Maintaining oxygen transport (i.e., oxygen delivery [Do₂]) satisfactory to meet the tissue metabolic needs is the goal of postoperative circulatory control. Oxygen transport is the product of cardiac output (CO) times arterial content of oxygen (Cao₂; i.e., hemoglobin concentration × 1.34 mL oxygen per 1 g hemoglobin × oxygen saturation), and it can be affected in many ways by the cardiovascular and respiratory systems, as shown in Figure 34-1. Low CO, anemia from blood loss, and pulmonary disease can decrease Do₂. Before altering the determinants of CO, including the inotropic state of the ventricles, an acceptable hemoglobin concentration and adequate oxygen saturation (Sao₂) should be provided, enabling increases in CO to provide the maximum available Do₂. As hemoglobin concentration increases, so does blood viscosity and, therefore, the work of the heart to eject the blood. In normal hearts (e.g., athletes), increasing hemoglobin levels to supranormal will increase performance, suggesting that in this setting the increased viscosity is less important than the increase in oxygen-carrying capacity.¹ This has not been examined in patients with cardiac disease. Model analysis of data from animal investigations suggests that maintenance of the hematocrit between 30% and 33% provides the best balance between oxygen-carrying capacity and viscosity.² This analysis also suggests that, in ischemic states, a hematocrit in this range may be desirable. Patients needing continued inotropic or mechanical support of ventricular function beyond the first few postoperative hours, especially those in need of intravascular volume expansion, should probably be transfused to a hematocrit in this range, bearing in mind that blood transfusion has been associated with decreased organ function and increased mortality in critically ill patients. A randomized trial suggested a transfusion threshold of 7 g, rather than 9 g, was associated with at least equivalent outcomes in critically ill patients who did not have acute myocardial infarction (MI) or unstable angina.³^–⁵ Neither of these studies identified cohorts of patients who had undergone cardiac surgery. Wu et al⁵ found transfusion for a hematocrit of 30% or lower in elderly patients with acute MI was associated with better outcome. This study supports the concept that this is the desirable hematocrit, especially in elderly cardiac surgery patients or those experiencing a complicated course.

Figure 34-1 Important factors that contribute to abnormal oxygen transport.

Hypoxemia from any cause reduces Do₂, and acceptable arterial oxygenation (Pao₂) may be achieved with the use of an increased inspired oxygen concentration (Fio₂) or positive end-expiratory pressure (PEEP) in the ventilated patient. Use of PEEP or continuous positive airway pressure in the spontaneously breathing patient may improve Pao₂ by reducing intrapulmonary shunt; however, venous return may be reduced, causing a decrease in CO, with DO₂ decreased despite an increased Pao₂ (Figure 34-2).⁶ It is important to measure CO as PEEP is applied. Intravascular volume expansion may be used to offset this damaging effect of PEEP⁷ (see Chapter 35).

Figure 34-2 Mean values ± standard error of the mean of arterial oxygen tension (Pao₂), intrapulmonary shunt (Q_S/Q_T), total static compliance, and oxygen transport, measured at the level of positive end-expiratory pressure (PEEP), resulting in maximum oxygen transport (“best PEEP”) compared with values obtained at 3 and 6 cm H₂O of PEEP above and below that level in 15 patients with acute respiratory failure. *P < 5.005 versus the value at best PEEP.

(From Suter PM, Fairley HB, Isenbery MD: Optimum end-expiratory pressure in patients with acute pulmonary failure. N Engl J Med 292:284, 1975.)

In patients with marginal arterial oxygenation, pulmonary function must be monitored closely to allow prompt therapy to be undertaken for abnormalities. Measurements of airway resistance and respiratory system compliance should be made. When resistance is increased, treatment of bronchospasm may improve the Pao₂ and CO, because decreases in intrathoracic pressure improve venous return. Treatment of lung overinflation may decrease pulmonary vascular resistance (PVR), benefiting right ventricular (RV) function.⁸ If compliance is decreased, application of PEEP or continuous positive airway pressure may help promote re-expansion of atelectatic areas and move the tidal volume to a more compliant section of the pressure-volume relation of the respiratory system⁹ (Figure 34-3). This will reduce the work expended by the patient during spontaneous efforts and may reduce PVR¹⁰ (see Chapter 35).

Figure 34-3 Pressure-volume diagram of elastic and resistive (nonelastic) work done on noncompliant lungs.

Breathing at ambient airway pressure and low lung volume (by T tube) (I) versus breathing with continuous positive airway pressure (CPAP) at increased lung volume (II). Solid line BCHJ is the elastic pressure-volume curve for the lung, determined by measuring transpulmonary pressures at the instant of zero flow. Hatched areas represent nonelastic work (BIC and HI′J), whereas elastic work is represented by BCD and HJK. With CPAP, both types of work are reduced. Without CPAP, the area ABDF is additional elastic work partly done by the patient and partly by elastic recoil of the chest wall, whereas with CPAP, the additional work represented by MHKL is mostly done by the CPAP system.

(From Katz JA, Marks JD: Inspiratory work with and without continuous positive airway pressure in patients with acute respiratory failure. Anesthesiology 63:598, 1985.)

Unexplained hypoxemia may be caused by right-to-left intracardiac shunting, most commonly by a patent foramen ovale. This is most likely to occur when right-sided pressures are abnormally increased; an example is the use of high levels of PEEP.¹¹ If suspected, echocardiography should be performed, and therapy to reduce right-sided pressures should be initiated.

Patients with pulmonary disease may experience dramatic worsening of oxygenation when vasodilator therapy is started because of release of hypoxic pulmonary vasoconstriction in areas of diseased lung.¹² Although CO may be increased, the worsening in Cao₂ will result in a decrease in Do₂. Reduced dosage of direct-acting vasodilators or trials of different agents may be indicated.

When Do₂ cannot be increased to an acceptable level as judged by decreased organ function or development of lactic acidemia, measures to decrease oxygen consumption (o₂) may be taken while awaiting improvement in cardiac or pulmonary function. For example, sedation and paralysis may buy time to allow reversible postoperative myocardial dysfunction to improve.

Temperature

Patients often are admitted to the intensive care unit (ICU) after cardiac surgery with core temperatures less than 35° C, especially after off-pump cardiac surgery. The typical pattern of temperature change during and after cardiac surgery and the hemodynamic outcomes are illustrated in Figure 34-4. Decreases in temperature after CPB occur, in part, because of redistribution of heat within the body and because of heat loss. Noback and Tinker¹³ found that administration of nitroprusside and the use of high flows (> 2.2 L/min/m²) during rewarming on CPB could improve the uniformity of rewarming and reduce this afterdrop from 4° C to about 2° C. Monitoring of body sites other than the blood and brain (e.g., urinary bladder, tympanic membrane temperatures) can help provide more complete rewarming, but the body temperature usually declines after bypass, especially when difficulties are encountered and the chest remains open for an extended period, and some degree of hypothermia is an almost unavoidable result.^14,¹⁵ Intraoperative use of newer forced air warming blankets or cutaneous gel pads¹⁶ can help reduce the temperature loss during and after surgery.

Figure 34-4 Nasopharyngeal temperature during and after cardiac surgery.

(1) Core (i.e., blood) cooling on cardiopulmonary bypass (CPB). (2) Core warming on CPB. (3) Afterdrop in temperature (T) after separation from CPB. (4) Rewarming after admission to the intensive care unit (ICU). Systemic vascular resistance (SVR) is increased, and carbon dioxide production (Vco₂) and oxygen consumption (o₂) are decreased on admission to the ICU because of residual hypothermia. During rapid rewarming, SVR decreases and Vco₂ and o₂ increase, which can cause marked cardiac and ventilatory instability. OR, operating room.

(From Sladen RN: Management of the adult cardiac patient in the intensive care unit. In Ream AK, Fogdall RP [eds]: Acute Cardiovascular Management: Anesthesia and Intensive Care. Philadelphia, JB Lippincott, 1982, p 495.)

The normal thermoregulatory and metabolic responses to hypothermia remain intact after cardiac surgery, resulting in peripheral vasoconstriction that contributes to the hypertension commonly seen early in the ICU.¹⁷ As temperature decreases, CO is decreased because of bradycardia, whereas oxygen consumed per beat is actually increased.¹⁸ Coagulation, platelet, and immune functions are impaired by hypothermia to potentiate postoperative bleeding and infection.¹⁹^–²¹ Other adverse consequences of postoperative hypothermia are large increases in o₂ and CO₂ production during rewarming.²² When patients cannot increase CO (i.e., O₂ delivery), the effects of this large increase in o₂ include mixed venous desaturation and metabolic acidosis. Unless end-tidal carbon dioxide is monitored or arterial blood gases are analyzed often to show the increased CO₂ production and guide increases in ventilation, hypercarbia will occur, causing catecholamine release, tachycardia, and pulmonary hypertension.²³ These effects of rewarming are most intense when patients shiver.²⁴ Shivering may be treated effectively with meperidine, which lowers the threshold for shivering. Muscle relaxation may provide more stable hemodynamics than meperidine but needs accompanying sedation to avoid having an awake and paralyzed patient.^25,²⁶

As the temperature increases, usually to about 36° C, the vasoconstriction and hypertension are replaced by vasodilation, tachycardia, and hypotension, even without hypercarbia. Often, over minutes, a patient who needs vasodilators for hypertension transforms into one requiring vasopressors or large volumes of fluid for hypotension. Volume loading during the rewarming period can help to reduce the rapid swings in blood pressure (BP) that may occur. It is important to recognize when these changes result from changes in body temperature to avoid attributing them to other processes that may call for different therapy.

Assessment of the circulation

Physical Examination

Surgical dressings, chest tubes attached to suction, fluid in the mediastinum and pleural spaces, peripheral edema, and temperature gradients can distort or mask information obtained by the classic techniques of inspection, palpation, and auscultation in the postoperative period. However, the physician should not be deterred from applying these basic techniques in view of the potential benefit. Physical examination may be of great value in diagnosing gross or acute pathology, such as pneumothorax, hemothorax, or acute valvular insufficiency, but it is of limited value in diagnosing and managing ventricular failure. For example, in the critical care setting, experienced clinicians (e.g., internists) using only physical findings often misjudge cardiac filling pressures by a large margin.²⁷ Low CO, in particular, is not consistently recognized by clinical signs, and systemic BP does not correlate with CO after cardiac surgery. Oliguria and metabolic acidosis, classic indicators of a low CO, are not always reliable because of the polyuria induced by hypothermia, oxygen debts induced during CPB causing acidosis, and medications or fluids given during or immediately after bypass.²⁸

Although clinicians are taught that the adequacy of CO can be assessed by the quality of the pulses, capillary refill, and peripheral temperature, there is no relationship between these indicators of peripheral perfusion and CO or calculated systemic vascular resistance (SVR) in the postoperative period.²⁹ By the first postoperative day, there is a crude correlation between peripheral temperature and cardiac index (CI; r = −0.60). Many patients arrive in the ICU in a hypothermic state, and residual anesthetic agents can decrease the threshold for peripheral vasoconstriction in response to this condition.³⁰ A patient’s extremities may therefore remain warm despite a hypothermic core or a decreasing CO. Even after temperature stabilization on the first postoperative day, the relation between peripheral perfusion and CO is too crude to be used for hemodynamic management.

Invasive Monitoring

Concepts regarding invasive monitoring with a pulmonary artery catheter (PAC) have been revolutionized in the past decade because of several studies in a variety of settings that fail to show a benefit from its use. In addition, there is a poor relation between filling pressures and end-diastolic volume, stroke volume (SV), or volume responsiveness. A recent review of patients admitted to medical ICUs in the United States demonstrated a reduction of more than 40% in PAC use in the 10 years before 2004.³¹ The same trend was evident in surgical patients, including those undergoing cardiac surgery. The PAC rarely is used in cardiac surgery patients in many other countries. Greater availability of high-quality bedside echocardiography, often performed by intensivists, has made this modality a technique of choice in the postoperative period. Measures of volume responsiveness in mechanically ventilated patients, such as pulse pressure or SV variability (from arterial waveform analysis devices), are widely recognized as more sensitive and specific indicators of the need for intravascular volume expansion than filling pressures.³²

Despite the lack of a proven benefit with PAC use, many patients in North America continue to have this monitor placed for cardiac surgery. Cardiac anesthesiologists believe the lack of evidence about the PAC may reflect the lack of a well-designed randomized trial. There are no such trials in cardiac surgery patients, probably attesting to the reluctance of cardiac surgeons and anesthesiologists to manage their patients without what they consider to be important information. After surgery, many cardiac surgical centers do not have in-house physicians, and surgeons believe the “objective” PAC data obtained over telephone are valuable. As less invasive tools such as echocardiography or arterial waveform analysis devices become better known and available, it seems likely that PAC use will diminish in cardiac surgery patients.

Specialized PACs have been developed that permit continuous mixed venous oxygen saturation (So₂) monitoring, continuous CO measurement, calculation of RV volumes and ejection fraction, or have either imbedded electrodes or channels to pass atrial or ventricular pacing wires. The ability to pace through a PAC is particularly valuable in patients undergoing “minimally invasive” procedures in which the surgeon does not have adequate access to the heart to place epicardial leads. The Svo₂ catheter helps evaluate the adequacy of Do₂ and allows continuous assessment of the response to therapy, which may affect Do₂ or o₂ (e.g., PEEP therapy). The trend in the Svo₂ may function as an early warning signal of worsening in the oxygen supply/demand relation as Do₂ declines or VO₂ increases. In the postoperative period, the Svo₂ does not correlate with CO because the latter is only one of the factors in the oxygen supply/demand relation.³³ On the continuous CO catheter, a wire coil warms the blood, passing by it at time intervals determined by an algorithm, and the measured changes in temperature at the tip of the catheter are used to provide a continuous display of the CO. Although the CO displayed needs gathering of information over several minutes and is therefore not as quick as conventional thermodilution, and it does not provide beat-to-beat SV, it avoids having to give injected volumes to the patient (which can add up to a significant amount every 24 hours) and provides trends that may give earlier warning than intermittent injections. The “volumetric” PAC-computer system (REF-1; Edwards Lifesciences, Irvine, CA) uses a high-sensitivity thermistor to permit calculation of accurate right-sided volumes³⁴ (see Chapter 14).

Echocardiography

Echocardiography is the technique of choice for acute assessment of cardiac function. Just as transesophageal echocardiography (TEE) has become essential for intraoperative management in various conditions, several studies document its utility in the postoperative period in the presence and absence of the PAC.³⁵^–³⁸ It provides information that may lead to urgent surgery or prevent unnecessary surgery, gives important information about cardiac preload, and can detect acute structural and functional abnormalities. Although transthoracic echocardiography can be performed more rapidly in this setting, satisfactory images can be obtained only in about 50% of patients in the ICU³⁹ (see Chapters 11 to 14).

Postoperative myocardial dysfunction

Studies using hemodynamic, nuclear scanning, and metabolic techniques have documented worsening in cardiac function after coronary artery bypass grafting (CABG) surgery.⁴⁰^–⁵³ Although improvements in myocardial protection, surgical techniques, and operative care have been reported, similar incidences of early biventricular dysfunction (90%) were reported between 1979 and 1990. All of these studies showed significant declines in left ventricular (LV) or biventricular (when measured) function in the first postoperative hours, with gradual return to preoperative values by 8 to 24 hours. In one study, this decline was evident in only half the patients⁴⁴; but in the other studies, more than 90% of patients showed at least a transient decrease in function. Decreased ventricular performance at normal or increased filling pressures occurs, suggesting decreased contractility. Similarly, “flattening” of the ventricular function curves is usually obvious, suggesting that preload expansion much greater than 10 mm Hg for CVP or 12 mm Hg for pulmonary capillary wedge pressure is of little benefit. In the classic study by Mangano,⁴⁵ patients with an LV ejection fraction of less than 0.45 or ventricular dyssynergy showed more marked and prolonged dysfunction than did those patients with normal ventricles.

Satisfactory myocardial protection is important to prevent postoperative dysfunction. In off-pump surgery, the idea is to preserve coronary perfusion; but during mechanical manipulation, changes in CO and BP can occur. For CABG with CPB, most surgeons use some combination of hypothermia and crystalloid or blood cardioplegia to arrest the heart and reduce its metabolism. Although there is little consensus that any one technique is preferable in all circumstances, cold intermittent crystalloid cardioplegia with systemic hypothermia is the most widely used technique clinically and in the reported studies. Salerno et al ⁵⁴ recommended continuous, warm, retrograde blood cardioplegia without systemic hypothermia. Mullen et al⁵⁰ suggested that blood cardioplegia had at least short-term benefit with less myocardial damage and better function; however, other studies of blood cardioplegia have showed mixed results⁴⁷^–⁵² (see Chapters 28 and 29).

Other proposed factors that contribute to postoperative ventricular dysfunction include myocardial ischemia,⁵⁵ residual hypothermia,^46,⁴⁷ preoperative medications such as β-adrenergic antagonists,⁵³ and ischemia/reperfusion injury (Box 34-1). Inflammatory cell activation from cytokine generation, upregulation of neutrophil adhesion molecules with neutrophil activation, oxygen free radical formation, and lipoperoxidation after ischemia/reperfusion injury may be important pathways accounting for the dysfunction. Multiple studies have showed the importance of limiting myocardial ischemia/reperfusion injury.^56,⁵⁷ Breisblatt et al⁴⁰ observed the timing of ventricular dysfunction, and recovery after CPB for CABG was similar to what had been suggested in animal models of reperfusion injury.⁵⁸^–⁶⁰ This nadir at 4 hours corresponds to the peak in cytokine levels. Cytokines can release nitric oxide from endothelium, which produces myocardial depression. Data evaluating complement inhibition with pexelizumab in improving outcomes represent a novel strategy⁶¹ (see Chapter 8).

Postoperative myocardial ischemia

Although intraoperative myocardial ischemia has been a focus, studies have shown that ischemia often occurs after surgery and is associated with adverse cardiac outcomes. Leung et al ⁵⁵ found the electrocardiographic (ECG) and segmental wall motion abnormality evidence of ischemia early after surgery in up to 40% of patients undergoing CABG surgery. Postbypass segmental wall motion abnormalities were significantly associated with adverse outcomes (e.g., MI, death; Figure 34-5). Surprisingly, these abnormalities most often appeared in the regions of the heart that had been revascularized. Hemodynamic changes rarely preceded ischemia; however, postoperative heart rates (HRs) were, as reported in other studies, significantly greater than intraoperative or preoperative values. Jain et al⁶² found major ECG changes in the 8 hours after cross-clamp release in 58% of CABG patients, and these changes were independent predictors of perioperative MI. Whether such changes occur because of surgery-reperfusion or events after CPB is not known. These findings do suggest that monitoring for ischemia must continue after revascularization. It may be that early recognition and treatment of ischemia or prophylactic medication can help prevent or reduce myocardial ischemia and dysfunction occurring after CABG surgery (see Chapters 6, 10, 12, 15, and 18).

Figure 34-5 Sequential changes in wall motion score, as measured by transesophageal echocardiography, in patients undergoing elective coronary artery bypass grafting. Score is defined as follows: 0 = normal; 1 = mild hypokinesis; 2 = severe hypokinesis with myocardial thickening; 3 = akinesis; and 4 = dyskinesis. Adverse outcomes were myocardial infarction, death, or congestive heart failure. The time periods were as follows: 1 = after tracheal intubation; 2 = before incision; 3 = after incision; 4 = before sternotomy; 5 = after sternotomy; 6 and 7 = internal mammary dissection; 8 = after pericardiotomy; 9 = immediately before bypass; 10 = unclamping of aortic side-clamp; 11 to 14 = off bypass; 15 = after chest closure; 16 = skin closure; 17 to 20 = intensive care unit (ICU) for the first 4 hours. Shaded area indicates the bypass period.

(From Leung JM, O’Kelly B, Browner WS, et al: Prognostic importance of postbypass regional wall motion abnormalities in patients undergoing coronary artery bypass graft surgery. Anesthesiology 71:16, 1989.)

Early recovery, or fast-tracking, of the cardiac surgical patient has led to some concern that ischemia will occur as patients awaken early after surgery in pain, especially because Mangano et al ⁶³ showed that sedation with a sufentanil infusion could reduce ischemia in this period. A randomized study by Cheng et al⁶⁴ dispelled this concern because awakening and extubation within 6 hours of CABG were not associated with more CK-MB (isoenzyme of creatine kinase with muscle and brain subunits) release or ECG changes than overnight ventilation. Wahr et al⁶⁵ showed that even with the use of propofol sedation, hemodynamic episodes (significant changes in HR and BP) were common in the 12 hours after surgery, and ST-segment changes occurred in 12% to 13% of patients.

Therapeutic interventions

Therapeutic interventions for postoperative biventricular dysfunction include the standard concerns of managing low CO states by controlling the HR and rhythm, providing an acceptable preload, and adjusting afterload and contractility. In most patients, pharmacologic interventions can be rapidly weaned or stopped within the first 24 hours after surgery.

Postoperative Arrhythmias

Patients with preoperative or newly acquired noncompliant ventricles need a correctly timed atrial contraction to provide satisfactory ventricular filling, especially when they are in sinus rhythm before surgery (see Chapters 4, 5, 10, 19, and 25). Although atrial contraction provides around 15% to 20% of ventricular filling, this may be more important in postoperative patients, when ventricular dysfunction and reduced compliance may be present. For example, in medical patients with acute MI, atrial systole contributed 35% of the SV.⁶⁶ The SV is relatively fixed in patients with ventricular dysfunction, and the HR is an important determinant of CO. Rate and rhythm disorders need to be corrected when possible, using epicardial pacing wires. Approaches to postoperative rate and rhythm disturbances are listed in Table 34-1. The use of a PAC with atrial or ventricular pacing electrodes or use of lumens for pacing wires can facilitate temporary pacing if epicardial wires are not functioning. Failing that, temporary transvenous pacing wires can be placed (see Chapter 25).

TABLE 34-1 Postoperative Rate and Rhythm Disturbances

Disturbance	Usual Causes	Treatments
Sinus bradycardia	Preoperative/intraoperative β-blockade	Atrial pacingβ-AgonistAnticholinergic
Heart block (first, second, and third degree)	IschemiaSurgical trauma	Atrioventricular sequential pacingCatecholamines
Sinus tachycardia	Agitation/pain	Sedation/analgesia
	Hypovolemia	Volume administration
	Catecholamines	Change or stop drug
Atrial tachyarrhythmias	Catecholamines	Change or stop drug
	Chamber distentionElectrolyte disorder (hypokalemia, hypomagnesemia)	Treat underlying cause (e.g., vasodilator, give K⁺/Mg²⁺)May require synchronized cardioversion or pharmacotherapy
Ventricular tachycardia or fibrillation	IschemiaCatecholamines	CardioversionTreat ischemia, may require pharmacotherapy
		Change or stop drug

Later in the postoperative period (days 1 through 3), supraventricular tachyarrhythmias become a major problem, with atrial fibrillation (AF) predominating. The overall incidence rate is between 30% and 40%, but with increasing age and valvular surgery, the incidence rate may be in excess of 60%.⁶⁷ There are many reasons for this, including genetic factors, inadequate atrial protection during surgery, electrolyte abnormalities, change in atrial size with fluid shifts, epicardial inflammation, stress, and irritation.⁶⁸ Randomized trials of off-pump coronary artery bypass have found a similar incidence of postoperative AF compared with on-pump CABG.^69,⁷⁰

Advanced age, a history of AF, and valvular heart surgery are the most consistently identified risk factors for AF.⁶⁸ Because AF is difficult to treat and potentially increases duration and cost of hospitalization, there is a great interest in effective therapy and prophylaxis.⁶⁷ Many studies have showed that β-blockade significantly reduces the incidence of postoperative AF and that withdrawal of β-blockers in patients receiving them before surgery is an important risk factor. Guidelines published by the American Heart Association, American College of Cardiology, and North American Society of Pacing and Electrophysiology recommend administration of β-blockers to prevent postoperative AF if there are no contraindications.⁷¹ Sotalol, which also has some class III actions, is also effective⁷² and is currently available in the intravenous form in North America (see Chapters 4 and 10).

Several studies have examined the use of amiodarone for prophylaxis or treatment and report both oral and intravenous amiodarone. Intravenous amiodarone is most often used in clinical practice because “acute” loading with oral therapy often is not feasible. Two pivotal studies of amiodarone deserve mention.

In the PAPABEAR study, oral amiodarone (10 mg/kg daily) or placebo was given 6 days before surgery through 6 days after surgery (13 days).⁷³ Atrial tachyarrhythmias occurred in fewer amiodarone patients (48/299; 16.1%) than in placebo patients (89/302; 29.5%) overall, in patients younger than 65 years (19 [11.2%] vs. 36 [21.1%], in patients older than 65 years (28 [21.7%] vs. 54 [41.2%]), in patients who had CABG surgery only (22 [11.3%] vs. 46 [23.6%]), in patients who had valve replacement/repair surgery (25 [23.8%] vs. 44 [44.1%]), in patients who received preoperative β-blocker therapy (27 [15.3%] vs. 42 [25.0%]), and in patients who did not receive preoperative β-blocker therapy (20 [16.3%] vs. 48 [35.8%]), respectively. Postoperative sustained ventricular tachyarrhythmias occurred less frequently in amiodarone patients (1/299; 0.3%) than in placebo patients (8/302; 2.6%) (P = 0.04).⁷³

In another study, Guarnieri et al ⁷⁴ evaluated 300 patients randomized in a double-blind fashion to intravenous amiodarone (1 g/day for 2 days) or to placebo immediately after cardiac surgery. The primary end points of the trial were incidence of AF and length of hospital stay. AF occurred in 67 (47%) of 142 patients on placebo versus 56 (35%) of 158 on amiodarone (P = 0.01). Length of hospital stay for the placebo group was 8.2 ± 6.2 days, and 7.6 ± 5.9 days for the amiodarone group.

After AF or other supraventricular arrhythmias develop, treatment often is urgently needed for symptomatic relief or hemodynamic benefit. The longer a patient remains in AF, the more difficult it may be to convert, and the greater the risk for thrombus formation and embolization.^68,⁷² Treatable underlying conditions such as electrolyte disturbances or pain should be corrected while specific pharmacologic therapy is being instituted. Paroxysmal supraventricular tachycardia (uncommon in this setting) can be abolished or converted by intravenous adenosine, and atrial flutter can sometimes be converted by overdrive atrial pacing by temporary wires placed at the time of surgery. Electrical cardioversion may be necessary if hypotension is caused by the rapid rate; however, atrial arrhythmias recur in this setting.⁶⁷ Rate control for AF or flutter can be achieved with various atrioventricular nodal blocking drugs, and conversion is facilitated by many of these drugs as well. Table 34-2 summarizes the various treatment modalities for supraventricular arrhythmias. If conversion to sinus rhythm does not occur, electrical cardioversion in the presence of antiarrhythmic drug therapy should be attempted or anticoagulation with warfarin started (see Chapters 4, 10, and 25).

TABLE 34-2 Treatment Modalities for Supraventricular Arrhythmias

Treatment	Specifics*	Indications
Overdrive pacing by atrial wires^†	Requires rapid pacer (up to 800/min); start above arrhythmia rate and slowly decrease	PAT, atrial flutter
Adenosine	Bolus dose of 6–12 mg; may cause 10 seconds of complete heart block	AV nodal tachycardiaBypass-tract arrhythmia
		Atrial arrhythmia diagnosis
Amiodarone	150 mg IV over 10 minutes, followed by infusion	Rate control/conversion to NSR in atrial fibrillation/flutter
β-Blockade	Esmolol, up to 0.5 mg/kg load over 1 minute, followed by infusion if tolerated	Rate control/conversion to NSR in atrial fibrillation/flutter
	Metoprolol, 0.5–5 mg, repeat effective dose q4-6h	Rate control/conversion to NSR in atrial fibrillation/flutter
	Propranolol, 0.25–1 mg; repeat effective dose q4h^‡
	Labetalol, 2.5–10 mg; repeat effective dose q4hr^‡	Conversion of atrial fibrillation/flutter to NSR
	Sotalol, 40–80 mg PO q12h	Conversion of PAT to NSR
Ibutilide	1 mg over 10 minutes; may repeat after 10 minutes	Rate control/conversion to NSR in atrial fibrillation/flutter
Verapamil	2.5–5 mg IV, repeated PRN^‡
Diltiazem	0.2 mg/kg over 2 minutes, followed by 10–15 mg/hr^§	Rate control/conversion to NSR in atrial fibrillation/flutter
Procainamide	50 mg/min up to 1 g, followed by 1–4 mg/min	Rate control/conversion to NSR in atrial fibrillation/flutter
		Prevention of recurrence of arrhythmias
		Treatment of wide-complex tachycardias¶¶
Digoxin^¶	Load of 1 mg in divided doses over 4–24 hours**; may give additional 0.125-mg doses 2 hours apart (3–4 doses)	Rate control/conversion to NSR in atrial fibrillation/flutter
Synchronized cardioversion	50–300 J (external); most effective with anteroposterior patches	Acute tachyarrhythmia with hemodynamic compromise (usually atrial fibrillation or flutter)

AV, atrioventricular; IV, intravenously; NSR, normal sinus rhythm; PAT, paroxysmal atrial tachycardia; PO, orally; PRN, as needed; SVT, supraventricular tachycardia.

* See specific drug monographs for full description of indications, contraindications, and dosage. Doses are for intravenous administration; use lowest dose and administer slowly in patients with hemodynamic compromise.

† Verify pacer is not capturing ventricle.

‡ Infusion may provide better control. This drug is less useful than diltiazem because of myocardial depression.

§ Limited experience; may cause less hypotension than verapamil.

¶¶ When diagnosis is unclear (ventricular vs. supraventricular) and there is no acute hemodynamic compromise (i.e., cardioversion not indicated).

¶ Rate of administration depending on urgency of rate control.

** Less useful than other drugs because of slow onset and modest effect.

In summary, AF is a frequent complication of cardiac surgery, but the incidence can be significantly reduced with suitable prophylactic therapy. β-Adrenergic blockers should be administered to patients without contraindication, and prophylactic amiodarone can be considered for patients at high risk for postoperative AF. Patients who are poor candidates for β-blockade may not tolerate sotalol, whereas amiodarone does not have this limitation. More studies need to be performed to better assess the role of prophylactic therapy in off-pump cardiac surgery. After AF occurs, there is a high incidence of recurrence, so treatment with specific continuing pharmacologic therapy is usually necessary.

Preload

The Frank-Starling law states that myocardial work increases as the resting length of the myocardial fiber increases.⁷⁵ In vivo, this implies that SV will increase with increasing end-diastolic volume, although there is a limit at which SV reaches a plateau (and possibly decreases), with further increases in end-diastolic volume caused by excessive muscle stretch. In the normal myocardium, the Frank-Starling mechanism is the most important mechanism for increasing CO, and hypovolemia is a common cause of decreased CO and hypotension in the perioperative period. Assessment of preload is probably the single most important clinical skill for managing hemodynamic instability. Preload rapidly changes in the postoperative period because of bleeding, spontaneous diuresis, vasodilation during warming, the effects of positive-pressure ventilation and PEEP on venous return, capillary leak, and other causes.

Direct assessment of preload is clinically feasible using echocardiography. Several studies have demonstrated a fair-to-good correlation between echocardiographic and radionuclide measures of end-diastolic volume, and there is a good correlation between end-diastolic area by TEE and SV.⁷⁶^–⁷⁹ Although the use of echocardiography to assess preload must always be tempered by the realization the clinician is viewing a two-dimensional image of a three-dimensional object, this is the most direct technique clinically available. Increased awareness of the value of TEE in the ICU and increased availability of echocardiography in general have made this modality a first choice in acute assessment of preload in the setting of unexplained or refractory hypotension. Without echocardiography, pressure measurements are used as surrogates for volume measurements. For example, in the absence of mitral valve disease, left atrial pressure is almost equal to LV end-diastolic pressure, and pulmonary artery occlusion pressure (PAOP) is almost equivalent to these two pressures. In patients without left atrial pressure catheters, the PAOP or, when shown to be equivalent to this latter number, the pulmonary artery diastolic (PA_d) pressure is used (see Chapters 5 and 14).

The use of PAOP as a measure of preload may be misleading in various settings, including after cardiac surgery, when changes in pressure may not accurately reflect changes in ventricular end-diastolic volume. Studies by Ellis et al ⁸⁰ and Calvin et al⁸¹ suggest that fluid therapy in postoperative patients could cause a large increase in LV end-diastolic volume, with minimal or no change in PAOP. Mangano et al^45,⁸² reported that fluid loading after CPB can uncover LV dysfunction and have also shown there is little benefit to be derived from exceeding a PAOP of about 12 mm Hg. Whether this is secondary to an open pericardium, which allows overdistention of the left ventricle, the use of PEEP causing RV distention, or other factors is unclear. Breisblatt et al⁸³ evaluated changes in ventricular pressure-volume relations after CABG surgery while keeping a low-to-normal PAOP (10 to 15 mm Hg). The increase seen in LV end-diastolic volume was not significant enough to explain the degree of ventricular dysfunction they noted. Bouchard et al⁸⁴ compared ventricular performance assessments from the PAC (i.e., LV stroke work index) with fractional area change and regional wall motion score index from TEE in 60 patients during and after cardiac surgery. They found no correlation between LV stroke work index and fractional area change and postulated that changes in ventricular compliance, loading conditions, and ventricular function alter the pressure-volume relation of the left ventricle in a manner that leads to discordant interpretations between the two techniques.

When ventricular compliance is normal and the ventricle is not distended, small changes in end-diastolic volume are usually accompanied by small changes in end-diastolic pressure. In patients with noncompliant ventricles from preexisting congestive heart failure (HF), chronic hypertrophy resulting from hypertension or valvular disease, postoperative MI, or ventricular dysfunction, small increases in ventricular volume may produce rapid increases in end-diastolic pressure, requiring therapeutic intervention.⁸³^–⁸⁹ Increased intraventricular pressure will increase myocardial oxygen demand (Mo₂) and decrease subendocardial coronary blood flow.⁹⁰ Myocardial ischemia may be the result. Increases in LV end-diastolic pressure are transmitted to the pulmonary circulation, causing congestion and, possibly, hydrostatic pulmonary edema. Although PAOP or left atrial pressure may not always show true preload, there are still good reasons to monitor them. The periods when patients are at particular risk for increases in end-diastolic pressure include awakening, endotracheal suctioning, and rapid-volume resuscitation. If myocardial ischemia or acute HF occurs, sudden increases in the end-diastolic pressure may result.

Many drugs may be used to reduce cardiac preload. Direct-acting vasodilators, especially nitroglycerin, increase venous capacitance, decreasing end-diastolic volume and pressure.⁸⁹^–⁹¹ Intravenous furosemide, besides its diuretic effect, increases venous capacitance.⁹² Diuretics are important to help remove the fluid that is mobilized in the days after surgery. In patients who may not tolerate the acute volume loss that is induced by the loop diuretics, a furosemide infusion allows a more gradual diuresis. Such infusions have been shown to be effective in patients with renal dysfunction.^93,⁹⁴ In a patient who appears refractory to diuresis, it is important to evaluate the circulatory state. Such refractoriness may suggest that a renal insult has been suffered, but it may also suggest underperfusion of the kidneys. In the latter case, the preload must be kept at the upper range of normal and CO augmented.

A new agent used in treatment of acute decompensated left-heart failure is recombinant B-type natriuretic peptide (nesiritide). The mechanism of action occurs by specific cell-surface receptors, stimulation of which increases levels of intracellular cyclic guanosine monophosphate (cGMP). The physiologic effects are mainly vasodilation, natriuresis, and renin inhibition. The result is balanced vasodilation, reducing preload and afterload, while simultaneously increasing SV and CO and promoting diuresis.⁹⁵ Nesiritide has not been studied widely in the cardiac surgery population, but it has been used anecdotally in patients with poor ventricular function and high filling pressures, pulmonary hypertension, or RV failure.⁹⁶ A related drug, human atrial natriuretic peptide, was shown in a small, randomized trial to reduce the need for dialysis and to improve dialysis-free survival after complicated cardiac surgery.⁹⁷

Angiotensin-converting enzyme (ACE) inhibitors also can cause venodilation and reduce preload. Alternatively, opioids or benzodiazepines, or both, used to reduce endogenous catecholamine release, should be considered in the patient who needs mechanical ventilation. Morphine causes histamine release, which directly induces venodilation. In the patient with oliguria and renal failure who is fluid overloaded, peritoneal dialysis, hemodialysis, or continuous hemofiltration or dialysis may be needed.^98,⁹⁹

Contractility

Contractility is a well-defined concept in vitro, where it can be measured by the velocity of shortening of isolated muscle strips. However, it has been more complex to quantify the contractility of the intact heart because it has been difficult to find a variable to measure contractility that is also independent of preload and afterload. The pioneering work of Suga and Sagawa ^100,¹⁰¹ has shown that contractility can be measured by the end-systolic elastance of the ventricle, defined as

in which P_ES is the end-systolic pressure, V_ES is the end-systolic volume, and V₀ is a dead-space term. E_ES is strictly determined by evaluating ventricular pressure-volume loops for different preloads or afterloads and by defining end-systole as the point in time at which the time-varying elastance is maximal.¹⁰² The slope of the line connecting the points at end-systole is E_ES. This parameter varies with changes in inotropic state but is nearly independent of preload and afterload. Routine measurement of end-systolic elastance is not clinically feasible. However, consideration of the above definition underscores the utility of TEE for the qualitative evaluation of contractility in the clinical setting. A decrease in contractility is manifested as some combination of a decrease in pressure or an increase in V_ES (i.e., a decrease in SV). End-systolic volume can be estimated using TEE. A large V_ES (implying a low ejection fraction) with a low or normal BP suggests a low value for E_ES and poor contractility. If BP is high, a large value of V_ES may be seen even if contractility is normal. By interpreting end-diastolic volume, end-systolic volume, and ejection fraction in the context of BP, an assessment of contractility is possible with TEE (see Chapters 5 and 12 to 14).

An alternative measure of contractility is the preload-recruitable stroke work, which is the slope of the line relating stroke work to preload.¹⁰³ In the operating room or ICU, it often is estimated by the extent of the increase in CO that accompanies an increase in preload and is not dependent on the availability of TEE. If preload is increased by a change in patient position, BP can be a surrogate for CO because it is unlikely that SVR will change in the short time needed for the position change. A change in BP, therefore, is proportional to the change in CO.

Therapy for decreased contractility should be directed toward correcting any reversible causes, such as myocardial depressants, metabolic abnormalities, or myocardial ischemia. If the cause of depressed myocardial contractility is irreversible, positive inotropic agents may be necessary to keep a CO satisfactory to support organ function (see Chapters 5, 10, 28, and 32).

Afterload

Afterload is a concept that is well defined in vitro, where it refers to the added tension imposed on isolated muscle strips with contraction, but it is harder to define in vivo. In analogy with in vitro studies, afterload can be equated with ventricular wall stress, expressed as the product of cavity radius times transmural pressure divided by wall thickness, as described by the law of Laplace. However, many investigators find this definition unsatisfactory because it implies the heart generates its own afterload and because afterload would be viewed as changing during the cardiac cycle.¹⁰⁴ If afterload is viewed as the external forces opposing ejection, possibly the best definition is the aortic input impedance, the complex ratio of pressure to flow, expressed in terms of magnitude, and the phase angle between flow and pressure for any given frequency. However, it has been difficult to quantitatively analyze the impact of impedance on overall cardiac performance (see Chapter 5). Sunagawa et al¹⁰⁵ proposed a simplified theory of ventricular-vascular coupling within the framework of the end-systolic pressure-volume relation. Using the definition of end-systolic elastance produces the following:

Equating SV with V_E − V_ES (ignoring the difference between end-ejection and end-systole), it is a matter of algebra to show that

This means that an increase in SV implies a decrease in P_ES. The interpretation of this from the perspective of the heart is that the work that can be done is finite, and that a greater SV can only be achieved by decreasing the end-systolic pressure, if contractility (E_ES) is fixed. At the same time, it is known that from the perspective of the vasculature, if the SV increases, BP increases when vascular tone does not change.

The application of an electrical law describing constant voltages and flows to the circulation, in which pulsatile flow is generated by a pump, has resulted in estimates of afterload that are questionable.¹⁰⁴ Although SVR is a component of impedance, it cannot be equated with it. The correct downstream pressure is probably not the CVP.¹⁰⁶ There is instead a critical opening pressure that should be used in calculating SVR that is not measurable in routine clinical care.¹⁰⁶ Clinical use of such calculated resistances is made complex by the relation of CO to body size; the normal resistance of a small patient is much higher than that of a large one. The use of a resistance index (i.e., using CI instead of CO) partly overcomes this problem, but it is not widely used.

Calculated SVR continues to be widely used in guiding therapy or drawing conclusions about the state of the circulation. This should only be done with caution, if at all. SVR is not a complete indicator of afterload. Even if SVR were an accurate measure of impedance, the response to vasoactive agents depends on the coupling of ventricular-vascular function, not on impedance alone. Hemodynamic therapy should be guided based on the primary variables, BP and CO. If preload is appropriate, low BP and low CO are treated with an inotropic drug. If BP is acceptable (and preload appropriate) but CO is low, a vasodilator alone or in combination with an inotropic drug is used. If the patient is hypertensive (with low CO), vasodilators are indicated; if the patient is vasodilated (low BP and high CO), vasoconstrictors are used (Box 34-2).

BOX 34-2 Hemodynamic Therapy Guidelines

Blood Pressure	Cardiac Output	Treatment
Low	Low	Inotrope
Normal	Low	Vasodilator ± Inotrope
High	Low	Vasodilator
Low	High	Vasopressor

Postoperative hypertension

Hypertension has been a common complication of cardiac surgery, reported to occur in 30% to 80% of patients from studies in the 1970s when CABG was common in patients with normal ventricular function.¹⁰⁷^–¹⁰⁹ The current population of older, sicker patients appears to have fewer problems with hypertension than with low-output syndromes or vasodilation. Although hypertension most commonly occurs in patients with normal preoperative ventricular function, after aortic valve replacement, or a prior history, any patient may experience development of hypertension. Multiple reasons contribute to postoperative hypertension, including preoperative hypertension, preexisting atherosclerotic vascular disease, awakening from general anesthesia, increases in endogenous catecholamines, activation of the plasma renin-angiotensin system, neural reflexes (e.g., heart, coronary arteries, great vessels), and hypothermia.¹¹⁰ Arterial vasoconstriction with various degrees of intravascular hypovolemia is the hallmark of perioperative hypertension.

The hazards of untreated postoperative hypertension include depressed LV performance, increased Mo₂, cerebrovascular accidents, suture line disruption, MI, rhythm disturbances, and increased bleeding.^108,^111,¹¹² Historically, therapy for hypertension in cardiac surgery was sodium nitroprusside because of its rapid onset and short duration of action.¹¹³ With multiple vasodilators available in the current era, sodium nitroprusside is no longer the drug of choice for many reasons. Nitroprusside is a potent venodilator, increasing venous capacitance (decreasing preload), and can produce arterial vasodilation, often leading to precipitous decreases in BP and a hyperdynamic cardiac state. Sodium nitroprusside can cause coronary arteriolar dilation with the potential for a steal phenomenon, resulting in myocardial ischemia.¹¹⁴ In patients with renal failure, the elimination of sodium nitroprusside is reduced, potentially leading to toxic effects of its metabolites cyanide or thiocyanate. This can occur if large doses are given to patients with normal renal function.

There are many alternative drugs to sodium nitroprusside for treating hypertension after cardiac surgery, including nitroglycerin,¹¹⁵ adrenergic-blocking agents such as β-adrenergic blockers,¹¹⁶ and the mixed α- and β-adrenergic blocker labetalol.¹¹⁷ Direct-acting vasodilators, dihydropyridine calcium channel blockers (e.g., nicardipine,¹¹⁸ isradipine,¹¹⁹ clevidipine,¹²⁰^–¹²³) ACE inhibitors,¹¹³ and fenoldopam (a dopamine₁ [D₁] receptor agonist)^124,¹²⁵ also have been used. Novel therapeutic approaches are listed in Table 34-3.

TABLE 34-3 Novel Vasodilators

Drug	Mechanism of Action	Half-Life
Nicardipine	Calcium channel blocker	Intermediate
Clevidipine	Calcium channel blocker	Ultra-short
Fenoldopam	Dopamine₁-receptor agonist	Ultra-short
Nesiritide	Brain natriuretic agonist	Short
Levosimendan	K⁺-ATP channel modulator	Intermediate