem, and telencephalon. One important distinction is that the USPs provide information rostral to T1, whereas the MSPs provide information rostral to C6 (Fig. 11-10). At Erb’s point, the response consists of an apparently triphasic (positive-negative-positive) nerve action potential, reflecting the passage of the mixed nerve volley through the brachial plexus. This component is usually labeled N11 for the large negative-going component. At the cervical C7 recording site, the main component is a negative peak occurring at 14-msec latency, N14, with an associated complex structure. It has been postulated that these waves are generated in the dorsal roots, dorsal horn, posterior columns, and structures of the lower brainstem. During spinal fusions, monitoring of the brachial plexus may also alert the surgeon and anesthesiologist to compressive positioning of the arms.

FIGURE 11-10 Median nerve evoked potentials (MSPs) produced by right median nerve stimulation (MD) stimulation. Data are recorded from Erb’s point on the right referenced to Erb’s point on the left (bottom trace); cervical C7, cervical C2 and C3 (left parietal cortex), all referenced to F_z. Note the increase in latency of the large negative wave first identified as N11 at Erb’s point, as the activity projects afferently.

Common Peroneal and Posterior Tibial Nerve Evoked Potentials

In the lower limb, nerves used to elicit cortical SEPs include the tibial, peroneal, and femoral. Spinal potentials are most consistently obtained through stimulation of the tibial nerve at the medial malleolus or peroneal nerve in the popliteal fossa.

SEPs recorded over the spine reflect the afferent volley traversing the dorsal columns. These responses can be recorded from electrodes attached to the skin over the spine, and they progressively increase in latency at more rostral recording locations. Spinal SEPs are relatively easy to obtain in children, with the amplitude and definition of the waves decreasing with increasing age, so that by the mid-teenage years, these responses are more difficult to obtain, as is the case with adults. More rostral recording locations reflect potentials arising in multiple ascending pathways, including the dorsal and dorsolateral columns, which lie primarily ipsilateral to the side of stimulation.

In our experience, SEPs are extremely sensitive and specific to spinal cord injury, whether it occurs in the dorsal or the ventral pathway. This is confirmed in the literature (Nuwer et al., 1995), where a false-negative rate of 0.063% was found for 51,263 spinal cases in which SEPs were the only potentials monitored. Furthermore, the negative predictive value (i.e., the likelihood of normal spinal cord function in the presence of stable SEPs) was 99.93%. This is a significant improvement over the 0.72% to 1.4% incidence of spinal cord injury reported for unmonitored cases (MacEwen et al., 1975).

Dermatomal Evoked Potentials

A disadvantage of SEPs produced by stimulation of large nerve trunks is that input to the spinal cord usually occurs over more than one level. This problem can be addressed by delivering the stimulus to a small cutaneous nerve that is believed to derive from a single dorsal root, or to the signature area of a particular dermatome. Significant disagreement exists concerning the cutaneous distributions of dermatomes, and care should be taken to stimulate the commonly accepted receptive fields of a root.

Pudendal nerve responses, a special type of dermatomal response, are particularly useful, especially in patients with spina bifida. The pudendal nerve carries sensory fibers from the penis, urethra, anus, and pelvic floor muscles and supplies motor innervation to the bulbocavernosus and pelvic floor muscles, the external urethral sphincter, and the external anal sphincter. Cortical responses to electrical stimulation of the dorsal nerve of the penis, the urethra, and the urinary bladder have all been described (Badr et al., 1982; Haldeman et al., 1982). Pudendal nerve responses are of similar morphology to the tibial nerve SEP (TSP) and are best recorded from the same area of the scalp (Fig. 11-11).

FIGURE 11-11 Pudendal nerve responses obtained from a male patient with tethered cord and symptoms referable to the pudendal nerve. All responses are recorded from P_z referenced to F_z. A, Responses obtained by stimulating the right branch of the dorsal nerve of the penis. B, Responses obtained by stimulating the left branch. Note the significant reduction in amplitude in response to the left-sided stimulation. C, Control data recorded during every procedure.

Motor Evoked Potentials (Descending Activity)

Because SEPs reflect function in the dorsal columns of the spinal cord, they do not directly assess the integrity of descending spinal motor tracts. It is possible to have focal damage to the motor areas in the spinal cord in which the SEPs remain normal. Accordingly, misleading results have occasionally been obtained when using SEPs alone for intraoperative monitoring and diagnosis (Lesser et al., 1986), but this is rare (Nuwer et al., 1995). MEPs, which may be either evoked EMGs or compound action potentials, can be used to test the integrity of the motor pathways through either electrical or magnetic stimulation (Merton and Morton, 1980; Barker et al., 1985). MEPs can be obtained via stimulation of the motor areas of the brain or spinal cord through the intact skin, direct stimulation of exposed neuronal tissue, or direct root stimulation (e.g., during the release of a tethered cord), and recording of a stimulus-related response either as a compound muscle action potential (CMAP) or as an efferent compound nerve action potential (CNAP) (Fig. 11-12).

FIGURE 11-12 A, Compound muscle action potentials (CMAPs). B, Compound nerve action potentials (CNAPs). Both muscle and nerve action potentials were produced by transcranial magnetic stimulation at scalp positions C₄, P₂, and C₂. The CMAPs were recorded from the abductor pollicis brevis muscle, and the CNAPs were recorded from the left median nerve at the wrist.

Transcranial electrical stimulation may be used to elicit motor responses. There is no general consensus about the location of recording electrodes, outside of specific muscle groups for evoked CMAPs or over the obvious peripheral nerves for evoked CNAPs, nor is there a general consensus concerning which class of these activities is more advantageous to record. CNAPs allow the patient to receive neuromuscular blockade agents. One of the most important stimulation parameters for eliciting reliable MEPs is the interstimulus interval (ISI) of a burst of stimuli (Taylor et al., 1993). It has been found that a burst of stimuli with ISIs between 2 and 5 msec (500 to 200 Hz) produce a maximal response by overcoming the depressed effect of general anesthesia (mentioned earlier) (Kalkman et al., 1995). The significant parameters and morphologic features of MEPs are response threshold, onset latency, central conduction time, and response size (Fig. 11-13).

FIGURE 11-13 Motor evoked potentials (MEPs) obtained by transcranial electrical stimulation by needle electrode at C₃ and C₄ during stabilization of a cervical fracture. Channels 7 and 8 (columns 1 and 2) are recorded from the left abductor pollicis brevis and gastrocnemius muscles from stimulation of the right brain (anode at C₄), while channels 9 and 10 (columns 3 and 4) are recorded from the same muscle groups on the right side when the left brain is stimulated (anode at C₃). In many younger infants, both sides of the brain can be stimulated simultaneously. Note the 250-Hz burst of stimulus activity at the beginning of each trace.

Brainstem Auditory Evoked Potentials

Monitoring of the function of cranial nerve VIII through the use of BAEPs assists in preserving hearing, locating cranial nerve VIII, or determining whether the overall function of the brainstem is altered. BAEPs are also sensitive to retraction on the frontal poles, most likely because of force transmission to the brainstem.

The classic BAEP consists of a minimum of five and a maximum of seven peaks. All occur with 10 msec of a brief click or tone presentation. Wave I is generated in the auditory portion of nerve VIII. Wave II is generated bilaterally at or in the proximity of the cochlear nucleus. Wave III is generated bilaterally from the lower pons near the superior olive and trapezoid body. Waves IV and V are probably generated in the upper pons or lower midbrain, near the lateral lemniscus or possibly near the inferior colliculus.

Waves I through V are relatively resistant to sedative medication and general anesthetics, but this places no constraints on the anesthesiologist. These waves are sensitive to temperature changes, with absolute and interpeak latencies increasing by approximately 0.20 msec. The latency of wave V is the primary concern in intraoperative monitoring of the BAEPs, because this is the most robust and easily identifiable of the waves in this response.

Visual System

VEPs are used to aid in determining the functional integrity of the visual system, primarily in the region of the optic nerves, chiasm, and optic radiations (Albright and Sclabassi, 1985). The recorded activity is generated either at the retina (electroretinogram) or at the cortex.

Stimulation of the visual system using a bright flash is not recommended for diagnostic purposes because of intersubject variability (Ciganek, 1961), except in select situations. In the operating room, this is a very helpful and effective technique (Fig. 11-14).

FIGURE 11-14 Intraoperative flash visual evoked potentials (VEPs), obtained between midline occipital (O_z) and vertex (C_z) electrodes. Data were from a 10-month-old girl who was operated on for a chiasmal glioma, which was 90% removed. Top: Response before resection. Bottom: Response after resection. The responses were obtained continuously during the procedure, were abnormal, and were highly variable from response to response, but they were unchanged during the procedure.

(From Albright AL, Sclabassi RJ: Cavitron ultrasonic surgical aspirator and visual evoked potential monitoring for chiasmal gliomas in children, J Neurosurg 63:138, 1985.)

Electromyography

The EMG is electrical activity produced in muscle fibers below the skin; it has a frequency content ranging from 15 to 150 Hz. Three types of electrodes are used to record the EMGs: fine wire electrodes, which have the highest impedance and the narrowest field of view; subdermal needles, which have an intermediate impedance and a larger field of view; and disk surface electrodes, which have the lowest impedance and the largest field of view (field of view means the integrated level of electrical activity). The recording techniques are essentially the same for all cranial nerves and all muscle groups. Subdermal platinum needle electrodes are used in bipolar recording configurations; that is, all recordings are made between a pair of electrodes inserted into the same muscle group. Bipolar recordings are used to minimize confusion regarding which cranial nerve or branch of a cranial nerve is producing the observed EMG. The electrodes are normally placed before the start of the procedure; occasionally, electrodes are placed in a sterile field by the surgeons.

These signals are listened to continuously for evaluation of nerve function both by the neurophysiologists and by the surgeons. Four categories of EMG activity are observed:

It is important to note that a sharply cut nerve may produce only a brief burst of activity; monitoring cannot be expected to replace extreme caution when working near the cranial nerves.

EMG is either spontaneous (e.g., anal sphincter activity produced by irritation of S3 to S5 roots during an untethering procedure involving the lower portion of the cauda equina) or evoked, of the type produced in selective rhyzotomy for the treatment of spasticity, for placement of pedicle screws, or for ongoing evaluation of the spinal cord. Evoked EMGs have a considerably larger amplitude (>1 mV) than sensory evoked potentials (<0.5 mcV) and therefore do not require averaging to extract them from the background noise.

Cranial Nerve Function

Cranial nerve function is monitored continuously during surgery for two reasons: first, to establish the location and orientation of the cranial nerves in the operative field, and second, to preserve functioning in the cranial nerves and their related brainstem nuclei (Sclabassi et al., 1993a). The major observed variables are the EMGs from the appropriate muscle group innervated by the cranial nerves of interest. In general, the cranial nerves ipsilateral to the operative side are monitored; when appropriate, bilateral activity is monitored.

In addition to monitoring the ongoing EMG activity related to the various cranial nerves, these cranial nerves may also be electrically stimulated. This is usually done to determine the location of the nerve in the operative field (because often the nerve is enveloped by tumor and may not be directly observable), or to determine the functional integrity of the nerve (Daube and Harper, 1989). The most common example of this procedure is the direct stimulation of nerve VII. The return path for the stimulating current is provided by a metal electrode inserted into the adjacent muscle mass. In some situations, when very precise localization of the nerve is required, bipolar stimulating electrodes are used. Most of the time, the question being asked is, Is the nerve there? These techniques are very useful when monitoring during surgery for posterior fossa tumors that extend to the floor of the fourth ventricle.