The prospect of heart surgery is frightening, and relatively “heavy” oral or intramuscular premedica-tion was often given in the past, particularly when patients had coronary artery disease with good left ventricular function . However, in current practice, most patients receive no sedative-hypnotic premedication until their arrival on the surgical unit, at which time most will receive small doses of intravenous midazolam.
Benzodiazepine sedative-hypnotics (diazepam, 5–10 mg orally), alone or in combination with an opioid (morphine, 5–10 mg intramuscularly or hydromorphone, 1–2 mg intramuscularly), were often used in the past. Longer acting premedicant agents (eg, lorazepam) are avoided by most practi-tioners to permit “fast tracking” of patients through their recovery.
The best practitioners of cardiac anesthesia formu-late a simple anesthetic plan that includes adequate preparations for contingencies. Many patients are critically ill, and there is little time intraoperatively to have an assistant search for drugs and equipment. At the same time, the anesthetic plan should not be excessively rigid; when problems are encountered with one technique, one should be ready to change to another without delay. Preparation, organization, and attention to detail permit one to more efficiently deal with unexpected intraoperative problems. The anesthesia machine, monitors, infusion pumps, and blood warmer should all be checked before the patient arrives. Drugs—including anesthetic and vasoactive agents—should be immediately avail-able. Many clinicians prepare one vasoconstrictor and one vasodilator infusion before the start of the procedure.
Cardiac surgery is sometimes associated with large and rapid blood loss, and with the need for multiple drug infusions. Ideally, two large-bore (16-gauge or larger) intravenous catheters should be placed. One of these should be in a large central vein, usually an internal or external jugular or subclavian vein. Central venous cannulations may be accomplished while the patient is awake but sedated or after induc-tion of anesthesia. Studies show no benefit from placing either central venous or pulmonary arterial catheters in awake (versus anesthetized) patients undergoing cardiovascular surgery.
Drug infusions should ideally be given into a central catheter, preferably directly into the catheter or into the injection port closest to the catheter (to minimize dead space). Multilumen central venous catheters and multilumen pulmonary artery cathe-ter introducer sheaths allow for multiple drug infu-sions with simultaneous measurement of vascular pressures. One intravenous port should be dedicated for drug infusions and nothing else; drug and fluid boluses should be administered through another site. The side port of the introducer sheath used for a pulmonary catheter can be used for drug infusions but serves better as a fluid bolus line when a large-bore introducer (9F) is used.Blood should be immediately available for transfusion if the patient has already had amidline sternotomy (a “redo”); in these cases, the right ventricle or coronary grafts may be adherent to the sternum and may be accidentally entered during the repeat sternotomy.
The electrocardiogram (ECG) is continuously monitored with two leads, usually leads II and V5. Baseline tracings of all leads may be recorded on paper for further reference. The advent of monitors with computerized ST-segment analysis and the use of additional monitoring leads (V 4, aVF, and V4R) have greatly improved detection of ischemic epi-sodes, as has the frequent intraoperative use of TEE.
In addition to all basic monitoring, arterial cannula-tion is always performed either prior to or immedi-ately after induction of anesthesia, as the induction period represents a time when major hemodynamic alterations may occur. Radial arterial catheters may occasionally give falsely low readings following ster-nal retraction as a result of compression of the sub-clavian artery between the clavicle and the first rib. They may also provide falsely low values early after CPB due to the opening of atrioventricular shunts in the hand during rewarming. The radial artery on the side of a previous brachial artery cutdown should be avoided, because its use is associated with a greater incidence of arterial thrombosis and wave distor-tion. Obviously, if a radial artery will be harvested for a coronary bypass conduit, it cannot be used as a site for arterial pressure monitoring. Other useful catheterization sites include the ulnar, axillary, and especially brachial and femoral arteries. A backup manual or automatic blood pressure cuff should also be placed on the opposite side for comparison with direct measurements.
Central venous pressure is not terribly useful for diagnosis of hypovolemia but has been customarily monitored in nearly all patients undergoing cardiac surgery. The decision about whether to use a pulmo-nary artery catheter is based on the patient, the pro-cedure, and the preferences of the surgical team. Routine use of a pulmonary artery catheter, once nearly universal in adult cardiovascular practice, is controversial. Pulmonary artery catheterization has declined precipitously in nearly all circumstances except adult cardiac surgery due to lack of evidence of a positive effect on patient outcomes. Left ven-tricular filling pressures can be measured with a left atrial pressure line inserted by the surgeon duringbypass. In general, pulmonary artery catheter-ization has been most of ten used in patientswith compromised ventricular function (ejection fraction <40–50%) or pulmonary hypertension and in those undergoing complicated procedures. The most useful data are pulmonary artery pressures, the pulmonary artery occlusion (“wedge”) pressure, and thermodilution cardiac outputs. Specialized cathe-ters provide extra infusion ports, continuous mea-surements of mixed venous oxygen saturation and cardiac output, and the capability for right ventricu-lar or atrioventricular sequential pacing. Given the risk associated with placing any pulmonary artery catheter, some clinicians opine that it makes sense to restrict pulmonary artery catheterization only to devices that offer these advanced capabilities.
The right internal jugular vein is the preferred approach for intraoperative central venous can-nulation. Catheters placed through the other sites, particularly on the left side, are more likely to kink following sternal retraction (above) and are not nearly as likely to pass into the superior vena cava as those placed through the right internal jugular vein.
Pulmonary artery catheters migrate distally during CPB and may spontaneously wedge with-out balloon inflation. Inflation of the balloon under these conditions can rupture a pulmonary artery causing lethal hemorrhage. Pulmonary artery cath-eters should be routinely retracted 2–3 cm during CPB and the balloon subsequently inflated slowly. If the catheter wedges with less than 1.5 mL of air in the balloon, it should be withdrawn farther.
Once the patient is anesthetized, an indwelling uri-nary catheter is placed to monitor the hourly output. Bladder temperature is often monitored as a mea-sure of core temperature but may not track core tem-perature well with reduced urinary flow. The sudden appearance of reddish urine may indicate excessive red cell hemolysis caused by CPB or a transfusion reaction.
Multiple temperature monitors are usually placed once the patient is anesthetized. Bladder (or rectal), esophageal, and pulmonary artery (blood) tempera-tures are often simultaneously monitored. Because of the heterogeneity of readings during cooling and rewarming, bladder and rectal readings are gener-ally taken to represent an average body temperature, whereas esophageal represents core temperature. Pulmonary artery temperature provides an accurate estimate of blood temperature, which should be the same as core temperature in the absence of active cooling or warming. Nasopharyngeal and tympanic probes may most closely approximate brain tem-perature. Myocardial temperature is often measured directly during CPB.
Intraoperative laboratory monitoring is mandatory during cardiac surgery. Blood gases, hematocrit, serum potassium, ionized calcium, and glucose measurements should be immediately available. The activated clotting time (ACT) approximates the Lee–White clotting time and is used to moni-tor heparin anticoagulation. Some centers routinely use thromboelastography (TEG) to identify causes of bleeding after CPB.
One of the most important actions in intraopera-tive monitoring is inspection of the surgical field. Once the sternum is opened, lung expansion can be observed through the pleura. When the pericardium is opened, the heart (primarily the right ventricle) is visible; thus cardiac rhythm, volume, and contractil-ity can often be judged visually. Blood loss and sur-gical maneuvers must be closely watched and related to changes in hemodynamics and rhythm.
TEE provides valuable information about car-diac anatomy and function during surgery.Two-dimensional, multiplane TEE can detect regional and global ventricular abnormalities, cham-ber dimensions, valvular anatomy, and the presence of intracardiac air. Three-dimensional TEE provides a more complete description of valvular anatomy and pathology. TEE can also be helpful in confirm-ing cannulation of the coronary sinus for cardiople-gia. Multiple views should be obtained from the upper esophagus, mid-esophagus, and transgastric positions in the transverse, sagittal, and in-between planes (Figure 22–2). The two views most com-monly used for monitoring during cardiac surgery are the four-chamber view (Figure 22–3) and the transgastric (short-axis) view (Figure 22–4). Three-dimensional echocardiography offers great promise for better visualization of complex anatomic fea-tures, particularly of cardiac valves. The following represent the most important applications of intra-operative TEE.
color-fl ow imaging ( Figure 22–5 ). Colors are usually adjusted so that fl ow toward the probe is red and fl ow in the opposite direction is blue. TEE also can detect prosthetic valve dysfunction, such as obstruction or regurgitation, and can detect vegetations from endocarditis. The TEE images in the upper mid-esophagus at 40–60° and 110–130° are useful for examining the aortic valve and ascending aorta ( Figure 22–6 ). The
valve annular diameter can also be estimated with reasonable accuracy. Doppler flow across the aortic valve must be measured looking up from the deep transgastric view ( Figure 22–7). The anatomic fea-tures of the mitral valve relevant to TEE are shown in Figure 22–8. The mitral valve is examined from themid-esophageal position, looking at the mitral valve apparatus with and without color in the 0° through 150° views (Figure 22–9). TEE is an invaluable aid to guide and assess the quality of mitral valve repair surgery. The commissural view (at about 60°) is par-ticularly helpful because it cuts across many scallops of the mitral valve.
(ie, looking for abnormal relaxation and restrictive diastolic patterns by checking mitral flow velocity or by measuring movements of the mitral valve annulus using tissue Doppler techniques); and regional systolic function (by assessing wall motion and thickening ab-normalities). Regional wall abnormalities from myo-cardial ischemia often appear before ECG changes. Regional wall motion abnormalities can be classified into three categories based on severity (Figure 22–10): hypokinesis (reduced wall motion), akinesis (no wall motion), and dyskinesis (paradoxical wall motion). The location of a regional wall motion abnormal-ity can indicate which coronary artery is experienc-ing reduced flow. The left ventricular myocardium is supplied by three major arteries: the left anterior descending artery, the left circumflex artery, and the right coronary artery (Figure 22–11). The areas of dis-tribution of these arteries on echocardiographic views
are shown in Figure 22–12. The ventricular short-axis mid view at the mid-papillary muscle level con-tains all three blood supplies from the major coronary arteries.
Computer-processed electroencephalographic (EEG) recordings can be used to assess anesthetic depth during cardiac surgery, and either the processed or “raw” EEG can be used to ensure complete drug-induced electrical silence (for brain protection) prior to circulatory arrest. These recordings are gen-erally not useful in detecting neurological insults during CPB. Progressive hypothermia (or progres-sively deepened anesthesia) is typically associated with EEG slowing, burst suppression, and, finally, an isoelectric recording. Most strokes during CPB are due to small emboli that are not likely to pro-duce changes in the EEG. Artifacts from the CPB roller pump may be seen on the raw EEG (due to piezoelectric effects from compression of the pump tubing) but can usually be identified as such by com-puter processing.
This modality provides noninvasive measurements of blood flow velocity in the middle cerebral artery, which is insonated through the temporal bone. TCD is useful for detecting cerebral emboli. Increased numbers of emboli detected by TCD or Doppler interrogation of the carotid artery have been asso-ciated with an increased risk of postoperative neu-robehavioral dysfunction.
Cardiac operations usually require general anesthe-sia, endotracheal intubation, and controlled ven-tilation. Some centers have used thoracic epidural anesthesia alone for minimally invasive surgery without CPB or combined thoracic epidural with light general endotracheal anesthesia for other forms of cardiac surgery. These techniques have never been popular in North America due to concerns about the risk of spinal hematomas following heparinization, the associated medical–legal consequences, and the limited evidence of an outcome benefit. Other cen-ters use a single intrathecal morphine injection to provide postoperative analgesia.
For elective procedures, induction of general anesthesia should be performed in a smooth, con-trolled (but not necessarily “slow”) fashion often referred to as a cardiac induction. Selection of anesthetic agents is generally less important than the way they are used. Indeed, studies have failed to show differ-ences in long-term outcome with various anesthetictechniques. Anesthetic dose requirements are variable and patient tolerance of inhaled anesthetics generally declines with declining ventricular function. Severely compromised patients should be given anesthetic agents in incremental, small doses. A series of challenges may be used to judge when anesthetic depth will allow intubation without a marked hypertensive response, while also avoiding hypotension from excessive anesthetic dosing. Blood pressure and heart rate are continuously evaluated following unconsciousness, insertion of an oral air-way, urinary catheterization, and tracheal intuba-tion. A sudden increase in heart rate or blood pressure may indicate light anesthesia and the need for more anesthetic prior to the next challenge, whereas a decrease or no change suggests that the patient is ready for the subsequent stimulus. Muscle relaxant is given after consciousness is lost. Reductions in blood pressure greater than 20% gen-erally call for administration of a vasopressor .
The period following intubation is often char-acterized by a gradual decrease in blood pressure resulting from the anesthetized state (often associ-ated with vasodilation and decreased sympathetic tone) and a lack of surgical stimulation. Patients will usually respond to fluid boluses or a vasocon-strictor. Nevertheless, the administration of large amounts of intravenous fluids prior to the bypass may serve to accentuate the hemodilution asso-ciated with CPB (below). Small doses of phenyl-ephrine (25–100 mcg), vasopressin (1–3 units), or ephedrine (5–10 mg) may be useful to avoid excessive hypotension. Following intubation and institution of controlled ventilation; arterial blood gases, hematocrit, serum potassium, and glu-cose concentrations are measured. The baseline ACT (normal <130 s) is best measured after skin incision.
Anesthetic techniques for cardiac surgery have evolved over the years. Successful techniques range from primarily volatile inhalation anesthesia to high-dose opioid totally intravenous techniques. In recent years, total intravenous anesthesia with short-acting agents and combinations of intra-venous and volatile agents have become most popular.
Th is technique was originally developed to circum-vent the myocardial depression associated with older volatile anesthetics, particularly halothane. But pure high-dose opioid anesthesia (eg, fentanyl, 50–100 mcg/kg, or sufentanil, 15–25 mcg/kg) pro-duces prolonged postoperative respiratory depres-sion (12–24 h), is associated with an unacceptably high incidence of patient awareness (recall) during surgery, and often fails to control the hypertensive response to stimulation in many patients with pre-served left ventricular function. Other undesirable effects include skeletal muscle rigidity during induc-tion and prolonged postoperative ileus. Moreover, simultaneous administration of benzodiazepines with large doses of opioids can produce hypotension and myocardial depression. Patients anesthetized with sufentanil (and other shorter acting agents) generally regain consciousness and can be extubated sooner than those anesthetized with fentanyl.
The drive for cost containment in cardiac surgery was a major impetus for development of anesthe-sia techniques with short-acting agents. Although the drugs may be costlier, large economic benefits resulted from earlier extubation, decreased inten-sive care unit (ICU) stays, earlier ambulation, and earlier hospital discharge (“fast-track” manage-ment). One technique employs induction with pro-pofol (0.5–1.5 mg/kg followed by 25–100 mcg/kg/ min), and modest doses of fentanyl (total doses of 5–7 mcg/kg) or remifentanil (0–1 mcg/kg bolus fol-lowed by 0.25–1 mcg/kg/min). Target controlled infusion (TCI) employs software and hardware (computerized infusion pump) to deliver a drug and achieve a set concentration at the effect site basedon pharmacokinetic modeling. For propofol the clinician sets only the patient’s age and weight, and the desired blood concentration on the Diprifusor, a TCI device widely available in countries outside North America. During cardiac surgery, this tech-nique can be used for propofol with a target con-centration of 1.5–2 mcg/mL. Whenever the very short-acting remifentanil is used for painful surgery, provision must be made for postoperative analgesia after its discontinuation.
Renewed interest in volatile agents came about fol-lowing studies demonstrating the protective effects of volatile agents on ischemic myocardium and an increased emphasis on fast-track recovery of car-diac patients. Selection of anesthetic agents is ori-ented to hemodynamic stability as well as early extubation (1–6 h). Propofol (0.5–1.5 mg/kg) or etomidate (0.1–0.3 mg/kg) is often used for induc-tion. Induction usually follows sedation with small doses of midazolam (0.05 mg/kg). Opioids are given in small doses together with a volatile agent (0.5–1.5 minimum alveolar concentration [MAC]) for maintenance anesthesia and to blunt the sym-pathetic response to stimulation. The opioid may be given in small intermittent boluses, by continuous infusion, or both (Table 22–1). To facilitate fast-track management, typical total doses of fentanyl and sufentanil generally do not exceed 15 and 5 mcg/kg, respectively, and some clinicians com-bine much smaller doses of fentanyl or sufentanil with an analgesic dose of hydromorphone or mor-phine administered toward the end of CPB. Some clinicians also administer a low-dose infusion of propofol (25–50 mcg/kg/min) for maintenance.
The major advantage of volatile agents or infusions of remifentanil or propofol, or both, is the ability to change the anesthetic concentration and depth rapidly. Isoflurane, sevoflurane, and desflurane are the most commonly used volatile anesthetics. Early laboratory reports of isoflurane inducing intracoro-nary steal have been overshadowed by later reports of myocardial protection. Isoflurane remains a com-monly used volatile agent. Nitrous oxide is gener-ally not used, particularly during the time interval between cannulation and decannulation, because of its tendency to expand any intravascular air bubbles that may form.
The combination of ketamine with midazolam (or diazepam or propofol) for induction and mainte-nance of anesthesia is a useful technique, particularly in frail patients with hemodynamic compromise. It is associated with stable hemodynamics, reli-able amnesia and analgesia, minimal postoperative respiratory depression, and rare (if any) psychoto-mimetic side effects. Ketamine and midazolam are compatible in solution, and may be mixed together in the same syringe or infusion bag in a 20:1 ratio. For induction, ketamine, 1–2 mg/kg, with mid-azolam, 0.05–0.1 mg/kg, is given as a slow intrave-nous bolus. Anesthesia can then be maintained by infusion of ketamine, 1.3–1.5 mg/kg/h, and mid-azolam, 0.065–0.075 mg/kg/h, or more easily with an inhaled agent. Hypertension following intubation or surgical stimulation can be treated with propofol or a volatile agent.
Muscle relaxation is helpful for intubation, to facilitate sternal retraction, and to prevent patient movement and shivering. Unless airway difficul-ties are expected, intubation may be accomplished after administration of a nondepolarizing mus-cle relaxant. The choice of muscle relaxant in the past was often based on the desired hemodynamic response. Modern, shorter acting agents such as rocuronium, vecuronium, and cisatracurium are commonly used and have almost no hemodynamic side effects of their own. Vecuronium, however, has been reported to markedly enhance bradycardia associated with large doses of opioids, particularly sufentanil. Because of its vagolytic effects, pan-curonium was often used in patients with marked bradycardia who were taking β-blocking agents, Succinylcholine remains appropriate for endotra-cheal intubation, particularly for rapid sequence induction. Judicious dosing and appropriate use ofperipheral nerve stimulator allow fast-tracking with any of these agents.