Extensive physiologic monitoring is employed during cardiac operations because virtually every major physiologic system required for life is affected. The reasons for physiologic monitoring are (1) to ensure patient safety in the absence of protective reflexes made ineffective by anesthetic drugs, (2) to enable pharmacologic and mechanical control of vital function, and (3) to diagnose acute emergencies that require immediate treatment. For example, morbidity as a consequence of breathing circuit disconnects, loss of oxygen from the hospital's central supply, or unrecognized esophageal or main stem endotracheal intubations can be prevented by capnography, pulse oximetry, airway pressure monitors, oxygen analyzers, and a stethoscope.
The senses of touch, hearing, and sight are the basic monitors. Electronic monitors are vigilance aids that supplement the anesthesiologist's perceptions. The selection of monitors is dictated by the utility of the data generated, expense, and risk. Routine or essential monitors that have been deemed cost-effective with low risk:benefit ratios include pulse oximetry, noninvasive blood pressure, capnography, temperature, ECG, precordial or esophageal stethoscope, and oxygen analyzers. These have been defined by the American Society of Anesthesiologists (House of Delegates, 1989) as essential monitors to be used in all surgical patients requiring anesthesia unless there are contraindications (eg, esophageal stethoscope during esophageal surgery) (Table 10-3). Other noninvasive and invasive monitors are used only with clear indication.
The growth in monitoring technology and sophistication is paralleled by an equal growth in cost. The balance between cost and enhancement of patient safety must be considered when additional monitoring is selected. It is difficult to justify a monitor that provides data that does not influence medical or surgical management.
Measurement of Blood Pressure
Blood pressure is the most commonly measured index of cardiovascular stability in the perioperative period. Anesthetics and surgery cause changes in blood pressure that are often rapid and may be great enough to cause harm unless anticipated and treated. However, a change in blood pressure that alters perfusion pressure may not change organ blood flow. This is because most vital organs can autoregulate blood flow in response to changes in mean arterial blood pressure, permitting constant blood flow over a range of perfusion pressures.18 In chronically hypertensive patients, the boundaries for autoregulation are shifted so that significant decreases in organ perfusion may occur with blood pressures in the considered “normal” range. Both the type and dose of anesthetic medications can affect the relationship between vital organ perfusion and blood pressure. For example, volatile anesthetics are potent vasodilators that tend to disrupt cerebral autoregulation in a dose-dependent manner to render blood flow more linearly dependent on blood pressure (Fig. 10-1).
Autoregulation maintains a constant cerebral blood flow between mean arterial blood pressures of 50 to 150 mm Hg in the conscious,unanesthetized state. Increasing doses of potent inhalation anesthetics produce a dose-dependent disruption of autoregulation owing to cerebral vasodilatation. (Modified with permission from Shapiro H: Anesthesia effects upon cerebral blood flow, cerebral metabolism, electroencephalogram and evoked potentials, in Miller RD [ed]: Anesthesia, 2nd ed. New York, Churchill-Livingstone, 1986; p 1249.)
Although noninvasive blood pressure monitoring suffices for most patients during routine noncardiac surgery, direct measure of arterial blood pressure with an indwelling arterial catheter is necessary for cardiac surgery in order to detect changes quickly, to measure nonpulsatile blood pressure during cardiopulmonary bypass, and to facilitate arterial blood sampling for laboratory analysis. The measuring system includes an intra-arterial catheter and low-compliance saline-filled tubing connected to a transducer with a pressure-sensing diaphragm. The transducer has a strain gauge that converts the mechanical energy (displacement of the diaphragm by a change in pressure) into an electric signal that typically is displayed as a pressure waveform with numeric outputs for systolic, diastolic, and mean pressures. The mean blood pressure is determined by calculating the area under several pulse waveforms and averaging over time. This represents a more accurate measure of mean arterial blood pressure than weighted averages of systolic and diastolic pressures.
The transducer requires a zero reference at the level of the right atrium. Any movement of the patient or the transducer that changes the vertical distance between the transducer and the right atrium affects the value of the blood pressure measured. If the transducer is lowered, the pressure diaphragm senses arterial blood pressure plus hydrostatic pressure generated from the vertical column of fluid contained in the tubing and displays a falsely high blood pressure. A transducer elevated above the zero reference level decreases the displayed blood pressure. A 1-cm column of water (blood) exerts a hydrostatic pressure equal to 0.74 mm Hg. Small changes in patient or transducer position have a relatively insignificant effect on arterial blood pressure measurements but have a more important effect on lower-amplitude pressure measurements, such as central venous, pulmonary artery, and pulmonary artery occlusion pressures.
The radial artery is the most common site for insertion of an intra-arterial catheter. Twenty-gauge catheters are preferred when the radial artery is used because larger catheters are more likely to cause thrombosis. The wrist is chosen for arterial cannulation most often because of dual blood supply to the hand, preventing ischemia in the setting of a catheter related thrombosis. A patent palmar arch allows the hand to remain perfused despite a potential thrombosis of either the radial or ulnar artery assuming adequate flow through the other vessel. Distal embolization remains a risk though. The radial is the more superficial vessel and generally considered easier to cannulate. The Allen's test was designed to assess ulnar and palmar arch blood flow during abrupt occlusion of the radial artery, but its value to predict morbidity with radial artery cannulation is equivocal.19 Other sites selected for the insertion of an intra-arterial catheterization include the brachial, axillary, and femoral arteries.
The contour of the arterial pressure waveform is different in central and peripheral arteries. The propagating pressure waveform loses energy and momentum with a corresponding delay in transmission, loss of high-frequency components such as anacrotic and dicrotic notches, lower systolic and pulse pressures, and decreased mean pressure.20 The changes in the pulse waveform can be attributed to damping, blood viscosity, vessel diameter, vessel elastance, and the effects of reflectance of the incident arterial waveform by the artery-arteriolar junction.21,22 The blood pressure waveform measured in the ascending aorta is minimally affected by reflected waves in contrast to measurement of blood pressure in the dorsalis pedis or radial artery.
The contour of the pressure waveform can also be affected by the physical construction of the monitoring system. A hyper-resonant response to a change in pressure, or ringing, occurs when the frequency response of the monitoring system (ie, extension tubing, catheter, and stopcocks) is close to the frequency of the pressure waveform.20 The natural or resonant frequency fn of a monitoring system is defined by
where C is compliance of the measuring system, L is the length of the tubing, D is the diameter of the catheter extension tubing, and p is the density of the solution.
To prevent ringing, the natural frequency of the monitoring system fn must be greater than the frequencies of the pulse waveform. Any process that decreases fn, such as narrow, long, compliant tubing, may cause ringing.23 Ringing increases the value of the systolic blood pressure and decreases the value of the diastolic blood pressure but generally does not affect the value of the mean arterial pressure.
Damping is the tendency of the measuring system, through frictional losses, to blunt the peaks and troughs in a signal.24 Kinks in the pressure tubing or catheter, stopcocks, and air bubbles contribute to damping. Overdamped systems underestimate systolic blood pressure and overestimate diastolic blood pressure. When long lengths of tubing are necessary, deliberate damping may improve the fidelity of the arterial waveform.
Testing a measuring system for ringing and damping ensures that an arterial contour is reproduced faithfully. A simple test involves a brief flush of the high-pressure saline-filled catheter extension assembly. Flush and release should produce a rapid return of the pressure waveform to baseline with minimal oscillations. A gradual return to baseline and loss of higher-frequency components of the waveform suggests overdamping. A rapid return to baseline followed by sustained oscillations suggests ringing.
The intraoperative ECG monitor has evolved from the fading-ball oscilloscope to a sophisticated microprocessor analog display. ECG signals are filtered digitally to eliminate electrical artifact produced by high-frequency (60-Hz) electrical power lines, electrocautery, patient movement, and baseline drift. The bandwidth filter modes are diagnostic, monitor, and filter. The diagnostic mode has the widest bandwidths (least filtered signal) and is preferred for detecting ST-segment changes caused by myocardial ischemia. Monitor and filter modes have progressively narrower bandwidths that effectively eliminate high-frequency interference and baseline drift but decrease the sensitivity of detecting ST-segment changes and decrease the specificity of ST-segment change to diagnose myocardial ischemia. Abnormal ST-segment depression (>1 mV) can occur from excessive low-frequency filtering and result in the misdiagnosis of myocardial ischemia. Filter modes are useful for detecting P waves and changes in cardiac rhythm in the presence of high-frequency interference.
The ECG is the most sensitive and practical monitor for the detection and diagnosis of disorders of cardiac rhythm and conduction and myocardial ischemia and infarction. Continuous monitoring of leads II and V5 is common (Fig. 10-2). Together these leads detect greater than 90% of ischemic episodes in patients with coronary artery disease who have noncardiac surgery.25
Standard intraoperative electrocardiogram (ECG) lead placement. Typically, leads II and V5 are monitored continuously.
Diagnostic criteria for myocardial ischemia based on the ECG are (1) acute ST-segment depression greater than 0.1 mV 60 ms beyond the J point or (2) acute ST-segment elevation greater than 0.2 mV 60 ms beyond the J point26 (Fig. 10-3). The normal ST-segment curves smoothly into the T wave. Flat ST segments that form an acute angle with the T wave or downsloping ST segments are worrisome for subendocardial ischemia. ST-segment elevation occurs with transmural myocardial injury but also may occur after direct current (dc) cardioversion and in normal adults. The lack of specificity of ST-T-wave changes for myocardial ischemia is a major limitation of intraoperative ECG monitoring. Pericarditis, myocarditis, mitral valve prolapse, stroke, and digitalis therapy may produce changes in the ST segment that mimic myocardial ischemia.
Automated ST-segment monitoring of the ECG can be used to detect intraoperative myocardial ischemia. General criteria for myocardial ischemia are ST-segment depression greater than 0.1 mV or ST-segment elevation greater than 0.4 mV that persists for longer than 1 minute. At fast heart rates, the ST-segment measurement point may occur on the upslope of the T wave, causing erroneous indication of ST-segment elevation.
Digital signal processing handles much larger quantities of information than the unaided eye and may increase the ability to detect ischemic episodes. ST-segment position analyzers automatically measure the displacement of the ST segment from a predetermined reference and enhance the ability to quantify changes in ST-segment position. Appropriate application requires accurate identification of the various loci in the P-QRS-T-wave complex. The operator defines the baseline and the J point of a reference QRS complex by movement of a cursor. New QRS-T-wave complexes are superimposed onto a predefined mean reference complex. Vertical ST-segment displacement is measured in millivolts and displayed graphically in 1-mV increments (Fig. 10-3). Because the accuracy of automated ST-segment monitoring is vulnerable to baseline drift and dependent on appropriate identification of the PR and ST segments, the diagnosis of myocardial ischemia is always verified by inspecting the actual ECG tracing.
Disturbances of rhythm and conduction are common during anesthesia and especially during cardiac surgery. Instrumentation of the heart, hypothermia, electrolyte abnormalities, myocardial reperfusion, myocardial ischemia, and mechanical factors such as surgical manipulation of the heart affect normal propagation of the cardiac action potential. Heart rate is measured by averaging several RR intervals of the ECG. The ECG may not sense the R wave of the selected lead if the electrical vector is isoelectric. A prominent T wave or pacemaker spike may be miscounted as an R wave by the ECG and artifactually double the rate. Usually, heart rate is best monitored by selecting the lead with an upright R wave and adjusting the sensitivity.
The QT interval can be measured only on hard copy. A normal QT interval is less than half the RR interval, but the QT interval must be corrected for heart rates higher than 90 or lower than 65 beats per minute. A prolonged QT interval increases the risk of reentrant ventricular tachydysrhythmias and may occur from hypokalemia, hypothermia, and drug effect (eg, quinidine or procainamide).
The ECG is monitored to confirm the electrical dormancy of the heart during aortic cross-clamping and perfusion with cold cardioplegia. Hypothermia decreases action potential conduction velocity, and high-dose potassium decreases the transcellular membrane potassium concentration gradient to prevent depolarization of cardiac muscle. During cardiopulmonary bypass and aortic cross-clamping, the loss and persistent absence of electromechanical activity suggest that myocardial oxygen consumption is maintained at a minimum.
Monitoring the ECG is most valuable when it begins before induction of general anesthesia. A hard copy of the pertinent leads permits comparison should a change be detected. An abnormal or marginal finding is less worrisome if it was present in the preoperative ECG and remains unchanged during the perioperative period. However, new-onset ST-T-wave changes or disturbances in rhythm and conduction suggest an ongoing active process that usually requires immediate attention.
Capnometry is the measure of carbon dioxide (CO2) concentration in a gas. The capnogram is the continuous graphic display of airway carbon dioxide partial pressure (Fig. 10-4). The capnogram is the single most effective monitor for detecting esophageal intubation, apnea, breathing circuit disconnects, accidental extubation of the trachea, and airway obstruction. Tracheal intubation is verified by detection of physiologic carbon dioxide concentrations in the exhaled gas. Changes in its contour reflect disorders of ventilation, carbon dioxide production, or carbon dioxide transport to the lungs. A steep increase in the phase 3 slope of the exhaled CO2 concentration suggests partial airway obstruction, either mechanical (eg, tube kinking) or physiologic (eg, bronchospasm). A progressive decrease in exhaled carbon dioxide concentration occurs with decreased CO2 production (eg, hypothermia), increased minute ventilation, or increases in physiologic dead space ventilation such as pulmonary embolus or a drop in cardiac output. A progressive increase in exhaled carbon dioxide concentration occurs with hypoventilation, increased CO2 production (eg, malignant hyperthermia), or increased delivery of CO2 to the lungs (eg, during weaning off bypass with an increase in pulmonary blood flow). Despite the complex interplay of mechanical and physiologic factors that affect the shape of the capnogram, with any abrupt change in contour, an acute change in the patient's cardiovascular, pulmonary, or metabolic state should be considered.
The normal capnogram: (1) inspired CO2 concentra-tion zero, (2) washout of anatomic dead space, (3) plateau represents alveolar gas CO2 content, and (4) beginning of inhalation.
Inhaled volatile anesthetics are different from other parenteral medications. They are delivered as a vapor through the breathing circuit. The clinical effect is determined by the partial pressure of this vapor in the brain, which is in equilibrium with the blood and the alveoli. Measuring the partial pressure of the volatile anesthetic at end expiration approximates the partial pressure in the alveoli and therefore the brain. Monitoring the end-tidal gas mixture adds precision to the administration of inhaled anesthetics and guards against inadvertent overdose.
The concentration of anesthetic gases is displayed as a percentage of the partial pressure of the anesthetic vapor to atmospheric pressure. This concentration can be measured using a variety of techniques, including mass spectroscopy, infrared spectroscopy, Raman spectroscopy, electrochemical and polarographic sensors, and piezoelectric absorption.27
Pulse oximeters were adopted universally into the practice of anesthesia almost immediately after their introduction despite the lack of published data demonstrating improved outcome with their use. The pulse oximeter is reusable, inexpensive, and noninvasive while providing continuous online data regarding the arterial hemoglobin saturation and pulse rate. The pulse oximeter can detect a decrease in oxyhemoglobin saturation even before changes in the color of the patient's skin or blood are evident.28 Its major limitations include a subjectivity to electrical interference and motion artifact as well as a high failure rate during periods of low flow or inadequate perfusion. In addition, pulsatile flow is required for proper function.29
The pulse oximeter is able to measure the percentage of oxyhemoglobin in arterial blood because of differences in the optical absorption properties of oxy- and deoxyhemoglobin. Using transillumination, oximetry emits wavelengths of 660 and 940 nanometers (nm). Oxyhemoglobin has a higher optical absorption in the infrared spectrum (940 nm), whereas reduced hemoglobin absorbs more light in the red band (660 nm). The ratio R of light absorbance at the two wavelengths is a function of the relative proportions of the two forms of hemoglobin. Rapid signal processing then permits reliable and rapid determination of the relative proportion of oxy- and deoxyhemoglobin. Calculation of arterial hemoglobin saturation is based on calibration algorithms derived from R ratios in healthy volunteers.
Arterial oxyhemoglobin saturation is distinguished from venous through the use of photoplethysmography by isolating the pulsatile component of the absorbed signal. The peaks and troughs in the blood volume pulse of the site being transilluminated produces a corresponding pulsatile effect on light absorption, rendering the calculated oxyhemoglobin saturation independent of nonpulsatile venous blood and soft tissue. Because the R values were determined in healthy volunteers, they are less accurate at oxyhemoglobin saturations below 70%. Motion artifact produces a high absorption of light at both wavelengths and an R value of approximately 1, which corresponds to an oxyhemoglobin saturation of approximately 85%.
Measurement of Temperature
Profound changes in body temperature during cardiac surgery are common, often deliberate, and affect vital organ function (Fig. 10-5). Anesthetized patients are poikilothermic. Intrinsic temperature regulation normally controlled by the hypothalamus fails during general anesthesia. Hypothermia occurs by passive and active heat loss. Passive mechanisms of cooling include radiation, evaporation, convection, and conduction. Active cooling usually occurs with extracorporeal circulation and with the use of cold or iced solutions poured into the chest cavity. Deliberate hypothermia during cardiac surgery is designed to arrest and cool the heart, decrease systemic oxygen consumption, and protect organs from hypoperfusion. Hyperthermia may result from preexisting fever, bacteremia, malignant hyperthermia, or overzealous rewarming during cardiopulmonary bypass. The monitoring and control of systemic temperature is important because either unintentional hypothermia or hyperthermia can be deleterious to the patient.
Time course of changes in nasopharyngeal (black) and bladder (blue) temperature during a cardiac operation performed on cardiopulmonary bypass employing deliberate hypothermia and deep hypothermic circulatory arrest (DHCA). Marked events: A = onset of deliberate hypothermia, B = onset of DHCA, C = end of DHCA, D = rewarming to 29°C, E = rewarming to normothermia.
Hypothermia after cardiopulmonary bypass is the result of ineffective or most commonly insufficient rewarming. Cold operating rooms; cold, wet surgical drapes; a large surgical incision; and administration of cold intravenous fluids also contribute. Hypothermia exacerbates dysrhythmias and coagulopathy, potentiates the effects of anesthetic drugs and neuromuscular blockers, increases vascular resistance, decreases the availability of oxygen, and contributes to postoperative shivering. The elderly are especially susceptible because of limited compensatory reserve. Although there is evidence to support the efficacy of mild therapeutic hypothermia for brain protection after cardiac arrest and subsequent resuscitation,30 the optimal target temperature of induced hypothermia for neuro-protection during cardiopulmonary bypass, or even the need for hypothermia at all remains highly contentious in cardiac surgery currently.31,32
Temperature typically is monitored from several sites during cardiac surgery. Blood temperature can be measured from the tip of a pulmonary artery catheter and within the cardiopulmonary bypass circuit (typically venous and arterial lines). Blood temperature is the first to change in response to deliberate hypothermia or active rewarming during cardiopulmonary bypass. Nasopharyngeal and tympanic temperatures reflect the temperature of the brain and closely track blood temperature because these sites are highly perfused. Rectal and bladder temperatures provide a measure of core temperature only at equilibrium. Esophageal temperature often underestimates core temperature because of the cooling effects of ventilation in the adjacent trachea. Axillary and inguinal temperatures are shell measurements and are impractical. The possibility of an increased gradient between the measured nasopharyngeal and actual brain temperature should be considered and careful attention should be made to prevent cerebral hyperthermia.33 Arterial inflow temperatures during rewarming should be limited to a maximum of 37 to 37.5°C to help prevent a large gradient between the inflow temperature and the brain temperature. Post operative hyperthermia has been associated with a worsened neurologic outcome.33
The degree and site of temperature changes are important indicators of an intact circulatory system. A persistent discrepancy in temperature between two sites may be a sign of malperfusion. Rewarming during cardiopulmonary bypass normally is associated with an increase in nasopharyngeal or tympanic temperature accompanied by a more gradual increase in temperature in organs with low perfusion. A persistently cold nasopharynx with a normal rate of increase in rectal temperature may be a result of aortic dissection and hypoperfusion of the head.
Measurement of Cardiac Output and Central Venous and Pulmonary Artery Pressures
Cannulation of the central venous circulation permits central administration of drugs, rapid administration of fluids through short, large-bore cannulas, and the measure of central venous pressure. The most commonly used site for central venous access is the internal jugular vein because of reliable insertion, ease of access from the head of the table, decreased risk of pneumothorax, and decreased risk of catheter kinking during sternal retraction. The subclavian vein is the preferred site for insertion of a central venous catheter for long-term intravenous total parenteral nutrition because of a decreased risk of blood-borne infection.34 The most important complication of internal jugular vein cannulation is inadvertent puncture or cannulation of the carotid or subclavian artery. Cannulation of the central venous circulation is confirmed by manometry, either electrical or mechanical, before insertion of a large-bore catheter. Ultrasound-guided cannulation of the internal jugular vein renders the procedure less dependent on anatomic landmarks and is associated with a decrease in the number of unsuccessful cannulation attempts35 (Fig. 10-6). With the advent and wide availability of portable ultrasound imaging devices, ultrasound-guided central venous cannulation is becoming commonplace to increase the success rate and decrease the risk of complications.36 TEE can be used to verify the presence of a wire in the right atrium or superior vena cava (SVC). It can also be used to verify correct placement of the central venous or pulmonary artery catheter tip.37
A two-dimensional short-axis image of the internal jugular vein (IJV) and carotid artery (CA) using a handheld ultra-sound transducer.
Central venous pressure (CVP) can be measured via the central venous catheter or a proximal port in a Swan-Ganz catheter. The pressure is best measured at end exhalation to exclude the transmission of intrathoracic pressure. The CVP tracing is normally composed of five waves. The A and C waves represent atrial contraction and isovolumetric right ventricular contraction. The X descent occurs with right ventricle ejection as the tricuspid annulus descends. The V and Y waves represent right atrial filling, first during late ventricular contraction and then during early diastole with tricuspid valve opening.
The CVP is traditional thought of as a measure of the overall volume status of a patient, with normovolemia represented by a CVP of 6 to 10 mm Hg during positive pressure ventilation. However, studies have consistently shown that the CVP does not correlate with measured circulated blood volume or fluid challenge responsiveness.38–41 The CVP value is dependent on the interaction between cardiac function and venous return. Venous return can be affected not only by the mean circulatory filling pressure (the equalization of pressure throughout the entire vascular system if heart was theoretically stopped temporarily) and the cardiac function but also by changes in venous tone. For example, after the induction of general anesthesia, a drop in CVP often represents anesthetic induced venodilation, not acute hypovolemia. The sensitivity of the CVP as a measure of volume status is affected by the body's advanced compensatory mechanisms designed to maintain homeostasis, and these significant changes in blood volume may not be associated with a change in the CVP. A 10% loss of blood volume can easily be compensated for by an auto-transfusion of volume from venous capacitance vessels without a change in the CVP and significant volume loading can occur without a concomitant rise in CVP as fluid accumulates in the same splanchnic venous beds.
Using the CVP as a measure of volume status is further complicated by the differences between intramural and transmural pressure. The pressure measured by an intravascular catheter represents the intramural pressure. The transmural pressure is the difference between the intramural pressure and the pressure surrounding the vascular structures. This difference represents the driving pressure for venous return whereas the measured intramural pressure may not. Increases in intrathoracic and intra-abdominal pressure as well as pericardial pressure will be transmitted to the intravascular space, increasing the CVP without increasing transmural pressure and thus without increasing venous return or indicating a degree of hypervolemia.
Beyond simply measuring pressure, the contour of the CVP waveform can provide other important clinical information. The presence of cannon A waves can confirm the diagnosis of a junctional rhythm in situations wherein the ECG is difficult to discern (also seen with tricuspid stenosis or right ventricular diastolic dysfunction). Large V waves during ventricular systole occur with significant tricuspid regurgitation. Deep X and Y descents can be seen during right sided volume overload, often in conjunction with tricuspid regurgitation. Right ventricular diastolic dysfunction can be suggested by a flattening of the Y descent.
Pulmonary artery catheters (PAC) are inserted via the central venous circulation through the right side of the heart with the catheter tip positioned just downstream of the pulmonic valve. The PAC can measure pulmonary artery pressure, pulmonary artery occlusion pressure, cardiac output, and mixed venous oxygen saturation and also permits the calculation of the derived values of systemic and pulmonary vascular resistance. The pulmonary artery occlusion pressure can be used as an index of left ventricular preload in the absence of mitral stenosis. However, the use of this pressure measurement to estimate preload is limited because of variability in left ventricular size and compliance. Pulmonary artery occlusion pressures (PAOP) can be affected by myocardial compliance, mode of ventilation, and ventricular afterload. Similar to the CVP, the PAOP can also be affected by increases in intrathoracic, intra-abdominal, and pericardial pressure. Increases in PA pressures or PAOP may often indicate myocardial dysfunction or myocardial ischemia (Fig. 10-7). Hemodynamic parameters derived from the pulmonary artery catheter have been shown not to be as sensitive or as specific for detecting myocardial ischemia as the ECG.42
Pulmonary artery occlusion pressure tracing at two time points. The acute onset of myocardial ischemia (B) was associated with ST-segment depression in ECG lead V5, increased pulmonary artery pressures, and a prominent v wave.
Complications associated with insertion of a PAC include infection, dislodgment of endocardial pacemaker wires, dislodgement of right atrial or ventricular clot or tumor, atrial and ventricular arrhythmias, pulmonary infarction, pulmonary artery rupture, perforation of the right atrium or ventricle, catheter entrapment, and heart block. The incidence of right bundle-branch block (RBBB) from catheter insertion is approximately 3% and may lead to complete heart block in patients with a preexisting left bundle-branch block (LBBB).43 Most PACs are now heparin-bonded to decrease the incidence of thrombus formation.44 Chronic indwelling PACs are associated with a progressive thrombocytopenia.45
Multiport pulmonary artery catheters equipped with a tip thermistor permit the measurement of pulmonary blood flow or cardiac output by thermodilution. Thermodilution cardiac output uses an indicator-dilution technique. The indicator, a known volume of cold saline, is injected rapidly into a proximal port in the right atrium. Cardiac output (CO) is calculated from the rate of change in blood temperature at the PAC catheter tip over time using the Stewart-Hamilton equation:46,47
where CO is cardiac output, V is the volume of the injectate, TB is the blood temperature at time 0, T1 is the injectate temperature at time 0, ΔTB(t) is the change in blood temperature at time t, K1 is the density factor, and K2 is the computation factor.
Thermodilution measures the degree of mixing that occurs between the cold injectate and blood. More mixing implies increased flow. Complete mixing of 10 mL of cold injectate with a circulating blood volume produces a small decrease in temperature at the catheter tip. Poor mixing, suggestive of slow, sluggish flow, produces a large decrease in temperature as the injectate bolus passes the thermistor. The derived value for cardiac output is inversely proportional to the area under the thermodilution curve. Rapid infusion of cold intravenous fluids at the time of measurement may falsely increase the derived cardiac output. Thermodilution measures right-sided cardiac output, which will not equal left-sided cardiac output in patients with intracardiac shunts.
Cardiac output may be monitored nearly continuously using a specialized PAC. The continuous cardiac output catheter intermittently heats blood adjacent to a proximal portion of the catheter and senses changes in blood temperature at the catheter tip using a fast-response thermistor. The method requires no manual injections, and values are acquired, averaged, and updated automatically every several minutes. The only true disadvantage to these catheters is the increased cost.
Mixed venous oxygen saturation (Sv̅O2) can be measured intermittently by manual blood sampling from the pulmonary artery port or continuously using a modified PAC equipped with an oximeter. Assuming normal oxygen consumption, a normal (Sv̅O2) generally denotes adequate oxygen delivery but does not provide information about the adequacy of perfusion to specific organs. A normal (Sv̅O2) may not reflect adequate tissue perfusion in patients with intracardiac shunts, sepsis, or liver failure. Oxygen consumption may change significantly during or after cardiac surgery as patients are warmed or as they begin to emerge from anesthesia, leading to a drop in (Sv̅O2). However, a significant decrease in (Sv̅O2) often signifies a decrement in oxygen delivery representing compromised cardiac output, anemia, or hypoxia.
Sv̅O2 provides an alternative method to calculate cardiac output if oxygen consumption is assumed based on a nomogram derived from resting awake subjects. By the Fick equation, cardiac output is equal to the rate of systemic oxygen consumption divided by the arterial-venous oxygen content difference:
where Vo2 is oxygen consumption, CO is cardiac output, Cao2 is the oxygen content in arterial blood, and CvO2 is the oxygen content in mixed venous blood. Errors may occur when oxygen consumption is well below the assumed nomogram in sedated, hypothermic, or anesthetized patients, with an overestimation of cardiac output.
Although routine use of a pulmonary artery catheter for monitoring patients during cardiac operation is debated, it does provide clinical information that is used to direct therapy in high-risk patients (Figs. 10-8 and 10-9). The pulmonary artery catheter appears to demonstrate continued utility in the care of patients with pulmonary hypertension and right ventricular dysfunction.48 Because an insidious decrease in (Sv̅O2) may be an early warning of impending circulatory insufficiency, ventricular dysfunction, ongoing bleeding, or impending tamponade, (Sv̅O2) pulmonary artery catheters may be particularly useful in the intensive care unit, where early deterioration can be detected and treated before an adverse event occurs.
Intraoperative hemodynamic recordings showing the time sequence of systemic severe vasodilation (A) and catastrophic pulmonary vasoconstriction–type (B) protamine reactions during the reversal of heparin anticoagulation in patients undergoing heart operation. Arterial blood pressure (ABP) and pulmonary artery pressure (PAP) decrease in parallel during systemic vasodilation. In contrast, an increase in PAP and central venous pressure (CVP) precedes the decrease in ABP during the pulmonary vasoconstriction–type reaction. The decreases in end-tidal carbon dioxide concentration (ETCO2) during the protamine reactions reflect the decrease in blood flow through the lungs.
Decreased left ventricular preload produced by graded estimated blood volume deficits (EBVs) was associated with a serial decreases in the mixed venous oxygen saturation (Sv̅O2), cardiac stroke volume (SV), left ventricular end-diastolic meridional wall stress (EDWS), left ventricular end-diastolic cavity cross-sectional area (EDA), and pulmonary artery occlusion pressure (PAOP). Patients with dilated cardiomyopathy displayed less change in SV and Sv̅O2 in response to equivalent EBV deficits. *p <.05 versus baseline value (ANOVA for repeated measures). (Modified with permission from Cheung AT, Weiss SJ, Savino JS: Protamine-induced right-to-left intracardiac shunting. Anesthesiology 1991; 75:904.)
Measurement of Electrolyte Concentration
Electrolyte abnormalities are common during and after cardiopulmonary bypass and are measured intermittently, typically using stat laboratory tests.49 The capability to detect and treat electrolyte disturbances is an important aspect of intraoperative care.
Abnormalities in sodium and water homeostasis are often associated with heart failure and compounded by hemodilution during surgery with or without cardiopulmonary bypass. Nonosmotic secretion of arginine vasopressin provoked by surgical stress, pain, hypotension, or nonpulsatile perfusion contributes to the development of hyponatremia by stimulating renal retention of free water. A 2- to 5-mEq/L decrease in the plasma sodium concentration is expected after beginning cardiopulmonary bypass but does not normally require treatment. Hypernatremia may be caused by excessive diuresis without free-water repletion or by the administration of hypertonic sodium bicarbonate solutions. An 8.4% sodium bicarbonate solution has an osmolality of 2000 mOsm/l or 6.9 times the osmolality of plasma. Hyperkalemia is common because of high-potassium cardioplegic solutions that are distributed into the systemic circulation. Hyperkalemia during cardiac surgery also may be caused by hemolysis, acidosis, massive depolarization of muscle, and tissue cell death. Increasing serum potassium concentration is manifested by peaked T waves, a widened QRS complex, disappearance of the P wave, heart block, and conduction abnormalities that may be life threatening. Very high concentrations of potassium used to provide cardioplegia inhibit spontaneous depolarization and produce asystole. Patients with diabetes mellitus are at increased risk for hyperkalemia because cellular uptake of potassium is mediated by insulin. Impaired renal excretion of potassium enhances hyperkalemia in patients with renal insufficiency. The initial treatment of hyperkalemia is aimed at redistributing extracellular potassium into cells, but the elimination of potassium from the body requires excretion by the kidneys or gastrointestinal tract. Insulin and glucose administration rapidly decrease extracellular potassium by redistributing the ion into cells. Alkalosis, hyperventilation, and beta-adrenergic agonists also favor redistribution of potassium into cells, but the response is less predictable. Calcium carbonate and calcium chloride antagonize the effects of hyperkalemia at the cell membrane. A typical intravenous dose of glucose and insulin for the acute treatment of hyperkalemia is 1 g/kg of glucose and 1 unit of regular insulin per 4 g of glucose administered.
Hypokalemia can occur during cardiac surgery from hemodilution with nonpotassium priming solutions, diuresis, or increased sympathetic tone during nonpulsatile perfusion. Intraoperative hypokalemia is exacerbated by preoperative potassium depletion from chronic diuretic therapy. Routine insulin administration for hyperglycemia and the frequent use of beta2-adrenergic agonists stimulate cellular uptake of potassium, also contributing to the frequent occurrence of hypokalemia in cardiac surgical patients. Hypokalemia predisposes to atrial arrhythmias, ventricular ectopy, digitalis toxicity, and prolonged response to neuromuscular blocking drugs. Hypokalemia is treated by slow administration of KCl in increments of 10 mEq, with potassium concentrations measured between doses.
Hypocalcemia decreases myocardial contractility and peripheral vascular tone and is associated with tachycardia.50,51 Hypocalcemia produces prolongation of the QT interval and T wave inversions, but significant arrhythmias owing to disturbances in ionized calcium concentration are not common. Hypocalcemia occurs soon after the onset of cardiopulmonary bypass but may resolve without treatment. Increasing serum concentrations of parathyroid hormone during cardiopulmonary bypass may explain, in part, the gradual increase in ionized calcium concentration to precardiopulmonary bypass levels.52 The etiology of cardiopulmonary bypass–induced hypocalcemia probably is multifactorial, but hemodilution and decreased metabolism of citrate after rapid blood transfusion are contributing factors. The routine administration of calcium salts without prior measurement of ionized calcium concentration poses the risk of hypercalcemia. Excessive calcium administration may increase the risk of postoperative pancreatitis and myocardial reperfusion injury.53
Magnesium deficiency is common in cardiac surgical patients, and acute magnesium supplementation decreases the incidence of postoperative cardiac dysrhythmias and overall morbidity after cardiac operations.54,55 However, measuring total plasma magnesium concentration has questionable clinical significance because the value primarily reflects the concentration of protein-bound magnesium and not physiologically active, ionized magnesium.56
Blood glucose control for cardiac surgery is controversial. Most would agree that treatment of severe hyperglycemia is warranted. Hyperglycemia has been shown to increase morbidity and mortality in nonsurgical patients after stroke and myocardial infarction as well as in patients after cardiac surgery.57–59 However, there are conflicting data regarding support for an aggressive treatment strategy designed at normalizing blood glucose. In a mixed surgical critical care population, intensive insulin therapy to maintain glucose ≤110 mg/dl reduced morbidity and mortality but with an increased incidence of hypoglycemia.60 In a meta-analysis of randomized clinical trials conducted in intensive care units, evidence suggested that intensive insulin therapy significantly increased the risk of hypoglycemia and offered no overall mortality benefit.61–63 Thus far, the literature has only focused on glucose control in the intensive care units. The benefit, if any, of glucose control during cardiac surgery currently remains unknown. No evidence exists supporting or refuting aggressive control. The stress of cardiac surgery, acute changes in catecholamine levels, large swings in body temperature, administration of steroids, and large volumes of intravenous fluids all affect blood glucose. The efficacy of intravenous insulin to reduce blood glucose is further complicated by hypothermia when insulin degradation is slowed as is insulin's ability to induce cells to increase glucose uptake (ie, insulin resistance). Over simplistic protocols that do not take into account the increased half life of intravenous insulin and its decreased effectiveness during cold conditions may result in patients receiving progressively increasing doses of insulin with little immediate effect. The slowly metabolized insulin will linger during hypothermia. With systemic rewarming, the large depots of insulin could produce severe hypoglycemia as well as clinically significant hypokalemia. Regardless of glucose management strategy, rapid and frequent measurement of blood glucose is essential.
Monitoring the Nervous System
Neurologic complications, including stroke, paralysis, cognitive dysfunction, blindness, and peripheral nerve injury, are second only to heart failure as the major cause of morbidity and mortality after cardiac surgery.64 Because a full neurologic examination is not possible during general anesthesia, the objective of intraoperative neurophysiologic monitoring is to provide a means for early detection of neurologic injury or impending injury to permit early intervention in an effort to restore function before the injury becomes permanent.
EEG is a measure of spontaneous electrical brain activity recorded from electrodes placed in standard patterns on the scalp. Electrical activity of individual electrodes is amplified and then recorded as continuous wavelets that have different frequencies and amplitudes. These data can be displayed as raw EEG or broken down into basic components of frequency and amplitude and displayed as a spectral analysis. A change in EEG activity from baseline during a procedure may indicate ischemia of the cerebral cortex as a consequence of hypoperfusion. Intraoperative EEG can also detect seizure activity masked by general anesthetics or electrocortical silence induced by deliberate hypothermia.
Intraoperative application of EEG can be justified for any cardiac operation where there is a risk for cerebral hypoperfusion such as combined cardiac and carotid endarterectomy procedures. EEG is very sensitive for detecting cerebral malperfusion and can be used to detect malperfusion of the aortic arch branch vessels during aortic dissection repair.65 Hypothermia produces dose-dependent slowing of the EEG. The predictable actions of hypothermia on EEG activity have made EEG a useful monitor and surrogate for brain temperature in order to determine the adequacy of deliberate hypothermia for circulatory arrest procedures.66 When performing intraoperative EEG, it is important to recognize that anesthetic agents can attenuate EEG amplitude and frequencies. Marking the EEG in the event of bolus sedative hypnotic administration or when the inhaled anesthetic concentration is changed will help to distinguish EEG changes caused by general anesthetics from changes as a consequence of neurologic injury.
The occurrence of awareness under anesthesia is a rare, approximately 0.2% with general anesthesia.67 In patients at high risk for awareness, the incidence may be near 1%.68,69 Cardiac surgery is associated with an increased risk of intraoperative awareness.69–73 The American Society of Anesthesiologists published a “practice advisory” intended to assist decision making pertaining to this issue.74 The consequences of anesthetic awareness can be significant, leading to long-term psychologic distress including post-traumatic stress disorder. This is caused by several factors, such as hemodilution of intravenous anesthetics from the bypass circuit, hemodynamic instability leading to the decrease or discontinuation of anesthetics, and the inability to measure end-tidal concentrations of inhaled anesthetics if used during bypass. There are several devices commercially available for monitoring anesthetic depth. The bispectral index or BIS monitor (Aspect Medical Systems) is the most widely used of these devices. By using a pad placed directly on the scalp, a frontal electroencephalogram signal is processed in order to calculate a dimensionless value of the patient's level of consciousness. The BIS values range between 0 and 100 with below 40 indicating a deep level of consciousness and between 40 and 60 considered adequate for most general anesthetics.75 It's clinical utility in preventing anesthetic awareness or reducing over-administration of anesthetics has been studied in several randomized trials with mixed results.67,75,76
Near Infrared Cerebral Oximetry
Although the technology has been in existence for over 25 years,77 only within the last 5 to 10 years has the use of continuous noninvasive cerebral oxygen saturation monitoring expanded. The device is based, like pulse oximetry, on the different absorption characteristics of oxygenated and deoxygenated hemoglobin when exposed to near infrared spectrum light (NIRS). Applied on each temple are two pads that each emits a near-infrared light signal that penetrates all tissues in the cranium using the transillumination characteristics of the skull. Signal contamination from extracerebral tissue saturations such as the bone and soft tissue are prevented from mixing with those from the brain and cerebral vasculature by using the concept of spatial resolution. Each pad employs two separate signal detectors at predetermined distances from the original light source. Signals from shallow tissues (extracerebral) return to the detector closer to the original light source and thus can be separated from the interpretation of signals from the deeper target tissues (intracerebral) that take a longer path.78 The sample size of the brain tissue monitored, however, is small and represents only approximately 1 ml of brain.79
The cerebral oximeter provides a continuous measurement of the percentage of oxygenated hemoglobin residing in the tested intracerebral tissue. Because the ratio of arterial to venous blood is 15:85,80 the cerebral saturation measured primarily reflects cerebral venous saturation, ie, the balance between oxygen supply and demand. The only cerebral oximeter approved for use in the United States currently is the INVOS 4100 and 5100 (Somanetics Corp, Troy MI).
A major critique of the system is that there is no established normative range or threshold for pathologic change. In a study of healthy elderly noncardiac surgical patients, the mean baseline value was 63+/8%.81 However, a wide range of interpatient variability has been found.82 Other factors that can lead to variability include hemoglobin concentration of the measured site and pad positioning. A 20% decline from baseline values is the most commonly used indicator of pathologic change. It is based mostly on the findings of neurologic change after carotid artery occlusion during carotid endarterectomy.83,84 Another dilemma facing practitioners is what to do when faced with low starting baseline cerebral saturations and how to determine when pathologic changes occur in these patients. It has been estimated that 7% of patients have baseline saturations below 50%.82
The most validated use of cerebral oximetry in cardiac surgery is as a predictor of postoperative neurocognitive dysfunction and extended hospitalization in those patients who have prolonged cerebral desaturations in the operating room.85–87 Reports on the diagnosis and treatment of cerebral desaturation and presumed hypoperfusion during cardiac surgery include a diverse list of causes such as anemia, hypocapnia, extreme neck turning, SVC occlusion, and low perfusion pressure.79,88,89 There is also a growing list of reports of its utilization in guiding length of cooling and conduct of antegrade cerebral perfusion during aortic arch repair and circulatory arrest (Fig. 10-10).90–93 Although cerebral oximetry represents an exciting tool to monitor the adequacy of brain perfusion, issues with sensitivity and specificity remain, likely preventing a more widespread adoption currently. Embolic events to regions beyond the small sample sizes of brain will clearly be missed. To date there is only on randomized prospective study evaluating cerebral oximetry monitoring with a defined treatment algorithm.87
Changes in left hemispheric (Channel 1) and right hemispheric (Channel 2) cerebral oxygen saturation prior to general anesthesia (baseline), deliberate hypothermia on cardiopulmonary bypass (cooling on CPB), deep hypothermic circulatory arrest (DHCA), and restoration of antegrade cerebral perfusion (after DHCA). Cerebral oxygen saturation was measured by near infrared spectroscopy using percutaneous sensors positioned on both sides of the forehead.
Motor and Somatosensory Evoked Potentials
The most common application of somatosensory evoked potential (SSEP) or motor evoked potential (MEP) monitoring in cardiothoracic surgery is for the detection of spinal cord ischemia during open or endovascular repair of the descending thoracic or thoracoabdominal aorta (Fig. 10-11).94–96 During these operations decreases in lower extremity SSEP or MEP amplitudes, indicating spinal cord ischemia, can be used to prompt interventions to increase arterial pressure, decrease lumbar CSF pressure, or reimplant segmental arterial branches in an effort to increase spinal cord perfusion to prevent paraplegia. The detection of reversible spinal cord ischemia by intraoperative MEP or SSEP monitoring may also identify patients who may be at risk for delayed postoperative paraplegia. Decreased SSEP and MEP amplitude have been shown to correlate with spinal cord ischemia, but the sensitivity and specificity of these techniques for detecting and reducing the incidence of spinal cord ischemia remains to be verified.97 Conditions other than spinal cord ischemia may produce MEP or SSEP changes.94
Intraoperative somatosensory evoked potential (SSEP) recordings from the lower (left panel) and upper (right panel) extremities that demonstrated intraoperative spinal cord ischemia during thoracoabdominal aortic aneurysm repair. Lower extremity SSEPs were generated by stimulation of the anterior tibial nerves at the ankles (left panel). Upper extremity SSEPs were generated by stimulation of the median nerves at the wrists (right panel). Bilateral disappearance of SSEP signals from the right (R) and left (L) lower extremities recorded at the cortex (R1, R2, R3, L1, L2, L3) and spine (R4, L4) with preservation of SSEP signals from the lumbar plexus (R5, L5) and popliteal nerves (R6, L6) indicated the acute onset of spinal cord ischemia. Upper extremity SSEP signals from the right (R) and left (L) brachial plexus (ERBS), cervical spine (N13), and cortex (N20) were maintained during the episode. The light grey tracings were the baseline SSEP signals used for comparison.
Intraoperative monitoring of SSEP is performed by placing stimulating electrodes on the skin adjacent to peripheral nerves in the arms or legs. Electrical stimulation of the peripheral nerves in the limbs generates action potentials that can be measured from recording electrodes over the lumbar plexus, brachial plexus, spine, brainstem, thalamus, and cerebral cortex. Motor evoked potentials (MEPs) are generated by transcortical electrical stimulation of the motor cortex that produces myogenic potentials that can be recorded in skeletal muscle. In theory, monitoring MEP should be more sensitive and specific than SSEP to detect spinal cord ischemia in the territory supplied by the anterior spinal artery. An advantage of SSEP monitoring is that it is relatively reliable and easy to interpret during general anesthesia without a contraindication to neuromuscular blockade. Although high concentrations of inhaled anesthetics, thiopental, or propofol can attenuate cortical SEP signals, a balanced general anesthetic with inhaled anesthetics maintained at a concentration of 0.5 MAC provide consistent conditions for monitoring intraoperative SSEP.
Detection of Cerebral Embolization
Intraoperative TEE, epiaortic ultrasound, and transcranial Doppler (TCD) are instruments that can be used to assess the risk and detect arterial embolic events (Fig. 10-12). The embolic burden to the cerebral circulation measured by quantitative TCD correlates with the incidence of intraoperative surgical manipulation and postoperative neurologic deficits.98 Intraoperative TEE can be applied to detect right-to-left intracardiac shunting through an atrial septal defect,99,100 intracardiac masses,101,102 or residual air within the cardiac chambers.103 Routine epiaortic ultrasonography or TEE to assess the degree of aortic atherosclerosis and guide the insertion of the aortic cannula and application of the aortic cross-clamp has been recommended for decreasing the risk of neurologic injury during CPB.104
Middle cerebral artery blood flow velocity measured intraoperatively using a 2-MHz transcranial Doppler ultra-sound transducer. The phasic velocity profile in the top panel was recorded before cardiopulmonary bypass. The irregular high-velocity, high-amplitude signals recorded in the lower panel indicate microemboli traveling through the middle cerebral artery immediately after ventricular ejection.