Patients undergoing cardiac surgery often receive multiple pharmacologic agents preoperatively. Most antihypertensive and antianginal cardiac medications should be continued preoperatively. Whether to continue or when to withhold preoperative antihypertensive therapy necessitates patient risk/benefit. Acute withdrawal of antihypertensive medications such as β-blockers or clonidine may result in potential perioperative hemodynamic instability, and these medications should be continued unless there are specific hemodynamic or other concerns. ,

However, decisions to continue angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor subtype I antagonists (ARA) are less clear. Patients in whom ACE therapy was maintained until the morning of surgery had an increased probability of hypotension at anesthesia induction. ARA-treated patients had more postinduction hypotension refractory to conventional vasopressors than those on other antihypertensive therapy (β-blockers, calcium channel blockers, or ACE inhibitors).

Most recommendations suggest discontinuing ARA before surgery. , Chronic ACE inhibitor use or ARA therapy less than 10 hours prior to general anesthesia was reported to increase risk of developing moderate hypotension but was not associated with increased postoperative complications. Further, intraoperative hemodynamic instability related to ACE or ARA therapy is not consistent in all studies. Patients treated long-term with ACE inhibitors with normal left ventricular function did not have altered hemodynamic instability during cardiac surgery. Although continuing therapy has potential beneficial effects, , continued use of these agents on the day of surgery may necessitate episodic vasopressor support. Although the continuation of preoperative cardiovascular medications is recommended, holding ACE inhibitors or ARA therapy is likely warranted.

Beyond lipid-lowering properties, statins possess pleiotropic effects that have been reported to reduce postoperative morbidity and mortality. Among patients undergoing percutaneous coronary interventions, Pascual reported that statin pretreatment was associated with reduced early ischemic events, primarily in those with high levels of inflammatory markers. Pretreatment with statins in patients undergoing coronary artery bypass grafting (CABG) reduced P-selectin levels, an adhesion molecule that plays a role in the pathogenesis of arteriosclerosis, below those of patients given a placebo. Berkan reported less use of inotropic agents among patients treated with statins, speculating that myocardial injury caused by cardiopulmonary bypass (CPB)–induced inflammatory changes was reduced. Others have reported a protective effect of statin pretreatment in reducing myocardial damage after CABG. , Based on these beneficial pleiotropic effects, statin therapy should be continued routinely in the perioperative period.

In patients with cardiovascular disease, the use of oral anticoagulants is common and includes both antithrombotic agents and antiplatelet agents. Oral antithrombotic agents include warfarin and other vitamin K antagonists (VKAs), as well as the non-vitamin K direct oral anticoagulants (DOACs). Warfarin is the vitamin K oral anticoagulant most often used in North America, and the DOACs include the direct thrombin inhibitor, dabigatran, and the factor Xa (FXa) inhibitors, apixaban, edoxaban, and rivaroxaban. The oral antiplatelet agents include aspirin and the P2Y12 receptor antagonists clopidogrel, prasugrel, and ticagrelor. Clopidogrel and prasugrel are prodrugs, of which clopidogrel is the most commonly used P2Y12 receptor antagonist. It is a prodrug that undergoes two metabolic steps to be activated and has a potential resistance reported to be around 30%. Ticagrelor is a direct-acting agent with no reported resistance. These agents will be considered separately. There are multiple guidelines currently reported regarding management strategies for surgery, including those from the American College of Chest Physicians (ACCP) in 2022 to inform best practices for the perioperative management of patients who are receiving a VKA and may require heparin bridging for the perioperative management of patients who are receiving a DOAC and for patients who are receiving one or more antiplatelet drugs and will be included. Considerations in managing the different anticoagulants are as follows.

P2Y12 inhibitors: Clopidogrel, prasugrel, ticagrelor.

The P2Y12 inhibitors clopidogrel, prasugrel, and ticagrelor are inhibitors of adenosine diphosphate (ADP)-mediated platelet activation. Although patients undergoing elective CABG within 3 to 5 days of clopidogrel treatment are at increased risk for bleeding, transfusion, and reoperation, there may be a protective effect in patients undergoing emergency surgery ( Fig. 4.1 ). However, variability of bleeding, especially with clopidogrel, is likely because of the potential for clopidogrel resistance that can occur in at least 30% of patients as clopidogrel is a prodrug and undergoes metabolic transformation. This is not the case with ticagrelor, as it is not a prodrug and does not require metabolic transformation to be active. Patients receiving clopidogrel within ∼3 days of surgery compared to 5 days or more had higher transfusion requirements (95% vs. 52%). However, others have suggested this does not occur. Ebrahimi and colleagues examined the effect of clopidogrel in patients with non-ST-segment elevation acute coronary syndromes (NSTE-ACS) requiring CABG in the Acute Catheterization and Urgent Intervention Triage strategy (ACUITY) trial. The trial enrolled 13,819 patients with NSTE-ACS undergoing early invasive management; 11.1% underwent CABG before discharge. Clopidogrel-exposed patients had a longer median duration of hospitalization than nonexposed patients but experienced fewer ischemic events within 30 days and had a similar occurrence of non-CABG-related major bleeding and post-CABG major bleeding. Multivariable analysis demonstrated that clopidogrel use before CABG was a predictor of reduced 30-day composite ischemia. Current ACCP guidelines suggest that for patients receiving a P2Y12 inhibitor drug, clopidogrel should be stopped 5 days before, ticagrelor 3 to 5 days before, and prasugrel 7 days before the surgery (conditional recommendation, low certainty of evidence).

Prevalence of composite outcome (reoperation or major bleeding) by number of days clopidogrel was stopped before surgery. Denominator for points on curve is the total number of patients who were exposed to clopidogrel ≤7 days before surgery ( n = 329); 89 patients experienced the composite outcome. Red line = clopidogrel exposure.

(From Berger JS, Frye CB, Harshaw Q, Edwards FH, Steinhubl SR, Becker RC. Impact of clopidogrel in patients with acute coronary syndromes requiring coronary artery bypass surgery: a multicenter analysis. J Am Coll Cardiol. 2008;52:1693-1701.)

For patients with coronary stents placed within the prior 3 to 12 months who are receiving dual therapy with a P2Y12 inhibitor and aspirin, the P2Y12 inhibitor should be held before elective surgery, and routine bridging therapy with a glycoprotein (GP) IIb/IIIa inhibitor, cangrelor, or low-molecular-weight heparin is not advised.

Several automated devices are available for anticoagulation management following heparin administration; the activated clotting time (ACT) monitor is the most common. Unfractionated heparin (UFH) is commonly administered on a weight-based protocol (300-500 units per kilogram) with the goal of achieving an ACT of 480 or more seconds prior to initiation of CPB. , In general, if an ACT of 480 or greater has not been obtained with doses of heparin (up to 500-600 units · kg ⁻¹ ), one should suspect heparin resistance.

Heparin resistance is generally suspected when excessive doses of heparin are required to produce the desired anticoagulation. Reports suggest that heparin resistance in cardiac surgical patients requiring CPB occurs in 4% to 22% of patients. , Preoperative risk factors include preprocedural heparin administration, thrombocytosis, and hyperfibrinogenemia.

One of the issues regarding heparin resistance is the actual definition. An early important report in cardiac surgery suggested that a need for more than 500 units/kg was consistent with heparin resistance. Defining heparin resistance is further complicated by the lack of knowledge defining an ideal ACT required for CPB. The number of 480 seconds has been repeated in many textbooks and guidance reports and has become a standard threshold value, although earlier studies suggest 350 seconds may provide satisfactory anticoagulation for CPB. Of note is that a target level of 2 to 4 units per milliliter of heparin is another potential goal to consider for coagulation monitoring.

Treatment options for heparin resistance include antithrombin (AT) III therapy or, potentially, the use of fresh frozen plasma (FFP). In most clinical situations with time constraints of preparation for CPB, use of FFP is impractical. Several clinical trials that reported AT III is effective in restoring heparin responsiveness for most patients exhibiting heparin resistance prior to CPB defined heparin resistance as an ACT < 480 seconds after 400 units · kg ⁻¹ of heparin.

Heparin-induced thrombocytopenia (HIT) is an immune-mediated prothrombotic response caused by IgG antibodies that develop to platelet factor 4 (PF4), a polypeptide stored in platelet alpha granules. Following heparin exposure, the IgG antibodies bind to platelet epitopes, causing aggregation and further activation that release highly prothrombotic microparticles. Following cardiac surgery and CPB, up to 50% of patients form HIT antibodies. Due to this unique immunologic response, IgG antibodies that produce HIT last approximately 3 months. Thus, patients with a previous history of HIT may not be at increased risk for subsequent antibody formation as this immunologic response is not a true amnestic response but rather a transient event due to platelet activation.

Despite the high incidence of antibody formation, clinically apparent HIT occurs in 1% to 5% following UFH. The initial assay to determine whether a patient has HIT is the enzyme-linked immunoassay (ELISA). However, the assay has a high sensitivity but less specificity. Results return as optical density. The definitive test is a functional assay, usually a serotonin release assay. Warkentin noted that patients with higher optical densities were at greater risk for true HIT, as documented by definitive testing with serotonin release assays.

HIT usually presents with a 50% decrease in platelet count or a thrombotic event after heparin exposure; however, heparin is a ubiquitous drug in cardiovascular medicine, and prior recent exposure may not be known. ELISA should be used for clinical decision-making for intraoperative anticoagulation. If cardiac surgery cannot be delayed until HIT assays are negative, other techniques are currently used. Alternative anticoagulation includes bivalirudin, a direct thrombin inhibitor, which is the agent most extensively studied for on-and-off pump cardiac surgery. Alternatively, plasmapheresis to decrease the IgG titer is an another strategy using heparin for cardiac surgery and then postoperatively switching to a direct thrombin inhibitor such as bivalirudin or argatroban ( Table 4.7 ). ,

Protamine is the primary agent for heparin reversal. Protamine, a highly basic polypeptide, forms a nonspecific acid-base, polyionic-polycationic interaction with heparin that neutralizes its anticoagulant effect. Indeed, mixing the two drugs ex vivo forms a precipitate, and the protamine heparin interaction may be associated with hypersensitivity reactions, including anaphylaxis. Many cardiac surgical centers reverse on an empirical basis, administering typically 1 mg of protamine for every 100 units of heparin. However, this routinely administered far too much protamine, and excess protamine can contribute to coagulopathy post-bypass. The optimal management for protamine dosing for reversal is using a point-of-care titration to determine the exact level required. Because the half-life of heparin is approximately 1 hour, protamine dosing, administering 0.5 mg to 0.7 mg of protamine for every 100 units of the initial heparin dose for CPB, is another potential strategy. ,

Light in the near-infrared (700-1300 nm) range has three important physical properties that make it useful for diagnostic assessment:

• •

  It penetrates tissue.

• •

  It is non-ionizing.

• •

  It is absorbed differentially by relevant chromophores depending on their oxygen-binding state.

When near-infrared light is emitted across a tissue (e.g., brain) and detected at its exit, absorption of the light can be used to calculate chromophore concentration using variants of the Beer-Lambert equation. All optical spectrometers consist of the same basic components: a light source of known intensity and wavelength, a light detector to measure the intensity of the light exiting the tissue, and a computer to translate the changes in light intensity into clinically useful information such as the concentrations of HbO ₂ , hemoglobin, or oxidized cytochrome aa3.

When photons impinge on biological materials, their transmission depends on a combination of reflectance, scattering, and absorption effects. A light source (light-emitting diode or laser source) emits near-infrared light that passes through a “banana-shaped” reflectance path in the frontal cerebral cortex to two to three detectors placed 3 to 5 cm from the emitter. Absorption occurs at specific wavelengths, determined by the molecular properties of the material in the light path. Optical path length for reflected light is linearly related to spacing between transmission and receiving sites, so many NIRS measurement instruments place the transmitting diode and light detector several centimeters apart on the head. Although this spacing results in a measurable signal intensity, it affects the amount and depth of tissue monitored. On a practical level, the available instruments space their transmitting and receiving sites differently, thus measuring different quantities and depths of tissue, which makes comparisons between instruments difficult ( Table 4.10 ).

Shallow arcs of light travel across skin and skull but do not penetrate the cerebral tissue. Deep arcs of light cross skin, skull, dura, and cortex. Subtracting the absorbance measured in the narrow arc from that measured in the deep arc leaves absorbance that is due to intracerebral chromophores. This is one of the distinguishing characteristics of cerebral oximeters compared with pulse oximeters. Cerebral oximeters use spatial resolution techniques to differentiate cortical from extracranial blood, whereas pulse oximeters differentiate pulsatile (arterial) from nonpulsatile (venous/capillary) blood. Cerebral oximetry measures predominantly venous saturation (75:25 or 85:15, depending on age and model used). NIRS could be a surrogate for jugular venous oxygen saturation (Sjvo ₂ ) monitoring without being invasive. It does not depend on pulse, blood pressure, or body temperature. This makes the technique ideally suited for monitoring oxygenation during CPB, hypothermic circulatory arrest, shock, or cardiovascular collapse.

Outcomes after heart surgery and NIRS monitoring.

The Austin study was a retrospective cohort study; no randomized trial of NIRS has been conducted in children. Dent and colleagues showed a correlation between prolonged low regional oxygen saturation (rSO ₂ ) (<45% rSO ₂ for ≥180 min) and new magnetic resonance imaging (MRI) abnormalities in a group of neonates who underwent the Norwood procedure with antegrade cerebral perfusion (ACP). In a prospective study of neonates with either single- or two-ventricle physiology undergoing surgery with CPB and ACP, the authors were unable to show an association between new white matter injury on postoperative MRI and prolonged low perioperative rSo ₂ %. However, it is important to note that in this study, 50% of patients in the single-ventricle group had prolonged low rSo ₂ (<45% for 240 min), one third of whom had new white matter injury in the postoperative period. No patients in the two-ventricle group had prolonged rSo ₂ % below 45% for 120 minutes. Although stroke and chorea are obvious neurologic abnormalities, subtle neurocognitive changes are difficult to establish. Thus, despite the lack of level 1A evidence that NIRS improves neurocognitive outcomes, we recommend intraoperative cerebral oximetry as a tool to optimize anesthesia (ventilation, oxygenation); perfusion (alpha-stat, pH-stat, hemoglobin, flow, temperature); and surgical techniques (cannulation). Fig. 4.4 shows the protocol recently published by Zaleski and Kussman from Boston Children’s Hospital.

Intraoperative cerebral oximetry protocol.

(From Zaleski KL, Kussman BD. Near-Infrared Spectroscopy in Pediatric Congenital Heart Disease. J Cardiothorac Vasc Anesth . 2020;34(2):489-500)

Primary RV dysfunction may occur after intracardiac surgery in neonates, infants, and children. Diagnosis of RV dysfunction is suggested by high right-sided filling pressures, liver distension, hypotension, tachycardia, reduced cardiac output, and systemic venous desaturation (low mixed-venous saturation). Treatment is directed toward improving oxygen delivery by increasing preload, augmenting contractility directly or indirectly, enhancing coronary perfusion, and reducing afterload.

The right ventricle is generally less responsive to inotropic support than the LV and, therefore, may require higher doses of inotropic agents. Epinephrine enhances RV contractility. By improving systemic arterial pressure, epinephrine can augment RV coronary blood flow. Maintaining a normal to slightly elevated systolic arterial pressure will maximize coronary perfusion and augment RV contractility. Milrinone, a phosphodiesterase-3 inhibitor, is a useful inotrope with pulmonary vasodilatory properties. In a randomized controlled trial of infants and children after heart surgery, a bolus of 75 µg/kg followed by a 0.75 µg/kg/min infusion reduced the occurrence of low cardiac output syndrome by 55%. The use of milrinone varies depending on institutional preferences; the loading dose usually ranges between 25 to 50 µg/kg during rewarming, followed by an infusion of 0.375 to 0.5 µg/kg/min. Communication with the surgeon is important prior to administering the bolus, even on CPB, because some surgeons are concerned with the hypotension that could result.

RV afterload can be decreased by mechanical ventilation with or without inhaled nitric oxide (iNO). Mechanical ventilation should be adjusted to optimize preload and decrease afterload. The right ventricle is extremely sensitive to alterations in intrathoracic pressure; therefore, ventilation that enables the lowest possible mean airway pressure should be the goal. Increased mean intrathoracic pressure increases RV afterload by direct compression of alveolar and extra-alveolar pulmonary vessels. iNO is an endothelium-derived smooth muscle relaxant used in neonates with persistent pulmonary hypertension and in pulmonary hypertension related to CHD. Nitric oxide (NO) decreases Rp and reduces intrapulmonary shunting, which may improve oxygenation. However, results of its use in the postoperative period are conflicting. In a randomized controlled trial of more than 100 infants at high risk for pulmonary hypertension, 20 ppm not only failed to show any benefit, it did not prevent pulmonary hypertensive crisis in children after congenital heart surgery. However, in a similar population, others have shown it to be effective even at lower doses. If inhaled nitric oxide (iNO) is used, NO ₂ levels should be monitored, and iNO should not be abruptly withdrawn because rebound pulmonary hypertension can occur.

If these measures are unsuccessful, ECMO should be implemented. EMCO unloads the right ventricle and favorably shifts the oxygen supply-demand ratio, often allowing the injured myocardium to recover.

After separation from CPB, the contractile state of the systemic ventricle may be depressed. Contributing factors include the preoperative condition of the myocardium (myocardial hypertrophy, elevated end-diastolic pressure, systolic dysfunction), response of the myocardium to the new loading conditions imposed by the operative repair, effects of hypothermia on myocardial compliance, suboptimal myocardial protection, and residual anatomic problems.

Systemic ventricular dysfunction is managed by optimizing preload, afterload, and heart rate. Tachycardia (>180-190 beats/min) may impair ventricular function in newborns and infants and should be treated with β-adrenergic blocking agents and, if necessary, vasopressors. When the heart rate is less than 120 to 130 beats/min, atrial or AV sequential pacing is appropriate.

In the preinduction period, ductal patency must be maintained with prostaglandins to ensure systemic cardiac output. Management depends on optimizing systemic oxygen delivery (by increasing cardiac output) and restricting Q ˙ P.

Before discontinuing CPB, ionized calcium levels and hematocrit must be optimized to ensure adequate oxygen-carrying capacity. Myocardial function is supported by the judicious use of inotropes. Tidal volume is increased to account for a reduction in lung compliance, and minute ventilation is adjusted to maintain normocarbia. After separation from CPB, Fi o ₂ is adjusted to maintain Sa o ₂ between 75% and 85% and an arterial Pao ₂ of 40 to 50 mmHg. Modified ultrafiltration has been shown to improve myocardial function, decrease lung water, and remove inflammatory mediators in patients with hypoplastic left heart physiology, as well as other complex malformations (see Chapter 2 ). Excessive Q ˙ P is less common in the immediate postbypass period. After modified ultrafiltration and before chest closure, the anesthesiologist should estimate Q ˙ P/ Q ˙ s and attempt to adjust Fi o ₂ and minute ventilation accordingly.

Chest closure can markedly reduce lung compliance and worsen hemodynamics. Leaving the chest open may improve Q ˙ P and heart filling by reducing mean airway pressure.

Lung isolation is helpful during thoracoscopic procedures in older children and adolescents. Lung isolation in infants and toddlers can be achieved with endobronchial intubation or bronchial blocker. The latter can be placed under fiberoptic or fluoroscopic guidance. Bronchial blocker or endobronchial intubation can be used for lung isolation in children. The former allows expansion of the surgical side, if necessary, by deflating the blocker to improve oxygenation. Both techniques are associated with complications and should be undertaken judiciously. Lung isolation in cyanotic pediatric patients requires close communication between the surgeon and anesthesiologist, as further reductions in oxygen saturation may be associated with ischemic changes on the ECG.

Cardiac reserve is the capacity to increase (or at least maintain) cardiac output in response to a variety of stressful sudden developments. It is determined by the integrative ability of cross-system mechanisms, including increased neurohumoral total (i.e., central and peripheral) body V o ₂ , increased ventricular afterload, and decreased ventricular preload. , That capacity is provided by all cardiac and extracardiac mechanisms for maintaining and increasing the force of ventricular contraction and cardiac output, primarily myocardial contractility and coronary blood flow. ¹ After cardiac surgery, the adequacy of cardiac performance alone is insufficient to ensure normal recovery and survival. Cardiac reserve also must be adequate.

Cardiac reserve inadequacy is most common in the first 6 to 12 hours after cardiac surgery, particularly during periods of increased V o ₂ (e.g., from agitation, tachypnea, or hyperthermia); suddenly increased ventricular afterload (e.g., from paroxysmal pulmonary arterial hypertension in a neonate); or acute reduction in ventricular preload (e.g., from sudden blood loss). Such cardiac reserve inadequacies can be categorized as “pre-pump” issues (i.e., lack of preload), pump issues (i.e., myocardial ischemia), or “post-pump” issues (i.e., vasoplegia), which can produce hypotensive emergencies or cardiac arrest early after cardiac surgery.

Cardiac reserve is also highly dependent on the patient’s preoperative condition. When disease causes reserves to be nearly maximally used to maintain adequate cardiac performance preoperatively, that which remains may be insufficient to meet the stresses of the intraoperative and postoperative period successfully. Reserves probably cannot be increased before the operation unless they have been acutely impaired by a reduced myocardial energy charge. Energy charge may be increased by the cardioplegic technique used in the OR. Limited cardiac reserves are also specifically compensated for by many features of early postoperative care.

CI (cardiac output expressed as L/min/m ² ) is one measure of adequacy of the cardiovascular subsystem, as evidenced by the oft-noted relationship between CI and survival (described by Dietzman and colleagues in 1969 ). In adults, a CI of at least 2.2 L/min/m ² or higher is generally adequate oxygen delivery in the early postoperative period ( Fig. 4.5 ). This is at the lower end of the range of normal (2.2–4.4 L/min/m ² ). In general, infants and small children appear to require a somewhat higher CI for normal recovery ( Fig. 4.6 ). In young patients, CI tends to be lower for about 4 hours after surgery than it was soon after discontinuing CPB; it begins to rise after 9 to 12 hours. ,

Relationship of cardiac index in the early hours after mitral valve replacement to the probability of cardiac death (UAB, 1975-1979). Solid line is the point estimate, and dashed lines are the 70% confidence limits.

(From Conti VR, Wideman F, Blackstone EH, Kirklin JW. Unpublished study; 1979.)

Relationship of early postoperative cardiac index (average of all early postoperative values) to probability of cardiac death in infants and small children. This graph suggests that convalescence cannot be considered normal in infants and small children unless cardiac index is about 2.0 to 2.2 L · min ⁻¹ · m ⁻² , somewhat higher than the value for adults.

(From Parr GV, Blackstone EH, Kirklin JW, Pacifico AD, Lauridsen P. Cardiac performance early after interatrial transposition of venous return in infants and small children. Circulation. 1974;50(2 suppl):II2-II8.)

Cardiac indices lower than these values are usually inadequate for maintaining normal recovery , ; this can be formalized in the inverse relation between CI early after surgery and the probability of hospital death. This relation can be refined by considering not only cardiac output but also mixed venous oxygen levels, with lower levels indicating a worsening prognosis at any given value of cardiac output. ²

Arterial blood pressure is an insensitive method of estimating adequacy of cardiac output early after surgery, primarily because systemic vascular resistance is usually altered from baseline. This may be related to increased levels of circulating catecholamines, plasma renin, or angiotensin II, or to other mechanisms. Systemic vascular resistance may result in normal, high, or low arterial blood pressure even when cardiac output is low.

Some patients tend to have low systemic vascular resistance and arterial blood pressure early after surgery, even when cardiac performance is good. This may occur more frequently in children with cyanotic heart disease, adults with diabetes, and patients with sepsis or drug interactions (especially preoperative use of ACE inhibitors). , Arterial hypotension is an indication for thoughtful evaluation.

Children are not convalescing normally when mean arterial blood pressure is >10% below normal for the patient’s age ( Table 4.14 ). For adults, particularly those older than 65 years, arterial blood pressure may mandate maintenance at or above commonly accepted normal values to ensure adequate perfusion of various organs, such as the brain, viscera, and kidneys.

Simple observation of pedal pulses is common in the postoperative period as a useful, but not infallible, method of estimating cardiac output adequacy in children, young adults, and those without peripheral vascular disease. Normal (grade 4) pedal pulses early after surgery are highly but not perfectly correlated with adequate cardiac output and a high probability of survival. , In adults aged ≥65 years, using the amplitude of pedal pulses to estimate perfusion adequacy is often confounded by the presence of peripheral arterial occlusive disease.

V o ₂ is the quantity (in mL) of oxygen consumed by the tissue per minute. Whole-body V o ₂ is infrequently calculated, but knowledge of it is useful; in some circumstances, it provides a better basis for prognostic and therapeutic inferences than cardiac output or mixed venous oxygen levels. Whole-body V o ₂ can be calculated by using a rearranged Fick equation. ³

The normal value for V o ₂ at 37°C is 155 mL/min/m ² (range 110–160 mL/min/m ² ); however, this value may increase severalfold during periods of stress. The value for whole-body V o ₂ in the patient recovering from cardiac surgery must be interpreted in light of body temperature: Residual hypothermia contributes to low V o _2, usually present within the first few hours after open heart surgery. This reduced V o ₂ is due in part to reduced capillary density (reduced area of capillary flow) and increased heterogeneity of capillary flow through the muscle mass and other body tissues in the early hours after CPB. Normally recovering patients operated on with mild hypothermic CPB generally require 2 to 8 hours for this effect to disappear and for peripheral perfusion to return to normal.

When V o ₂ is appreciably lower than the normal level for the existing body temperature, a hazardous condition exists; indeed, one useful definition of shock is “a condition characterized by an acute reduction in V o ₂ .” Abnormally low V o ₂ may result from reduction in or extreme heterogeneity of capillary flow (of which “no reflow” is an extreme example) in one or more organs of the body (sometimes termed a reduction in capillary density ), lengthening of the diffusion path between capillaries and cells, or intracellular metabolic derangement. One or all of these may exist in patients early after cardiac surgery. When important V o ₂ reduction relative to temperature persists for more than a few hours, the probability of death increases.

A further consideration is oxygen extraction (O ₂ ER), which is the fraction of D o ₂ per min (expressed as O ₂ ER = V o ₂ /D o ₂ ). Normal O ₂ ER is approximately 25% (or a ratio 0.25) but can increase to meet the metabolic demands of the body’s tissues (i.e., maintain V o ₂ ).

Mixed venous oxygen level, generally expressed as oxygen tension (Pv o ₂ ) or saturation (Sv o ₂ ) within the venous system, can be a useful index of circulatory adequacy because it somewhat reflects mean tissue oxygen levels. When Pv o ₂ is <30 mmHg, oxygen delivery is impaired, probably secondary to inadequate cardiac output; when it is <23 mmHg, the inadequacy is likely to be severe ( Fig. 4.7 ). However, normal or near-normal venous oxygen levels are not reassuring as to the adequacy of cardiac output unless V o ₂ is approximately normal for the existing body temperature; that is, normal mixed venous oxygen level does necessarily denote normal D o ₂ , whereas an abnormal level usually portends poor oxygen delivery.

Relationship of mixed venous Po ₂ to probability of acute cardiac death in infants and young children. Convalescence cannot be considered normal if the value is less than about 28 mmHg.

(From Parr GV, Blackstone EH, Kirklin JW. Cardiac performance and mortality early after intracardiac surgery in infants and young children. Circulation. 1975;51(5):867-874.)

After CPB, changes in venous oxygen level may be useful in detecting low cardiac output, one of many causes of decreased oxygen delivery. Identifying decreases may be of particular value in patients coming to surgery with a high severity-of-illness index; however, the method of measurement in real time is imperfect. In a nonsurgical but critically ill ICU population, Jain and colleagues found a weak relationship between CI and Sv o ₂ , as measured from a Swan-Ganz catheter. Other studies of indwelling fiberoptic reflectance oximetry have found no or only a very weak relationship between Sv o ₂ or Pv o ₂ and measured CI. , In contrast, more recent studies of the utility of monitoring Sv o ₂ value as a response to inotropic therapy have observed that persistently low Sv o ₂ values (below 60%) after cardiac surgery were associated with higher 30-day and 1-year mortality.

As postoperative critical care has evolved, intraoperative echocardiography and the use of the Swan-Ganz catheter have become standard routines. , In lower-risk patients, some centers have begun to use a central-line catheter that can obtain a venous oxygen sample from the SVC to be used interchangeably with Sv o ₂ or as a trend correlate. However, caution is warranted because of concerns about accuracy across the spectrum of contemporary cardiac surgery patients. ,

The acid-base status of blood is a frequently used, but somewhat nonspecific and insensitive, indicator of cardiac output adequacy. Metabolic acidosis during and after cardiac surgery is most commonly a result of lactic acidemia. Lactate production is a byproduct of anaerobic metabolism, which most often occurs under conditions in which oxygen delivery is suboptimal (typically secondary to low cardiac output), as is oxygen consumption. Our understanding of lactate metabolism has grown considerably over the last 30 years. Lactate measurement remains an important tool for clinicians in the ICU, but lactate elevation should be thought of as a sign of metabolic stress and not assumed to represent an oxygen debt. The presence of hyperlactatemia portends worse patient outcomes and should trigger an evaluation for the underlying cause.

Lactate has been shown to be an important metabolic fuel for both the heart and brain during periods of stress. Under normal conditions, the heart derives approximately 60% to 100% of its energy from free fatty acids and the remainder from glucose and lactate (0%–20% from each). However, myocardial uptake and utilization of lactate has been shown to increase during times of metabolic stress, such as during exercise, shock, and β-adrenergic stimulation, and can account for up to 60% of the cardiac oxidative substrate. The human brain also becomes a lactate consumer during periods of increased metabolic demand. Lactate has been shown to account for approximately 7% of cerebral energy requirements under basal conditions and up to 25% during exercise. Thus, lactate appears to be an important mobile fuel source for aerobic metabolism during times of stress, as it can be rapidly created and exchanged between tissues.

Lactate is formed from pyruvate in the cytosol of the cell following this equation:

During glycolysis, glucose is broken down through several chemical reactions to form 2 molecules of pyruvate as well as 2 molecules of ATP. Cytoplasmic pyruvate can then undergo either (1) oxidative phosphorylation via the citric acid cycle in the mitochondria or (2) enzymatic conversion to lactate, a process that does not require oxygen. The conversion of lactate to pyruvate is held in equilibrium by the enzyme lactate dehydrogenase, which favors the production of lactate with a lactate:pyruvate ratio of approximately 10:1 under normal conditions. Therefore, any condition that increases pyruvate production will also increase lactate formation.

The average lactate turnover at a physiologic steady state is approximately 20 mmol/kg per day, with 70% to 75% of this being cleared by the liver. , The two major mechanisms of lactate clearance are (1) oxidation via pyruvate and the citric acid cycle and (2) gluconeogenesis in the liver and kidneys via the Cori cycle. , At rest, approximately half of available lactate is metabolized through oxidation and 75% to 80% during exercise.

Hyperlactatemia can occur in any clinical setting where lactate production exceeds lactate clearance. It is a common occurrence in patients undergoing cardiac surgery, with a reported incidence as high as 80% during the perioperative period, and it has been associated with worse patient outcomes. , A lactate level of 3 to 5 mmol/L is typically used to define hyperlactatemia (0–2 mmol/L is considered normal).

Lactic acidosis can be classified into two categories (type A and B) based on the presence or absence of clinical signs of tissue hypoxia. This distinction can provide a useful framework for identifying causes and appropriate management in patients with hyperlactatemia.

• •

  Type A lactic acidosis is defined by lactate accumulation in the setting of either global or regional tissue hypoxia. In these scenarios, lactate is overproduced and underutilized as a result of impaired mitochondrial oxidation.

• •

  Type B lactic acidosis is defined by lactate elevation in the absence of cellular hypoxia. It is further subdivided into type B1, which occurs in specific disease processes; type B2, which is caused by a drug or toxin; and type B3, which is caused by inborn errors of metabolism.

An important and likely underappreciated mechanism of type B lactic acidosis is adrenergic-driven accelerated aerobic glycolysis. When carbohydrate metabolism exceeds the oxidative capacity of the mitochondria, intracellular pyruvate concentration increases, which in turn increases lactate production by mass effect. Under stressful conditions, epinephrine-dependent stimulation of the β ₂ -adrenoceptor augments the glycolytic flux, both directly and through enhancement of sarcolemmal Na ⁺ K ⁺ -ATPase. Disorders associated with elevated endogenous epinephrine levels, such as severe asthma, sepsis, severe trauma, cardiogenic or hemorrhagic shock, and pheochromocytoma, can cause hyperlactatemia through this mechanism. Aerobic glycolysis and tissue hypoxia are not mutually exclusive and, in many situations, may both contribute to hyperlactatemia. ^,

The diagnostic utility of lactate is diverse. It functions as a marker of resuscitation, it can identify patients with occult hypoperfusion, and it can provide important prognostic information. In the clinical setting, hyperlactatemia is commonly used as a marker of hypoperfusion and tissue ischemia. , Postoperatively, its presence is associated with a complicated recovery and should trigger the clinician to evaluate the patient for a new shock state or evidence of tissue hypoxia. Resuscitative efforts to normalize serum lactate are often designed to improve oxygen delivery to reduce tissue hypoxia. These interventions traditionally include initiation of volume resuscitation, addition of inotropes and vasopressors, and transfusion of packed RBCs.

A patient with elevated lactate should be evaluated for evidence of inadequate perfusion. Capillary refill, urine output, and mental status are easily assessed at the bedside, and abnormal findings should raise concern about inadequate perfusion. Central venous oxygen saturation <70% or a venoarterial C o ₂ difference >6 mmHg may signify inadequate oxygen delivery.

Patients with evidence of global hypoperfusion should be thoroughly evaluated for the underlying cause of shock. Information from a pulmonary artery catheter, such as CI, filling pressures, and mixed venous oxygen saturation, along with an echocardiogram to evaluate ventricular and valvular function, can be invaluable for diagnosing cardiogenic shock and tamponade physiology. Chest tube outputs should be evaluated, and if evidence of significant postoperative bleeding is found, the patient should be evaluated for coagulopathy. Any coagulopathy should be treated, and the patient should be resuscitated with blood products. If bleeding persists, the patient may require re-exploration.

In the absence of global hypoperfusion, the patient should be evaluated for regional ischemia, such as mesenteric ischemia, limb ischemia, or compartment syndrome, and surgical remediation should be initiated when warranted. Serial lactate testing can guide ongoing resuscitation efforts, as lactate clearance has been associated with improved outcomes; failure to clear lactate is a strong predictor of death. ^,

If there is no clinical evidence of hypoperfusion, interventions aiming at increasing D o ₂ are not likely to be of any benefit and may potentially cause harm. In this situation, non-hypoxic causes of lactic acidosis should be considered. The patient’s medical history and medications should be reviewed for potential causes of increased lactate production or impaired clearance.

• •

  Lactate measurement is an important tool in the cardiac surgery ICU.

• •

  Lactate elevation should be thought of as a sign of metabolic stress and not assumed solely to represent an oxygen debt requiring intervention to improve oxygen delivery.

• •

  The presence of hyperlactatemia portends worse patient outcomes and should trigger an evaluation for both hypoxic and nonhypoxic causes, with rapid intervention if required.

In the intact ventricle, afterload is defined as systolic wall stress. This is the analog of the load that resists shortening in the isolated papillary muscle. Other things being equal, increased afterload results in decreased stroke volume. In the intact ventricle, afterload is related to (1) ventricular transmural pressure during systole, (2) ventricular wall curvature, as determined by ventricular volume (La Place effect), (3) ventricular wall thickness, and (4) the shape of the ventricle.

Ventricular wall determinants of afterload are not likely to change during and early after surgery. Instead, acute changes in LV and RV afterload are usually produced by changes in intraventricular pressures during systole. These changes are equated with changes in proximal aortic and pulmonary arterial systolic pressures. During and early after surgery, proximal pulmonary arterial pressures can be monitored directly (by using a pulmonary artery catheter), but proximal aortic pressures are not. They must be inferred from measured radial (or brachial or femoral) artery pressures. As many factors contribute to the magnitude of systolic amplification, systolic blood pressure is usually higher at the radial artery than in the ascending aorta, except in the presence of peripheral vasoconstriction secondary to low cardiac output or high-dose α-adrenergic agents. In most instances, systolic pressure variability between the aorta and peripheral arteries is not clinically important, but an awareness of it is advantageous in some situations. Mean pressures are similar in the two areas.

A tendency toward arterial hypertension is present in as many as 90% of adult patients soon after surgery, related to increased systemic arteriolar resistance. , This complication increases (1) ventricular afterload and thereby decreases stroke volume, (2) aortic wall tension and thereby increases the likelihood of tearing aortic purse-string sutures and/or aortotomy suture lines, and (3) LV metabolic demands, which can exacerbate any latent myocardial ischemia. Nonetheless, there is a surprising lack of robust data to guide blood pressure management after cardiac surgery. Although mean arterial blood pressure (but typically not systolic pressure) is monitored due to the interrelations mentioned earlier between peripheral and central arterial pressures, contemporary early postoperative blood pressure management seeks to target a MAP of 65 to 90 mmHg and a systolic blood pressure <140 mmHg. , However, the patient’s preoperative blood pressure must be considered, and to avoid cerebral complications, markedly hypertensive patients must not be rendered hypotensive through overtreatment. For example, a higher target MAP may be indicated in patients with baseline hypertension or abnormal cerebral autoregulation. , Conversely, a lower MAP may be desirable in patients with poor ventricular function that is sensitive to afterload or to protect a fragile aorta.

Negative intrathoracic pressure can increase LV load-resisting shortening by increasing LV transmural pressure. Positive-pressure ventilation negates this effect, but labored spontaneous ventilation may augment afterload, which in turn may decrease cardiac output. , Sodium nitroprusside has traditionally been used for pharmacologic management of postoperative hypertension in a cardiac surgery ICU (see Appendix), but nitroglycerin may be preferred when myocardial ischemia is present, as it can decrease coronary resistance. Other agents include low-dose labetalol (5–10 mg) if an adequate heart rate and functional epicardial pacing wires are present or hydralazine (5–20 mg) when rapid normalization of blood pressure is required. Here again, it is important that the perioperative specialist judiciously avoids overshoot and iatrogenic hemodynamic swings. A reasonable approach is to set an initial MAP target of 10% above the normal value.

A newer agent being used in the postoperative cardiac surgery patient is clevidipine. Clevidipine quickly reduces blood pressure by selectively acting on the L-type Ca2+ channels on arteriolar smooth muscle. Clevidipine has a potential advantage over other agents (including nicardipine) due to its ultrashort action and rapid hydrolysis by blood and extravascular tissue esterases, and it does not depend on hepatorenal metabolism and excretion. , In addition to its rapid action and metabolism, clevidipine may be more cost-effective than nitroprusside in patients undergoing cardiac surgery and in those with an aortic dissection; however, this requires additional study. ,

When a change in stroke volume cannot be explained by a change in end-diastolic fiber length (a preload or pre-pump issue) or load-resisting shortening (an afterload or post-pump issue), it is considered to result from a change in the contractile state (a pump issue). Contractility in a given ventricle can be acutely depressed or increased.

Bedside assessment of myocardial contractility and the resultant quantification of ventricular pump function are desirable goals postoperatively. The simplest way to represent the heart’s capacity as a pump is to determine any of several modifications of the Frank-Starling mechanism. For instance, the measured change in cardiac output (or stroke volume) with a fluid or blood-product challenge serves as a surrogate for contractile function. In truth, this is not reflective of intrinsic contractile properties of the myocardium because the pressure-volume relationship is affected not only by preload but also by load-resisting shortening, myocardial compliance, and intact vagal and sympathetic reflex activity. Changes in instantaneous ventricular or aortic pressure over time may reflect myocardial contractility, but this value is exquisitely sensitive to afterload and preload.

The relationship between ventricular pressure and volume (the pressure-volume loop) is currently the nearest approximation to an in vivo assessment of contractility. Additionally, the area within the loop represents stroke work. The end-systolic pressure and the pressure at end-diastole of several different loops represent contractility and stiffness, respectively. The loops are composed of four segments: isovolumic contraction, ejection, isovolumic relaxation, and filling ( Fig. 4.10 ). When ventricular volume or resistance is altered, a group of points at end-systole fall along a line, the slope of which Suga and colleagues called Emax . Emax is an index of contractility at zero volume ( Fig. 4.11 ); changes in the slope in a steeper direction reflect increased inotropy and a shift rightward represents negative inotropy. With the use of catheters for LV pressure measurement and TEE for instantaneous border detection, pressure-volume loops, and Emax can be interpreted online in the ICU.

Diagrammatic representation of a pressure-dimension relationship of the left ventricle, on which events of the cardiac cycle have been indicated.

(From Foex P, Leone BJ. Pressure-volume loops: a dynamic approach to the assessment of ventricular function. J Cardiothorac Vasc Anesth. 1994;8:84.)

When resistance to ejection is altered, pressure-dimension loops at end-systole extend to a straight line termed the end-systolic pressure-dimension line. Slope of this line is an index of contractility. Increase in inotropy causes an increase in the slope of the line. It can also be seen that an increase in inotropy causes widening of the loop as ejection shortening is increased. Extrapolation of end-systolic pressure-dimension line to zero pressure defines V ₀ (or D ₀ ), the dimension the ventricle would attain if intracavitary pressure became zero.

(From Foex P, Leone BJ. Pressure-volume loops: a dynamic approach to the assessment of ventricular function. J Cardiothorac Vasc Anesth. 1994;8:84.)

During and early after cardiac surgery, one or both ventricles are typically what limits cardiac performance. It is usually advantageous to keep this concept clearly in mind when considering and treating patients. When the AV valves are normal, the most important indicator is the relationship between left and right atrial pressures because this represents the closest available approximation of ventricular end-diastolic pressure and, by implication, sarcomere length. When the cardiac valves are normal, the ventricle with the highest corresponding atrial pressure is more commonly the one limiting cardiac performance. Echocardiography can often provide useful, supportive information. In an analysis of more than 1.2 million patients undergoing cardiac surgery, the use of echocardiography (particularly in higher-risk patients) was associated with lower mortality and identification of important valvular pathologies needing intervention.

The optimal target for the postoperative patient is sinus rhythm at a normal heart rate (see Table 4.15 ). The normal compensatory response to increased O ₂ demand is increased heart rate. However, in older adults and patients with diseased myocardium, this response is often absent. In this situation, it may be prudent to manipulate heart rate in otherwise normally recovering patients with slow sinus (or junctional) rhythm to approximate AV electrical synchrony (even if not the same as normal sinus rhythm), with the goal of improving cardiac output. For this, atrial pacing via two temporary atrial leads placed during surgery is used (see D o ₂ equation earlier). Atrial pacing is also helpful in suppressing premature atrial and ventricular beats and may limit the onset of an established arrhythmia.

Disturbances of cardiac rhythm also may contribute to low cardiac output. Junctional (AV nodal) rhythm reduces cardiac output by 10% to 15%. Atrial contraction may contribute more in patients who have abnormal ventricular compliance and who are, therefore, more reliant on atrial contraction for adequate preload. Nonetheless, a junctional rhythm is less efficient than sinus rhythm because the atrial contribution to ventricular filling is absent with junctional rhythm. A junctional rhythm is usually transient, and its effects are easily overcome by atrial pacing (unless the rate is rapid); its presence does not connote an immediate risk.

Bradyarrhythmias caused by edema or damage to the AV node or His bundle, hypoxemia, or drugs can lower cardiac output. Tachyarrhythmias in the form of AF or flutter, or paroxysmal atrial tachycardia may also result in hypotension. Risk for tachyarrhythmia increases during infusion of catecholamines. A complete discussion of postoperative rhythm disturbances and their treatment is found later under “Cardiac Arrhythmias.” ,

The surgeon’s responsibility for performing an adequate operation demands that he or she, with the perioperative team, continue to search for evidence of this postoperatively, particularly when the patient has signs and symptoms of a low cardiac output state. Using the methods described in this and other chapters, the postoperative care team should search for residual intracardiac or extracardiac shunting, pathway obstructions, valvular regurgitation, graft or conduit dysfunction, or cardiac compression. If the operation is found to be inadequate in any of these regards, prompt reoperation is usually indicated.

Myocardial dysfunction was once thought to explain low cardiac output after cardiac surgery when atrial pressures were elevated above the usual postoperative values—in the absence of any other explanation. The availability of two-dimensional echocardiography in the OR and ICU, particularly TEE, enables direct visualization of ventricular wall motion and assessment of end-diastolic and end-systolic volumes. , These kinds of studies support the inference that low cardiac output is caused by myocardial injury or stunning or by impaired cardiac reserve in the presence of increased stress. This inference can be supported by the finding of increased creatine kinase-myocardial band isoenzyme or troponin in the serum (see also the Fourth Universal Definition of Myocardial Infarction ).

Increased RV afterload may appear quickly as a result of a sudden rise in pulmonary artery pressure and vascular resistance. Consequently, during an episode of paroxysmal pulmonary hypertension (e.g., intratracheal suctioning, coughing, or ventilatory dyssynchrony), cardiac output may fall rapidly, and “sudden” death may occur, particularly in neonates and infants. These outcomes are probably not purely the result of increased RV afterload because they reflect impaired RV reserve as well.

Increased left ventricular afterload may result from a sudden elevation of systemic arterial pressure (hypertension) during suctioning, restlessness, or hypoxia. Hypertension can result in increased bleeding and strain vascular anastomoses, and cause a sudden increase in left ventricular afterload that, combined with impaired left ventricular reserves, can lower cardiac output soon after surgery. Sustained increase in systemic vascular resistance and left ventricular afterload is present early after cardiac operations in at least half of adults who undergo surgery for acquired heart disease. ,

Often, disturbances in afterload (afterload mismatch) and preload (preload reserve) are neither independent nor isolated events. Increased RV afterload leads to decreased left ventricular preload and interventricular dependence. Similarly, increased left ventricular afterload leads to decreased RV preload. A common situation in which a corrective operation leads to increased left ventricular afterload is restoration of mitral valve competence or closure of a VSD. In a physics analogy, mitral regurgitation or left-to-right flow through a VSD represents a pair of resistors in a parallel circuit in which R _T = 1/r ₁ + 1/r ₂ , where r ₁ and r ₂ represent resistances in the two outflow streams and R _T is total resistance. Closure of 1 outflow (r) increases downstream resistance to ventricular shortening; by inference, wall tension and myocardial oxygen consumption increase.

The most important intraoperative risk factor for low cardiac output soon after surgery, other than an incomplete operation, is a discrepancy between the duration of any global myocardial ischemia and the efficacy of the measures used for myocardial management (see complete discussion in Chapter 3 ). This is often reflected in the finding of long global myocardial ischemic time as a risk factor for low cardiac output and for death postoperatively. Coronary air embolization during CPB is said to affect cardiac performance after the operation adversely, but most air entering the coronary arteries while the heart is not supporting the circulation passes quickly into the coronary sinus and may have little deleterious effect.

The extent of the whole-body inflammatory response to CPB affects the heart as well and relates to the probability of low cardiac output after surgery. This is one aspect of the correlation between CPB duration and postoperative death; another is that long CPB duration is sometimes the result, rather than the cause, of poor cardiac performance.

Low cardiac output due to surgical factors can result from (1) an incomplete or inadequate operation; (2) acute myocardial injury, with or without necrosis, from impaired coronary blood flow resulting from, for example, failure of CABG; (3) incomplete relief of ventricular inflow or outflow obstruction; (4) important residual or created AV valve or semilunar valve regurgitation that may increase the stroke volume requirements of the ventricle; and (5) a residual VSD or large left-to-right shunt that may similarly increase left ventricular stroke volume requirements.

Fine polyvinyl catheters are sometimes placed temporarily in the left atrium, right atrium, occasionally the pulmonary trunk via the right ventricle, and uncommonly, the LV to investigate whether an inadequate or incomplete operation is causing the low cardiac output; temporary epicardial atrial and ventricular wires are placed for control of heart rate. More commonly, TEE with color-flow Doppler imaging has been used to detect an incomplete or inadequate operation.

The possibility that residual left-to-right shunting is contributing to low cardiac output must always be considered after repair of congenital heart lesions and acquired or iatrogenic septal defects. Left-to-right shunting can be estimated in terms of the pulmonary (Qp) to systemic (Qs) flow ratio by simultaneously removing samples from the radial artery, right atrium, and pulmonary artery. Given the materials now used as patches for VSD repair, an appreciable left-to-right shunt (Qp/Qs >1.5) early postoperatively must be assumed to represent an incomplete repair or an overlooked defect. The simplified shunt equation is:

where Sa o ₂ is the oxygen saturation percentage of arterial blood, SRA o ₂ is the volume of blood withdrawn from the right atrium, and SPA o ₂ is the volume of blood withdrawn from the pulmonary artery.

TEE with color-flow Doppler imaging can often identify a left-to-right shunt, but quantitation is not precise. Intraoperative TEE has become standard of care during adult cardiac surgery procedures and for guiding intracardiac repair of CHD for patients ≥3 kg in size.

When postoperative cardiac output is low, preload-appropriate fluid administration is the most common hemodynamic intervention. Consideration of what type of fluid and how much to infuse is typically driven by the presence of one or more of the following factors: existing or ongoing blood loss, alteration in vascular capacitance due to vasodilation with rewarming, secondary to an inflammatory response from CPB, and reduced ventricular compliance due to myocardial stunning or ischemia. , Conversely, excessive fluid administration can result in heart failure, pulmonary edema, acute kidney injury, and other organ dysfunction. ,

Given the patient-specific and dynamic variations in physiology after surgery, a target pressure within 15 mmHg of the higher of the two atrial pressures is reasonable. If a patient requires more than 3 L of crystalloid, this should prompt the perioperative team to consider alternative diagnoses. As an example, for a patient whose left ventricular wall is unusually thick or has lost its contractility or compliance, it may be helpful to raise mean left atrial pressure to 20 mmHg. However, the tendency toward pulmonary edema is increased when left atrial pressure is elevated to this level. When the RV is the limiting factor in cardiac performance, right atrial pressure usually can be raised advantageously only to about 18 mmHg. Above this, a descending limb on the Starling curve usually becomes apparent, and cardiac output falls. Also, the tendency toward whole-body fluid retention, pleural effusion, and ascites is increased by high right atrial pressure. When left ventricular performance is the limiting factor and systemic arterial blood pressure is more than 10% above normal (see Table 4.14 ), vasodilating agents should be used to reduce left ventricular afterload to a level between normal and 10% above normal.

RV dysfunction refers to impaired RV filling or ejection, but it is not associated with overt signs or symptoms of heart failure. RV failure describes reduced forward flow through the pulmonary circulation, resulting in a low cardiac output state, as evidenced by hypotension, hypoxemia, cool extremities, neurologic manifestations, and signs of systemic congestion (jugular venous distension, hepatojugular reflux, peripheral edema, pericardial effusion, oliguria, congestive hepatopathy/splenomegaly, ascites, anasarca). Acute RV failure is a “…progressive, often rapid, syndrome characterized by systemic congestion secondary to impaired RV filling and/or reduced RV flow output.” Both RV dysfunction and RV failure limit cardiac performance and are associated with significant morbidity and mortality in postoperative cardiac surgery patients. ,

Contemporary management of RV dysfunction or RV failure includes optimizing heart rate (sinus rhythm, ideally) and volume status (right arterial pressure of 5–12 mmHg), avoiding hypotension (MAP >65), and using inotropes, inodilators, and/or vasodilators to reduce RV afterload to improve cardiac stroke volume. ,

There is insufficient information to determine the superiority of one agent over another. Traditionally, nitroprusside (0.5–3.0 µg/kg/min), nitroglycerin (0.5–3.0 µg/kg/min), or phentolamine (1.5–2.0 µg/kg/min) have been used; however, these agents may not be advisable in the hypotensive patient. Inodilators (agents with inotropic and vasodilatory properties) such as dobutamine and milrinone can improve hemodynamics in RV failure and have the additional benefit of reducing PVR, although an additional vasopressor (vasopressin or norepinephrine) will probably be needed to support systemic blood pressure. ,

In infants, maintaining near-anesthesia for 24 to 48 hours with fentanyl or another intravenously administered agent may minimize paroxysms related to pulmonary artery hypertension and the consequent increase in RV afterload. Alternatively, neonates and infants could be given the long-acting α-receptor–blocking agent phenoxybenzamine. It is administered first at the commencement of CPB. An additional dose is usually given about 12 hours after the patient returns to the ICU.

Heart rate is adjusted to optimal levels when necessary by using atrial pacing, ventricular pacing when AF is present, or AV sequential pacing when AV dyssynchrony or dissociation is present. When tachyarrhythmias are present, pharmacologic means of control may be used.

If these relatively simple measures do not quickly bring cardiac performance to an adequate level, inotropic agents are begun (see Appendix 4B ), despite well-known disadvantages. There is no ideal inotropic agent, nor are there specific indications for specific agents. In their review, Doyle and colleagues classified inotropic drugs according to their effect on intracellular cAMP: cAMP-independent drugs include calcium, digoxin, and α-adrenergic agonists; cAMP-dependent agents include epinephrine, dobutamine, and isoproterenol. These are β-adrenergic agonists that, coupled with dopaminergic drugs (dopamine), have variable effects on peripheral resistance. Phosphodiesterase inhibitors (e.g., milrinone) and calcium sensitizers (levosimendan) , may enhance contractility while producing myocardial relaxation (lusitropism) and relaxing vascular smooth muscle. They are not susceptible to receptor downregulation.

Dopamine, although used less commonly in contemporary practice, may initially be infused at 2.5 µg/kg/min. This dose can be increased to 15 to 20 µg/kg/min if needed; however, if a favorable response is not obtained at 10 µg/kg/min, it is not likely to be obtained at higher doses. A previous notion held that the advantage of dopamine was that it increases renal blood flow and cardiac contractility, but more recent data and expert consensus indicate that it does not deter organ dysfunction and may be associated with increased rates of AF. Dopamine is typically initiated at low doses (2–4 µg/kg/min) to maintain systemic peripheral vascular resistance, whereas higher doses (>6 µg/kg/min) increase peripheral resistance. Dobutamine is gradually added in similar doses and appears to augment myocardial blood flow and afterload reduction to a greater degree and blood pressure to a lesser degree.

Occasionally, hypotension exists in the presence of normal and adequate cardiac output. Under that special circumstance, norepinephrine can be administered through a central venous catheter; a low initial dose (0.01–0.05 µg/kg/min) is often sufficient. In cases of progressive shock, larger doses can be used, but potentially at the expense of digital ischemia, mesenteric ischemia, and other complications. Vasopressin can be administered alone as a first-line agent or in combination (0.01–0.06 U/min) with other agents to achieve desired blood pressure in patients who develop vasoplegic shock after cardiac surgery. Although vasopressin may improve clinical outcomes, its superiority over norepinephrine is debatable.

Epinephrine is the catecholamine of choice for some, but its powerful vasoconstricting effects at higher doses necessitate thoughtful use, as any inotropic agent has the potential to increase myocardial oxygen demand and is potentially arrhythmogenic. When an insufficient response is obtained from other drugs, or if excessive tachycardia develops, epinephrine is added or substituted. The drug is initially infused at a dose of 0.01 to 0.05 µg/kg/min, which may be increased as needed.

Milrinone is also useful in patients with low cardiac output after cardiac surgery because it combines a peripheral vasodilatory action with its inotropic effect. This drug differs from catecholamines in structure and mode of action in that it is a phosphodiesterase enzyme inhibitor in cardiac and vascular (including pulmonary) tissue, not a β-adrenergic receptor agonist. Milrinone can be used to treat RV dysfunction or pulmonary hypertension, but it may require coadministration with a vasopressor to maintain an adequate MAP. In addition, the half-life of milrinone (30–60 min) and its impact on platelet function must be considered before it is used. , Administration can be initiated with a loading dose of 5 µg/kg over 10 minutes (though often initiated without a bolus to minimize hypotension), followed by a maintenance dose of 0.125 to 0.5 µg/kg/min.

Additionally, 10% calcium chloride is administered in a dose of 0.1 mmol/kg, with supplemental doses, if the ionized serum calcium level is lower than 1.2 mmol/L.

Implantable VADs, whose use was first described by Cooley and colleagues in 1969, are used for support after cardiac operations, as a bridge to transplant, and as a bridge to recovery. Shorter-term, temporary cardiopulmonary support devices are referred to as ECLS, defined as a “…set of therapies that focus on oxygenation, carbon dioxide removal, cardiac support, or a combination thereof.” ,

Low cardiac output accompanying reduced ventricular function after cardiac surgery occasionally prohibits separating the patient from CPB. A thorough investigation for correctable causes of low cardiac output should be made and may include the use of Doppler ultrasound flow velocity or electromagnetic flow probes to ensure aortocoronary bypass graft patency and adequacy of flow, and the use of intraoperative TEE to evaluate the accuracy of repairs and segmental ventricular wall motion. Preload and afterload should be optimized, and appropriate pharmacologic agents should be administered. When these measures are insufficient, patients with preoperative shock or anticipated postoperative difficulty in being weaned from CPB may require a short-term ECLS device. Insertion of a long-term implantable VAD under these circumstances is inadvisable.

Temporary support of the failing circulation allows further major operative intervention to be postponed until the patient’s condition improves. The temporary assist system can be used as a bridge to ventricular recovery or as a bridge to more durable mechanical circulatory support if weaning is not possible. The options, timing, and management of temporary ECLS have evolved over the past 60 years, such that we now have access to a range of devices, allowing for a more nuanced approach to the treatment of medically refractory cardiac and cardiopulmonary failure.

The ECLS devices presently available include the intra-aortic balloon pump (IABP), ECMO, and percutaneous or central VADs. With any short-term ECLS device, the use should be considered in the context of the patient’s age, comorbidities, neurologic function, and prospects for long-term survival and quality of life. The following discussion relates to use of various temporary ECLS devices for cardiopulmonary failure after cardiac operations.

The IABP is the most widely used temporary ECLS device today. First described by Moulopoulos and colleagues in 1962, the IABP was designed to produce diastolic augmentation of coronary and systemic blood flow. The IABP employs the principle of diastolic counterpulsation, which augments diastolic coronary perfusion pressure, reduces systolic afterload, favorably affects the myocardial oxygen supply/demand ratio, and augments cardiac output. Kantrowitz and colleagues performed the first IABP procedure in humans in 1968.

An IABP is used in adult patients with inadequate cardiac performance not responsive to optimized preload, afterload, and heart rate or to moderate doses of inotropic and vasopressor support (see Appendix 4B ). The rapid inflation of the IABP during diastole provides augmented diastolic coronary perfusion pressure; rapid deflation immediately presystole reduces left ventricular afterload (through both passive and active mechanisms with a Venturi-like effect) and left ventricular myocardial oxygen requirement. This action enhances coronary perfusion and, by optimizing left ventricular hemodynamics (particularly factors related to ventricular interdependence), an IABP may also augment RV function. The absolute magnitude of support provided by an IABP depends on a multitude of factors (including balloon size and optimal positioning, aortic and systemic vascular compliance, and heart rate), but even in the best circumstances, remains relatively modest, at 0.5 to 1.0 L/min of additional cardiac output.

Whenever possible, the decision to insert an IABP (if not already placed preoperatively) is made in the OR rather than postoperatively. Postoperatively, an IABP may be preferable to higher-dose catecholamines for patients with severe left ventricular dysfunction with inadequate cardiac output with preserved oxygenation or severe ventricular arrhythmias. This technique has led to the survival of some patients who would otherwise have died ; however, more recent trials and guidelines have questioned routine IABP use and its impact on longer-term survival. , ,

An IABP is commonly inserted in cases of MI with low cardiac output or shock. Preoperative insertion may also be helpful in patients with unstable angina, left main disease with ongoing ischemia or ischemia leading to ventricular arrhythmia. In the era of more complex arterial revascularization for ischemic heart disease (see Chapter 9 ), IABP support is helpful intraoperatively for pre-bypass support in patients with low ejection fraction. For patients with acute mitral regurgitation or ventricular septal rupture, insertion upon diagnosis is often lifesaving. Perhaps the most frequent indication for preoperative IABP insertion is poor perfusion from low cardiac output. Notably, better survival rates have been found when the IABP was placed preoperatively. ,

Arterial puncture is generally used for IABP insertion, despite the somewhat higher prevalence of vascular complications, because of the technical ease of balloon insertion and removal. When the patient is still on CPB (or the femoral pulse cannot be palpated because of hypotension), ultrasound guidance is advisable. If ultrasonography is not available, the traditional technique of making a small incision in only the skin over the femoral artery can be used. Through this incision, the femoral artery can often be palpated, and the arterial puncture technique applied.

Contraindications for IABP placement include moderate or worse aortic valve insufficiency, thoracic aortic aneurysmal disease, and acute or chronic aortic dissection. In addition, significant aortoiliac occlusive disease and thoracic or abdominal aortic aneurysm greatly increase the risk for vascular complications or insertion failure when the femoral route is used ; in these instances, circulatory support by ECMO should be considered.

The timing of the IABP begins in a 1:1 (heartbeat to balloon activation) ratio with ventricular diastole, as judged by ECG and arterial pressure pulse signals, and is largely automated with modern devices. The timing of the IABP balloon is important, as early or late deflation reduces the effectiveness of the device and can increase left ventricular afterload, thereby increasing myocardial oxygen demand.

Once the patient’s hemodynamic state improves, consideration can be given to weaning the patient from the IABP, possibly as early as 6 to 12 hours after insertion, provided that infusion of inotropic and vasopressor agents has been stable or reduced. The IABP ratio is then progressively reduced to 1:2 or 1:3, along with ongoing reassessment of the patient’s physiologic response to reduced support. In most postoperative patients, the final reduction can be achieved within 12 to 48 hours.

The balloon can then be removed by using either a closed method or a preclosure device. Most vascular complications occur when the balloon is inadvertently inserted through the superficial rather than the common femoral artery. After removing the balloon percutaneously using the closed method, firm pressure is applied to the groin and held for half an hour. If circulation to the leg becomes impaired or a hematoma becomes apparent, prompt exploration in the OR is indicated.

Circulation in the leg distal to the site of balloon insertion is observed systematically. If signs of ischemia appear, generally, the balloon is removed.

Since the first descriptions of CPB in the 1950s and subsequent reports of successful use outside of the OR for ECMO in the 1970s, there has been a rapid growth in the use of this technology in patients with extreme cardiopulmonary failure. Other indications for ECMO include biventricular failure of multiple etiologies, massive pulmonary embolus or other refractory obstructive shock states, and warming the severely hypothermic patient. ,

The primary ECMO configurations are venoarterial (VA-ECMO) and venovenous (VV-ECMO), which are used for cardiopulmonary support and respiratory support, respectively. In this chapter, the focus is on the use of VA-ECMO for cardiac surgery patients with postcardiotomy shock. In current VA-ECMO systems, a centrifugal pump removes blood from the patient’s venous system via an inflow cannula. The blood is oxygenated and ventilated through a temperature-regulated polymethylpentene hollow-fiber oxygenator/heat exchanger; it is filtered and returned to the patient’s arterial system via an outflow cannula.

Cannula placement for venous drainage and arterial inflow is limited only by vessel size and accessibility. Depending on cannula size and position (i.e., central or peripheral cannulation), flows of 5 to 10 L/min are achievable. In the postcardiotomy patient, a central cannulation strategy is an optimal first-line strategy due to the convenience of being able to switch the existing CPB circuit while leaving the aortic and central venous cannulae in place.

Several key management aspects distinguish ECMO from other ECLS therapies. The use of ECMO is associated with both bleeding and thrombotic complications. As VA-ECMO requires an oxygenator, it generally necessitates anticoagulation (except for short periods) due to the risk for oxygenator thrombosis and arterial thromboembolism, in particular when the patient is not on full support. The converse also is true: Using anticoagulation to maintain ECMO integrity increases the risk of bleeding complications.

As with the other ECLS therapies, ECMO can provide full cardiopulmonary support, but it does not completely drain all blood returning to the heart. Further, in peripherally cannulated ECMO, the retrograde nature of blood return can increase afterload and, therefore, left ventricular distention. Patients with very poor left ventricular systolic function have an increased risk for blood stasis and intracardiac thrombosis. Therefore, if the LV is distended, the patient will probably require support with inotropes to help maintain ejection, targeting a minimum pulse pressure of 10 to 15 mmHg. If pharmacologic support is insufficient, the next step is to consider IABP placement, atrial septostomy, or a surgical left ventricular drainage vent for left ventricular decompression. More recently, the use of a transaortic VAD device (such as the Impella®, described later) in conjunction with ECMO has been reported to provide effective emptying of the LV and additional afterload reduction.

When patients who are peripherally cannulated have residual or recovering heart function but insufficient lung function, deoxygenated blood can be ejected from the heart. As a consequence, oxygenated and deoxygenated blood can mix in the aorta, resulting in a “north-south” or “Harlequin” syndrome due to a hypoxic upper body (including heart and brain) and hyperoxic lower body. , In addition to clinical findings, detecting this phenomenon requires arterial sampling from a catheter placed in the right arm (the left arm might be accurate, depending on where the interface between oxygenated and deoxygenated blood exists within the aorta). If this condition is severe and persistent and requires ongoing support, the team may need to consider adding a venovenous arterial cannula (for arterial VV-ECMO to improve venous drainage from the right heart) or converting to central cannulation. ,

When planning for weaning from VA-ECMO, the perioperative team should assess whether both biventricular performance and pulmonary function are adequate to ensure successful decannulation of the ECMO circuit. There is no one universal methodology on how best to perform this assessment , ; however, an institutional procedure designed around a multidisciplinary team should be developed.

In regard to outcomes related to the use of VA-ECMO for postcardiotomy shock, high-quality evidence supporting its use remains sparse. Data from contemporary published series indicate that 40% to 60% of these patients are decannulated from ECMO, whereas only 20% to 40% survive to hospital discharge. Given that ECMO is typically used in emergency and salvage situations as a bridge to recovery, durable device implantation, or transplantation, perhaps this outcome is not surprising; that said, the alternative would be certain death. In light of these difficult decision-making situations, it is increasingly advocated that expertise from supportive and palliative care professionals be sought early in the process of supporting any patient with ECLS. Importantly, one study has indicated that those patients who survive beyond 30 days of ECMO support have reasonable longer-term (5-year) survival and better health-related quality of life.

Catheter-mounted temporary VADs were described as early as 1975, with the first in-human use reported in 1990. , The microaxial concept evolved into today’s Impella pump (Abiomed, Danvers, MA), initially made available for clinical use in 2000. The Impella system includes an external controller and a catheter-based pump that houses the electrical power and controller connections and an inlet for continuous heparin or bicarbonate flushing of the implanted motor. The pump comes in two sizes for left ventricular support (CP and 5.5) and one size for RV support (RP). Support with the larger CP device is intended for short duration (generally 4–6 days), whereas the larger 5.5 device can be used for up to 14 days.

Impella devices can be inserted through a peripheral artery (either femoral or axillary, although axillary access better facilitates extubation and mobilization) or directly into the ascending aorta over a wire and positioned retrograde through the aortic valve. The RP version is inserted through the femoral vein and right heart to the pulmonary artery, unloading the right ventricle into the pulmonary artery. These devices can be inserted with fluoroscopic guidance in a cardiac catheterization laboratory, hybrid OR, or a standard OR with portable fluoroscopy. Echocardiographic guidance is typically added for more precise positioning.

Impella devices are contraindicated for patients with severe aortic stenosis or insufficiency, left ventricular thrombus, tamponade or left ventricular rupture, conditions precluding placement (such as a mechanical aortic valve or severe obstructive peripheral vascular disease), or a contraindication for anticoagulation. Contraindications specific to the RP device include major venous or RV thrombus, inferior vena cava filters, and anatomic anomalies of the pulmonic or tricuspid valve that preclude placement.

The degree of support is titratable and can be set on the controller by adjusting the RPMs by set levels (P1–P9). The perioperative team needs to monitor the position of the device across the aortic valve; current devices accomplish this automatically by using differential pressure sensors. Also important is monitoring for potential complications, such as hemolysis, limb ischemia, aortic or mitral valve injury, stroke, arrhythmia, and vascular injury. , Recovery of cardiac function can be assessed by scaling down the support level: The device can be removed when satisfactory hemodynamics and organ perfusion are maintained at low (P1 or P2) settings over a period of hours.

Similar to the IABP, Impella CP pumps are commonly used as a means of support in centers without surgical VAD capability, either as a bridge to recovery or to allow safer transportation to a VAD center. The utility of this device continues to be investigated and refined. One RCT comparing the efficacy of a previous-generation device (Impella LP 2.5) with an IABP in nonemergency, high-risk percutaneous coronary intervention (the PROTECT II Study) was stopped early due to concerns about futility; nonetheless, the trial did reveal a trend toward superior hemodynamic support and fewer adverse events with the Impella. Similarly, in an RCT investigating the management of acute MI complicated by cardiogenic shock (the EMPRESS Trial), the Impella was not superior to an IABP in terms of 30-day mortality. An ongoing RCT (SURPASS Impella 5.5 trial; registration NCT05100836) is examining the impact of the Impella 5.5 device on outcomes.

The TandemHeart® (LivaNova, London, UK) is a percutaneous, centrifugal left-side VAD. It uses a 21F drainage catheter inserted through the femoral vein and right atrium and into the left atrium via a transseptal puncture. A centrifugal pump and control console unload the left heart, decreasing left ventricular end-diastolic pressure and myocardial oxygen consumption. A separate 15F to 19F arterial catheter inserted into the common femoral artery provides the inflow in a retrograde direction. In essence, this approach is similar to percutaneous ECMO, with the key difference being that the TandemHeart does not require an oxygenator (although one can be used if required).

Contraindications for TandemHeart use include right or left atrial intracardiac thrombus, severe peripheral arterial or aorto-iliac atherosclerotic disease that precludes safe arterial cannula placement, or contraindication for anticoagulation. Aside from the potential access site issues of bleeding and limb ischemia, device-specific complications include left or right atrial perforation causing tamponade during insertion, cannula migration causing selective pulmonary vein intubation, tamponade from perforation of the left atrium, or shunting from backward migration into the right atrium. , Accordingly, many institutional protocols require fixation of the device and cannulae and limitation of leg mobility, in addition to routine mechanical circulatory support care, including monitoring for distal limb ischemia or emboli, insertion-site bleeding and hematoma, and evidence of hemolysis.

In patients with improved native heart function and minimal inotropic support requirements, weaning from the TandemHeart is accomplished by decreasing the pump speed and assessing hemodynamic response over time. Data on the effectiveness of this device versus others continue to evolve; one small meta-analysis failed to demonstrate the benefit of this device over an IABP.

The CentriMag (Abbott Laboratories, Abbott Park, IL) and Rotaflow (Getinge, Gothenburg, Sweden) devices (and a new but less-well-studied product, the LivaNova Revolution pump) are examples of intermediate duration centrifugal pump systems that can be used as bridge-to-decision mechanical circulatory support (either recovery, palliation, or conversion to a durable VAD). These devices were developed to overcome key issues with older-generation blood pumps, including shear stress resulting in hemolysis and thrombus, friction with heat generation, turbulent flow, and stasis. The Rotaflow pump is magnetically suspended on a sapphire bearing with no shafts or seals, whereas the CentriMag is completely magnetically levitated with no bearings, shafts, or seals. Both pumps have very small prime volumes and can be used for left ventricular, RV, or biventricular assist support. In addition, an oxygenator can be used or removed, and initial biventricular support can be simplified to univentricular support if partial cardiopulmonary recovery has occurred.

An additional potential advantage is that the use of central cannulation permits the use of shorter, large-bore cannulae that produce better hemodynamics (higher flows, better ventricular decompression). As a result, these systems are usually indicated for patients who present with postcardiotomy cardiogenic shock or who are unable to achieve adequate hemodynamic support from peripheral mechanical circulatory support systems. Specific care for the centrally cannulated mechanical circulatory support patient includes monitoring for bleeding complications, ensuring cannula and/or tubing securement, and avoiding kinking along the length of the extracorporeal tubes during mobilization. Importantly, as the cannulae are fixed to access sites, cannula migration should not be a problem.

In the ICU, moving toward extubation and active mobilization should be the goal when using these devices. Although inability to anticoagulate is a major concern, the flow characteristics and low stasis issues associated with these devices produce very few thromboembolic events, and there have been reports on managing these pumps without anticoagulation for several days to weeks after initial implantation. , The weaning strategy is similar to that of the Impella and TandemHeart devices.

The complex scientific, surgical, ethical, philosophical, and financial considerations necessarily involved in the decision to implement ECLS do not permit a simple listing of indications for these therapies. There is general agreement that survival is negatively affected by a low cardiac output (CI ≤1.5 L/min/m ² ) and elevated ventricular diastolic pressure (reflected in high left or right atrial pressure) when CPB is continued for an hour longer than usual, even with IABP support and catecholamine administration in moderate doses. Trial separation from CPB is likely to fail (usually reflected in a progressive drop of systemic arterial blood pressure, elevation of atrial pressures, reduction in venous oxygen saturation, and metabolic acidosis).

Patients with continuing low cardiac output in the ICU despite IABP and catecholamine support should also be considered for temporary ECLS, even if the patient must be transferred to another institution. Use of temporary ECLS is appropriate for patients of any age whose myocardial complications are expected to resolve within a few days, and for patients younger than 70 years of age whose myocardial complications may not resolve but who are acceptable candidates for durable VAD or cardiac transplant. In the latter group, the temporary device is used for support while it is determined whether the major organ systems are functioning and whether the patient is neurologically intact.

The care of patients on ECLS is labor intensive. Continuous bedside care, often by more than one nurse, is initially required, as is the ready availability of a cardiopulmonary perfusionist or other personnel capable of managing the extracorporeal assist system. Bleeding is frequently a nuisance and can even be a major complication, requiring frequent monitoring of ACT, which is used to monitor heparin effect; the desired ACT is approximately 160 to 200 seconds (assuming a control level of 100–120 seconds). Bleeding from the primary operative site is controlled by reversing heparin with protamine. Platelets, FFP or other blood products, and pharmacologic agents to promote normalization of the blood clotting subsystem are administered as indicated. A heparin-bonded extracorporeal circuit is adequate to prevent clotting in the short term. When bleeding from the operative site has ceased or slowed (usually within 12 hours postoperatively), heparin therapy can be restarted. Many centers now use bivalirudin for anticoagulation during ECMO support as accumulating data demonstrate its safety and possible benefit regarding patient survival.

Smooth operation of the system requires frequent infusion of blood products to maintain adequate atrial pressures. Body temperature is monitored continuously, and measures are taken to heat or cool the blood in the extracorporeal circuit or the patient’s body with a heating or cooling blanket. The ECLS may be fitted with ports for hemofiltration or access for continuous venovenous renal replacement therapy to remove excess extracellular fluid if there is marked edema. Ventilatory assistance is required.

Separating the patient from temporary ventricular assistance requires judgment and patience. There is a tendency to remove systems too early to avoid complications directly related to prolonged extracorporeal circulation. Moreover, simply removing the temporary device in anticipation that the heart will sustain adequate function is usually unsuccessful. Instead, the heart’s ability to support the circulation should be tested by reducing flow into the extracorporeal circuit, thereby raising atrial pressures and allowing flow through the supported ventricle. However, there is a limit to which flow in an extracorporeal circuit can be reduced without introducing danger of clotting. Flow of less than about 1 L/min should be avoided. Anticoagulation levels should be maintained, and the testing interval should be brief. Cardiac function is monitored using echocardiography and continuous measurement of arterial and pulmonary artery pressures, atrial pressures, cardiac output, and Sv o ₂ .

If the heart cannot sustain adequate cardiac output under conditions of low-flow bypass, the separation process should be abandoned, full flow resumed, and plans made for later attempts at separation or conversion to an implantable device. The multidisciplinary team should collaborate on the process and timing of ECLS discontinuation and plan ahead in case the patient is unable to achieve adequate cardiac output in the absence of support. This could, in some instances, include a plan not to re-escalate to ECLS therapy. Thus, it is appropriate to establish a clear and transparent process and to “set the stage” with all team members, including the patient, family, and caregivers. This process should be undertaken early, frequently, and with the appropriate documentation

Ventricular electrical instability includes premature ventricular contractions (PVCs), ventricular tachycardia, and ventricular fibrillation. Whether postoperative patients with frequent PVCs and some types of ventricular tachycardia are at risk for developing ventricular fibrillation remains controversial. Despite this, it is prudent to assume that such arrhythmias soon after surgery place the patient at increased risk for sudden death. Detecting such arrhythmias requires continuous ECG monitoring during the first 24 to 72 hours postoperatively.

Unifocal isolated PVCs occurring outside the T waves fewer than 4 to 6 times per minute does not appear to increase the risk for cardiac death; however, the perioperative care team must remain vigilant for ischemia, hypothermia, abnormalities in potassium (or other electrolytes), and other postoperative complications, particularly in those patients with low left ventricular function or known history of ventricular arrhythmias. Routine use of Class IC antiarrhythmic medication (see Appendix 4C ) is generally not required and indeed may cause harm.

In patients with postoperative ventricular tachycardia and an ejection fraction <40%, early electrophysiologic study may be indicated. , Paradoxically, patients in whom such studies reveal acute suppression of PVCs or ventricular tachycardia with the administration of drugs have a good prognosis without drug treatment; these patients can probably be discharged from the hospital on no drug therapy. By contrast, patients whose ventricular arrhythmias cannot be suppressed with drugs have a relatively poor prognosis, and consideration should be given to an implantable cardioverter-defibrillator.

Often, PVCs or bursts of ventricular tachycardia can be suppressed by using temporary electrodes to establish atrial or ventricular pacing without the need for drug therapy. When the hemodynamic state is impaired by ventricular tachycardia, immediate direct current cardioversion (100 J initially in adults and, if ineffective, 200 J) is indicated. When these measures are ineffective, advice of an experienced electrophysiologist is indicated.

Postoperative atrial tachyarrhythmias remain one of the most common occurrences after cardiac surgery. The perioperative team should be aware of several important scientific statements and guidelines, in particular the ACC, American Heart Association (AHA), and Heart Rhythm Society (HRS) guideline for managing AF, the AHA scientific statement on AF occurring during acute hospitalization, the European Society of Cardiology and European Association for Cardio-Thoracic Surgery guidelines for diagnosing and managing AF, The Society of Thoracic Surgeons 2017 clinical practice guidelines for the surgical treatment of AF, and the Society of Cardiovascular Anesthesiologists and European Association of Cardiothoracic Anaesthetists practice advisory for managing perioperative AF in patients undergoing cardiac surgery.

Postoperative AF in the setting of cardiac surgery is common, affecting 20% to 60% of patients depending on the type of surgery (less prevalent after CABG, more common with isolated or concomitant valve surgery). , It is most likely to occur on postoperative days 2 to 4. The traditional belief has been that new postoperative atrial fibrillation (POAF) is often self-limited and benign, although this dogma has been challenged by new evidence that patients with POAF after cardiac surgical intervention have a fivefold higher risk for permanent AF and ischemic stroke that persists to 10 years after surgical intervention. , , , Further, although findings on the degree of complications are inconsistent, it is generally understood that the development of POAF after cardiac surgery is associated with longer hospitalization, greater short- and long-term morbidity (such as renal complications, infections, and bleeding), , , greater mortality, recurrent hospitalizations, and consequently greater costs of care. , ,

The adequacy of pulmonary function early after surgery, while the patient is still intubated and receiving controlled ventilation, is judged by the response of Pa o ₂ and Pa co ₂ )to the minute volume of ventilation and ventilating gas mixture being used. The response of the arterial oxygen levels can be expressed as the alveolar-arterial oxygen difference, P(A − a)O ₂ , which in intact humans is normally only a few mmHg. Nearly all patients early after cardiac surgery have abnormally large alveolar-arterial oxygen differences due to intrapulmonary right-to-left shunting of 3% to 15%. This assumes that no right-to-left intracardiac shunting is present. Response of the carbon dioxide level can be expressed as the minute volume of ventilation required to maintain Pa co ₂ at 30 to 45 mmHg, usually about 15 to 20 mL/kg in both adults and children.

When emerging from anesthesia in the OR or ICU, the patient can be converted to a spontaneous mode of mechanical ventilation, either synchronized intermittent mandatory ventilation or pressure support ventilation. The patient’s respiratory rate and other hemodynamic parameters can be used to judge the adequacy of cardiac and pulmonary function. In an adult, adequacy of lung function is indicated by a patient-triggered respiratory rate of 8 to 12 breaths/min with a tidal volume of 6 to 8 mL/kg. and adequate oxygenation, ventilation arterial blood gas parameters, and end-expiratory pressure.

PEEP may be used routinely or as indicated, with a setting of 5 to 8 cm H ₂ O for adults and children older than 12 years and 4 cm H ₂ O for younger patients. Unless the hemodynamic state is suboptimal, using PEEP does not alter it, even in infants. Studies suggest that PEEP is associated with larger lung volumes, fewer perfused but nonventilated alveoli during ventilation, and smaller P(A − a)O ₂ after extubation.

Appropriate PEEP levels need to be considered in patients with chronic obstructive lung disease (to avoid air trapping and barotrauma in those with bullous lung disease) and in infants and children who have undergone a Fontan operation or cavopulmonary anastomosis (to avoid further elevation of jugular venous pressure).

Continuous positive airway pressure (CPAP) , may be used in infants once their cardiovascular state is stable, obviating the need for intermittent positive-pressure breathing and intermittent mandatory ventilation. Should intermittent positive-pressure breathing be required initially, the infant is transferred to intermittent mandatory ventilation and sometimes to CPAP for several hours before extubation.

Traditionally, continued intubation after cardiac surgery has been recommended to maintain more precise control of cardiopulmonary physiology. However, early extubation, either in the first 6 hours in the ICU (fast-track) , or in the OR (ultra-fast-track) facilitated by lower opioid doses and neuromuscular blockade reversal, leads to shorter ICU and hospital stays and potentially lower healthcare costs. , Therefore, a protocol beginning with anesthesia induction can be designed to expedite awakening and spontaneous breathing. This requires perioperative team readiness and a culture of safety to ensure appropriate surgical pain management and preparedness for reintubation, day or night.

Once the patient has been extubated, useful indices of pulmonary function include Pa co ₂ and Pa o ₂ levels, hemodynamics, and visually estimated work of breathing. A Pa co ₂ level <45 mmHg in adults, <50 mmHg in young children, and <55 mmHg in infants generally indicates an adequate minute volume of ventilation. Higher values indicate inadequate alveolar ventilation. Of note, Pa o ₂ is often mildly depressed after extubation when the cardiac surgery was performed with CPB and usually remains so for the first few days ( Fig. 4.14 A and B). This is caused by the somewhat widened alveolar-arterial oxygen difference. These measurements are valuable, but when a patient is comfortable and breathing easily and slowly, and the chest radiograph is within normal limits of a postoperative patient, it is highly probable that pulmonary function is adequate and initial recovery will be satisfactory.

A, Arterial oxygen tension (Pao2 ), venous admixture (Q˙va/Q˙t) , and arteriovenous oxygen content difference [C(a − v)O ₂ ] measured preoperatively and at intervals early postoperatively in 10 adults undergoing operation with cardiopulmonary bypass. B, Preoperative and postoperative values for minute ventilation (V˙ E) , frequency (f), and tidal volume (Vt) in the same patients. Air- and O ₂ -breathing results have been combined, and mean values for the group at each time are shown.

(Data from Rea HH, Harris EA, Seelye ER, Whitlock RM, Withy SJ. The effects of cardiopulmonary bypass upon pulmonary gas exchange. J Thorac Cardiovasc Surg. 1978;75(1):104-120.)

After cardiac surgery with CPB, the lungs are more likely to have some degree of dysfunction, albeit typically mild and transient. Pulmonary dysfunction is caused in part by the absence of pulmonary blood flow during total CPB or by its near absence during partial CPB, either of which produces very low shear stresses in the pulmonary capillaries. This appears to accentuate neutrophil activation because neutrophils appear to be exquisitely sensitive to shear stress. Leukocytes activated by the general damaging effects of CPB incite an inflammatory response in the pulmonary vasculature. During total CPB and lung ischemia, plasma thromboxane B ₂ levels increase, which may contribute to a pulmonary vascular inflammatory response. The cytokines interleukin-6 and interleukin-8 increase with CPB and may also contribute to membrane damage and neutrophil activation in the lung. , The alveolar-capillary barrier becomes more permeable than normal, , and after cardiac surgery with CPB, macromolecules enter the pulmonary interstitium and ultimately the alveoli, promoting the development of pulmonary edema.

Postoperative disturbances in lung function and increases in alveolar polymorphonuclear leukocytes have been linked to an important reduction in pulmonary surfactant activity. In addition, multiple and not-completely-defined factors encourage the development of large areas of atelectasis, either segmental or occasionally lobar; in particular, the left lower lobe has a strong tendency to atelectasis, even in patients who are otherwise recovering normally.

In some patients, pulmonary dysfunction is caused by direct trauma to the lungs. In addition, the high intraoperative tidal volumes (>8 mL/kg) and ventilatory driving pressure (the difference between the plateau pressure and the level of PEEP) have been associated with ventilator-induced lung injury and postoperative pulmonary complications. , ,

A patient’s inability to cough and clear secretions (due to poor inspiratory effort secondary to poor pain control) during or early after surgery also contributes to dysfunction. Left (and occasionally, right) phrenic nerve injury may occur even after carefully performed cardiac operations, increasing the tendency to pulmonary dysfunction early after surgery. However, left lower lobe atelectasis is considerably more common than left phrenic nerve paralysis. Markand and colleagues found the left phrenic nerve to be paralyzed in only 11% of patients who developed left lower lobe atelectasis early after open-heart operations. In most patients, the left phrenic nerve recovers within 6 months of paralysis. In patients with persistent diaphragmatic paralysis, diaphragm pacing is being investigated. ,

Mechanical obstruction of the lower trachea or the bronchi can produce pulmonary dysfunction that may go unrecognized unless care is taken to identify and treat it. Localized or more extensive pulmonary edema may develop in the presence of low or normal left atrial pressure, no doubt related to changes in pulmonary venular and capillary permeability, the causes of which are only partially understood. This phenomenon seems to be more marked in older adults. Less commonly, frank pulmonary hemorrhage that develops as CPB is discontinued or early thereafter can cause serious bronchial obstruction and contribute in a major way to pulmonary dysfunction. This is probably a more severe result of the same factors that lead to pulmonary edema in patients with normal or low left atrial pressures.

Patient-specific risk factors for pulmonary dysfunction after cardiac surgery have long been recognized, but formal identification and quantification are rare. In an observational study, young age at operation was identified as a risk factor, particularly when the patient was younger than about 2 years old ( Fig 4.15 ). Lell and colleagues made similar observations. Higher risk at a young age is associated with the increased tendency of the very young to develop whole-body edema after CPB. A study by Gallagher and colleagues, supported by broad anecdotal experience, found that older age , particularly older than 60 years, also has been associated with a higher prevalence of pulmonary dysfunction after cardiac surgery.

Relationship between age at operation and duration of CPB (represented by solid isobars and their dashed 70% confidence limits) to probability of pulmonary dysfunction after cardiac surgery. Nomogram depicts specific solutions of a multivariable equation; value entered for C3a was 882 ng · mL ⁻¹ (see original publication for details).

(From Kirklin and colleagues. )

Chronic obstructive lung disease is an important risk factor for postoperative pulmonary dysfunction and increases the overall risk of the operation because it predisposes patients to increased work of breathing and air trapping. Preoperative pulmonary arterial hypertension , even when associated with low pulmonary arteriolar resistance, predisposes infants to pulmonary dysfunction and also to paroxysms of pulmonary arteriolar constriction and pulmonary hypertension soon after surgery. Congenital morphometric pulmonary abnormalities, such as alveolar hypoplasia frequently present in patients with CHD or Down syndrome, increase the prevalence of postoperative pulmonary dysfunction.

Protective mechanical ventilation strategies using low tidal volume or high PEEP levels improve outcomes for patients who have had surgery. High intraoperative ventilatory driving pressure (defined earlier) and changes in the level of PEEP are associated with more postoperative pulmonary complications. However, traditional belief has been that low tidal volumes promote alveolar collapse in poorly ventilated, dependent regions of the lung with resultant lung trauma secondary to the repetitive collapse and reopening of alveolar units, leading to ventilator-induced lung injury. , As a result, there has been some enthusiasm for the use of an “open-lung strategy.”

The open-lung strategy refers to ventilation management that uses mechanical ventilation during surgery along with recruitment maneuvers to prevent alveolar collapse during the procedure. , The multinational Protective Ventilation in Cardiac Surgery (PROVECS) trial, an RCT of 493 patients, investigated whether an open-lung perioperative ventilation strategy that combined mechanical ventilation during CPB, perioperative recruitment maneuvers, and higher PEEP levels was protective against postoperative pulmonary complications, compared with usual care. The intervention was not found to be superior. Another trial (FLOWVENTIN HEARTSURG; German Clinical Trials Register DRKS00018956) is investigating whether a newer ventilation mode that increases control of inspiratory and expiratory airway flows will result in less postoperative pulmonary dysfunction.

The type of oxygenator used is associated with the amount of pulmonary dysfunction generated by the operation, with oxygenators other than membrane-type oxygenators being risk factors for pulmonary damage. This is less common with contemporary oxygenators used in clinical practice. The use of filters in the arterial tubing may reduce pulmonary dysfunction postoperatively. Longer-duration CPB is also a risk factor, probably related to its direct correlation with an increase in the patient’s extracellular water , ( Fig. 4.16 ). A higher level of C3a generated by complement activation during CPB is a risk factor that may be related to elevated neutrophil activation. Use of external cardiac cooling , particularly by ice slush rather than cold saline, increases the prevalence of left phrenic nerve paralysis and hence the tendency toward postoperative pulmonary dysfunction.

Relationship between duration of CPB and increase in interstitial fluid (ECF-PV) 4 to 6 hours after operation. The x’s represent patients undergoing closure of left-to-right shunts; circles represent those undergoing operation for valvar heart disease. Patients with heart failure were not included. The regression equation is: Ln (ECF − PV) = −3.27 + 0.83 Ln (CPB time); r = 0.86; P < 0.001 . CPB, Cardiopulmonary bypass; ECF, extracellular fluid; Ln, natural logarithm; PV, plasma volume.

(Data from Cleland J, Pluth JR, Tauxe WN, Kirklin JW. Blood volume and body fluid compartment changes soon after closed and open intracardiac surgery. J Thorac Cardiovasc Surg. 1966;52(5):698-705.)

Postoperative events can increase the probability of pulmonary dysfunction. Elevated left atrial pressure , with consequent elevation of pulmonary capillary and venular pressures, aggravates the tendency toward increased lung water and pulmonary dysfunction. The longer the patient is on a ventilator, the greater the chances of ongoing pulmonary dysfunction. Elevating the left hemidiaphragm because of phrenic nerve paralysis predisposes patients to continuing pulmonary dysfunction, particularly small patients. ,

The treatment goal for the pulmonary subsystem is the patient’s earliest possible return to extubated spontaneous breathing and ambulation. These benefits include a decline in intrapleural pressure and an increase in left ventricular end-diastolic diameter (or volume), improved ventricular systolic function (due to increased preload associated with the shifting of some blood volume toward the chest), and improved cardiac output.

For patients with ARDS, ICU and mechanical ventilation management mirrors the principles of management in critically ill noncardiac patients. Maintaining low tidal volume ventilation is the mainstay of therapy, and vigilant fluid management, minimization of blood product transfusion, appropriate nutrition, and early physical rehabilitation may improve outcomes. , In cases of refractory hypoxemia, rescue therapies such as inhaled nitric oxide or fixed-dose epoprostenol (although of unclear survival benefit), neuromuscular blockage, recruitment maneuvers, and escalation to ECMO (in appropriately selected patients) may be required. , , ,

Stable patients are usually extubated either in the OR or during the early hours after the operation. Otherwise, if pulmonary function is good, extubation should be delayed only under special circumstances, including the presence of cardiac assist devices and the possibility of early reoperation. Early extubation can prevent complications such as ventilator-associated infections and prolonged ventilation dependence.

In general, criteria for extubation include stable and satisfactory cardiac performance, lack of important cardiac arrhythmia, and appropriate awakening with satisfactory neurologic status. There should be no anticipation of return to the OR (e.g., for bleeding) and satisfactory mechanics and ventilatory lung function, as assessed by arterial blood gas analysis and a clinical estimate of inspiratory force and volume. Increasing emphasis has been placed on the advisability of routinely extubating the normally recovering patient within 6 hours of arrival in the ICU. , , Using standardized protocols that leverage the interdisciplinary team can facilitate fast-track extubation after uncomplicated cardiac surgery procedures.

Atelectasis and the associated loss of functional alveolar units may be particularly problematic in patients with pre-existing obstructive or restrictive lung disease. In such patients, the Fi o ₂ should be titrated to target Pa o ₂ >70 mm and normocarbia, with a goal pH of 7.35 to 7.45. Standard prophylactic measures include chest physiotherapy and use of devices to encourage deep inspiration to recruit atelectatic areas. , Several studies have found that using noninvasive ventilation and maintaining prophylactic nasal positive airway pressure at 10 cm H ₂ O for at least 6 hours after cardiac and aortic procedures may avoid the need for reintubation in higher-risk patients. Recent findings suggest that the use of a high-flow nasal cannula (without additional dedicated positive pressure) also may maintain oxygenation and mitigate against the need for reintubation ; a large ongoing RCT (Reducing Reintubation Risk in High-Risk Cardiac Surgery Patients With High-Flow Nasal Cannula) is examining the utility of this approach in the postoperative cardiac surgery patient.

Treating patients with pulmonary dysfunction who are still intubated during the early hours after surgery is largely an intensification of usual management, as specific effective therapy is not available. The Fi o ₂ level is appropriately adjusted upward if P(A − a)O ₂ is large. Minute volume of respiration (not only measured but also judged visually by chest wall excursion) and respiratory rate are adjusted to maintain Pa co ₂ at 30 to 35 mmHg. The airway is kept clear by appropriate tracheal suctioning.

When hemoconcentration develops because of plasma leakage from the intravascular space into the interstitial space of the lungs (and at times into all other organs and the pleural and peritoneal spaces), administering concentrated serum albumin can help to counteract this trend; however, the benefit of colloids after cardiac surgery remains unclear. , The timely use of diuretics is useful because they reduce extracellular fluid volume and, in turn, extravascular lung water; nonetheless, appropriate flow to the kidneys must be preserved to avoid a prerenal acute kidney injury.

All the while, efforts to reduce ventilatory support and work toward extubation of the patient must continue (see Appendix 4F ). Conducting spontaneous breathing trials and decreasing levels of pressure support appear equally effective in patients who are difficult to wean from the ventilator.

A prolonged period of endotracheal intubation is necessary when:

• •

  Criteria for extubation are not met

• •

  Neurologic complications are present

• •

  Severe dysfunction or hemodynamic instability of the cardiac subsystem is present

• •

  Persistent chest drainage or a residual cardiac defect makes early return to the OR likely

During periods of prolonged ventilation, active decision-making is required regarding use of continuous versus intermittent sedation and neuromuscular blockade, depending on hemodynamic stability, effectiveness of ventilation, and timing of ventilator weaning. Commonly used sedation and paralytic agents are described in Tables 4.16 and 4.17 .

When patients have agitated delirium after cardiac surgery, particularly during ventilator weaning, the priority is evaluating cardiac and pulmonary subsystems to ensure that agitation is not an indicator of important dysfunction. , If these subsystems have satisfactory function, the next level of intervention targets communication between the patient and nursing or physician staff or both to ensure that adequate pain management has been achieved and nonpharmacologic approaches to allay patient anxiety are in use. When these interventions are insufficient, using pharmacologic agents at the lowest effective dosage can be considered. These may include propofol, dexmedetomidine, and, in patients with a known history of substance abuse or withdrawal, benzodiazepines (e.g., diazepam, lorazepam, midazolam) used judiciously. , Antipsychotic medication (e.g., haloperidol) should be reserved for patients whose pain is well managed but who remain at a safety risk to themselves or the healthcare provider team.

In those unusual circumstances when intubation in older children and adults is necessary for more than 7 to 10 days, consideration is given to tracheostomy. Some data suggest that early tracheostomy may be of benefit in patients who are likely to require prolonged mechanical ventilation after their surgery. , Tracheostomy has few disadvantages, and it usually increases both patient comfort and ventilation effectiveness. Further, tracheostomy can be helpful in weaning the patient from the ventilator and is not associated with higher rates of sternal infection. , , Tracheostomy is rarely performed in neonates, infants, and young children because it is difficult to manage in this age group.

When the patient meets all criteria for extubation, reintubation is usually unnecessary. However, positive pressure ventilation or reintubation should be considered when:

• •

  Pa co ₂ increases to >50 mmHg over 4 hours in the absence of appropriate respiratory compensation for metabolic alkalosis.

• •

  Signs of decreasing cardiac output are noted.

• •

  Signs of exhaustion from spontaneous breathing are observed.

• •

  Excessive pulmonary secretions with ineffective coughing occur.

When the situation is borderline, proper management requires careful observation by senior members of the team. A competent professional should perform reintubation.

Three different classification systems have been used to define CSA-AKI ( Table 4.18 ). The Risk, Injury, Failure, Loss, and End-stage (RIFLE) kidney disease classification defines three severity grades for CSA-AKI (risk, injury, and failure) based on changes to serum creatinine and urine output and two clinical outcomes (loss, end-stage). , The Acute Kidney Injury Network (AKIN) group sought to increase the sensitivity of the RIFLE criteria by recommending that a smaller change in serum creatinine (≥26.2 µmol/L) be used as the threshold to define the presence of AKI. In 2012, the Kidney Disease: Improving Global Outcomes (KDIGO) AKI working group published further criteria for defining and classifying AKI that is meant to serve as a combination of AKIN and RIFLE definitions. Specific to cardiac surgery patients are:

• •

  Increase in serum creatinine by ≥26.5 μM (0.3 mg/dL) within 48 hours; or

• •

  Increase in serum creatinine to ≥1.5 times baseline within 7 days.

Preoperative renal function impairment that is intractable to therapy considerably increases the risk for acute renal failure soon after surgery. Therefore, preoperative evaluation includes assessment of renal function.

The predictive value of serum creatinine and GFR on the risk for postoperative dialysis has been quantified by Mehta and colleagues for adult patients undergoing cardiac surgery. Although GFR is traditionally the standard measure of renal function, its formal measurement is frequently not feasible in the clinical setting. A useful approximation of creatinine clearance can be made with the Cockcroft-Gault equation , :

where Cr = creatinine. For women, multiply the result by 0.85 to account for smaller muscle mass.

Chronic heart failure (preoperative NYHA class IV) considerably increases postsurgical risk for acute renal failure, as does congenital heart disease after cardiac surgery in older patients. A renal lesion is known to exist preoperatively in many such patients. This does not appear to be the case in young patients; for example, renal failure is rare in infants and children undergoing repair of tetralogy of Fallot. Further, when cardiac output is importantly reduced after cardiac surgery in neonates and infants, acute renal failure rates as high as 8% to 10% have been reported, , possibly because immature kidneys may have less ability to concentrate urine in the context of reduced renal blood flow. Compared with older patients, infants may develop more tissue hypoxia during and early after CPB, with a resulting increase in production of potassium, blood urea nitrogen, and other substances, some of which may be nephrotoxic. Uric acid levels, for example, were shown by Hencz and colleagues to rise to nephrotoxic levels (10 mg/dL) in some patients within 24 hours of operation, and the levels were significantly higher in patients younger than 3 years of age than in older children.

In adults, the development of CSA-AKI results from the interplay between patient susceptibility to renal injury and perioperative renal insults. Multiple factors can affect the kidney, including hemodynamic instability, inflammation, hemolysis, and nephrotoxic agents. In the setting of cardiac surgery, it is believed that renal injury arises intraoperatively during CPB. , Indeed, a renal tubular injury signal is measurable shortly after the initiation of CPB.

The evolution of ischemic CSA-AKI is conceptually divided into initiation, extension, maintenance, and repair phases. The initiation phase is the period immediately after an ischemic insult initiates reversible tubular injury and GFR drops abruptly. At the tissue level, ATP depletion, reactive oxygen species generation, and inflammatory mechanism induction contribute to sublethal epithelial and endothelial damage. , , If the ischemic insult is brief and mild, the injury remains sublethal, and organ function recovers rapidly. Prolonged or severe ischemia followed by reperfusion, however, leads to an extension phase characterized by microvascular damage, tubule cell death, desquamation, and luminal obstruction.

Further decline in GFR ensues, and the injured endothelial and epithelial cells amplify inflammatory cascades. The maintenance phase is characterized by a dynamic equilibrium between waning injury and evolving repair mechanisms. The balance between cell survival and death determines the severity and duration of this phase.

Karkouti and colleagues conducted a study of 3500 adult cardiac surgery patients from seven hospitals to identify potentially modifiable risk factors associated with CSA-AKI. Using multivariate logistic regression analysis, the investigators identified three potentially modifiable variables that were independently and strongly associated with AKI: preoperative anemia, perioperative RBC transfusions, and surgical reexploration. In addition to these potential targets, several recent clinical practice guidelines and expert consensus statements have been generated to provide an evidence-based synthesis of strategies for preventing CSA-AKI. , An example of a preoperative, intraoperative, and postoperative approach provided by the Harrington Heart and Vascular Institute–University Hospitals Taskforce on Preventing CSA-AKI in the Perioperative Period identifies six components for the perioperative team to consider (see also Appendix 4G ):

• 1.

  Preoperative: Identify patients at risk for CSA-AKI (low/moderate/high risk) and provide patient medication education.

• 2.

  Intraoperative: Identify AKI risk factors during OR timeout and consider medications and hemodynamic management.

• 3.

  Postoperative: Assess risk factors for AKI handed over to the team upon ICU arrival.

  • a.

    Implement goal-directed therapies for high-risk patients by using a hemodynamic algorithm in addition to blood sugar and medication management.

  • b.

    Implement ongoing evaluation of renal function, hemodynamic status, and volume status.

  • c.

    Conduct outcome measurement and monitoring.

The following sections provide a framework for this approach to patient management aimed at minimizing the risk for CSA-AKI (see Engelman and Shaw for a protocol example).

Preoperative identification of patients at risk for CSA-AKI is a key step in mitigating it. The perioperative team performs a kidney health assessment , that includes a full medical and medication history, determination of baseline kidney function, a history of previous episodes of AKI after illness, diagnostics, procedures or surgery, recent nephrotoxin exposure, and assessment for anemia. A complete urinalysis should be undertaken, and consultation with a nephrologist is advisable. Further considerations include optimizing glycemic control by maintaining blood glucose at 80 to 180 mg/dL, limiting aminoglycoside antibiotics, using vancomycin judiciously, and holding ACE inhibitors and angiotensin receptor blockers (ARBs) for up to 48 hours preoperatively. In addition, consider holding all nonsteroidal antiinflammatory drugs (NSAIDs) for 1 week before surgery when possible.

The use of a dynamic predictive risk score can help the perioperative team classify patients as having either a low, moderate, or high risk. The patient’s risk for perioperative CSA-AKI will be documented within the medical record and should be discussed at the OR presurgical huddle and during handoff to the ICU team initially responsible for patient care immediately after surgery.

Key considerations highlighted in recent guidelines include avoiding hypothermic perfusion (>37°C) and using a goal-directed oxygen delivery strategy that includes appropriate hemodynamic monitoring and fluid management.

Rewarming the patient on CPB may result in differential tissue perfusion and has been proposed as a mechanism for renal dysfunction after cardiac surgery. Thus, avoiding hyperthermia during this phase of the operation is recommended. , Similarly, inadequate D o ₂ has been associated with the development of CSA-AKI. A seminal single-center, prospective observational study of 1048 patients on CPB found that oxygen delivery lower than D o ₂ that was indexed to body surface (D o _2i ) of <272 mL/min/m ² was independently associated with CSA-AKI that required renal replacement therapy. This finding was further refined in subsequent analyses, indicating that a D o _2i of ≥270 mL/min/m ² is required to avoid stage 1 AKI. , , The intraoperative team should seek to adjust arterial flow and hematocrit levels to maintain D o _2i above this level. , , At present, there is no evidence that a dopamine infusion during CPB or the use of mannitol provides any protection against CSA-AKI.

During the first 24 to 48 hours after a cardiac procedure, the kidney remains vulnerable to ongoing or new insults. Hemodynamic perturbations, bleeding, and administration of drugs or agents with nephrotoxic side effects can compound renal dysfunction after CPB. The ICU team should review a patient’s preoperative risk score during sign-out with OR anesthesia staff and then calculate a new risk score based on a dynamic predictive scoring tool at the time of ICU admittance. The postoperative risk score should be documented in the initial ICU history and physical and then incorporated into the genitourinary section of the patient’s daily assessment and plan in the progress note. As with the intraoperative phase, ongoing care should include optimizing glycemic control, using an insulin infusion to maintain blood glucose at 120 to 180 mg/dL (see endocrine section later), limiting aminoglycoside antibiotics, and avoiding NSAIDs and IV radiocontrast agents. Consideration should also be given to withholding ACE inhibitors and ARBs in patients who are initially oliguric or classified as having moderate or high risk for CSA-AKI. The use of a urine catheter for hourly monitoring of urine output should be maintained, with the goal of the patient producing a minimum of 0.5 to 1.0 mL/h/kg.

The choice of fluid, in terms of type (balanced versus saline crystalloid or crystalloid versus albumin), remains varied and controversial. , , Regardless, excessive fluid administration is associated with venous congestion and increased rates of AKI. Fluid administration and other agents should be goal-directed to ensure adequate kidney blood flow. The perioperative team should seek to optimize intravascular blood volume by using continuous hemodynamic hourly monitoring (either invasive or noninvasive) with administration of fluid challenges with lactated Ringer’s solution for responsiveness if the patient is oliguric; when the CI is <2.0 L/min/m ² and the CVP is <8 to 10 mmHg; if the PADP is <14 mmHg (if a pulmonary artery catheter is being used); or if stroke volume variation in a ventilated patient with normal sinus rhythm is >13% (if using an arterial pressure waveform monitor). Diuretics should be considered if CVP is >15 mmHg or PADP is >20 mmHg.

All patients with oliguria or deemed at high risk for CSA-AKI should be treated with the postoperative hemodynamic algorithm known as the Kidney Disease Improving Global Outcomes (KDIGO) Bundle. This practice was validated by the PrevAKI RCT, in which 1046 high-risk patients were identified through two new renal biomarkers: insulin-like growth factor-binding protein 7 (IGFBP7) and tissue inhibitor of metalloproteinases-2 (TIMP-2). These biomarkers are involved in G1 cell cycle arrest and are able to identify patients at high risk for CSA-AKI through a clinically available assay. An assay result of [TIMP-2]·[IGFBP7] >0.3 has been associated with CSA-AKI.

In the PrevAKI study, patients with a positive biomarker result were randomized to an intervention protocol that included using fluids, diuretics, and inotropes to maintain a systolic blood pressure 100 to 130 mmHg and/or a MAP of 65 to 90 mmHg, a CI >2.2 L/min/m ² , urine output of at least 0.5 mL/kg/h (using lean body mass), and Sv o ₂ >65%. Within this study protocol, the investigative team used a hemodynamic catheter-based cardiac monitoring system that combined pulse-wave contour analysis, transpulmonary continuous thermodilution, and central venous oxygen saturation monitoring to optimize volume status. Further, ACE-I, ARB medications, NSAIDs, and radiocontrast agents were avoided in high-risk patients for at least 48 hours postoperatively. In addition, an intensive insulin infusion order set was implemented for the first 12 to 24 hours to allow for optimal blood glucose management and to avoid hyperglycemia, defined as >180 mg/dL for more than 3 hours.

Patients in the intervention arm of the PrevAKI trial had lower rates of AKI versus controls (55.1% vs. 71.7%, respectively; odds ratio 0.48, 95% CI 0.29–0.80) and had significant reductions in moderate to severe AKI as well (29.7% versus 44.9%). There were no differences between the groups in the need for renal replacement therapy in those who developed CSA-AKI despite the use of the KDIGO Bundle. Similarly, the 90-day mortality, ICU LOS, and major adverse kidney events endpoints did not differ statistically between treatment groups.

Lastly, in patients with RV dysfunction, venous hypertension can have deleterious effects on renal function due to a reduced transrenal pressure gradient. It is prudent for the perioperative team to provide appropriate inotropic support to the patient with RV dysfunction or RV failure and to consider renal replacement therapy for volume management (particularly in patients with a CVP >15 or a PADP >20).

A renal function panel should be monitored at least every 12 hours during the first 24 hours for all elective cardiac surgery cases, regardless of risk score. Serum creatinine and hourly urine output should be monitored until no further indicator of AKI persists. It is reasonable to consider reassessing AKI risk in all elective cardiac surgery patients by using a dynamic predictive scoring tool after the first 24 hours.

As outcomes related to CSA-AKI can be impactful, from both patient outcome and healthcare resource utilization viewpoints, it is suggested that the surgical and perioperative teams monitor AKI rates as a component of ongoing quality efforts.

Patient and surgical risk factors for postoperative stroke include history of stroke, peripheral vascular disease, diabetes, hypertension, dialysis dependency, severe chronic lung disease , previous cardiac surgery, preoperative infection, urgent operation, CPB lasting >2 hours, need for intraoperative hemofiltration, and high transfusion requirement. , ,

In a large retrospective review of more than 45,000 patients undergoing CABG over almost 3 decades, Tarakji and colleagues found that the risk for stroke peaks by the second postoperative day. Others have reported that the risk is reduced after that but can persist for as long as 5 years. Whitlock and colleagues classified postoperative strokes as early or late, based on etiology. Gaudino and colleagues classified stroke as intraoperative, early postoperative (first 7 days), or late (beyond 7 days).

Approximately 40% to 50% of perioperative strokes are discovered when the patient awakens from anesthesia and are therefore attributed to an intraoperative event. , Broadly, these can be subclassified as secondary thromboembolic events (60%, arising from aortic, cardiac, or CPB sources) or hypoperfusion events (40%). The most common etiology of early postoperative stroke events is new POAF (see “ Atrial Fibrillation ” earlier). , Late stroke can be attributed to the baseline atherosclerotic burden and the patient’s risk profile; persistent POAF after cardiac surgery remains a long-term risk. , , ,

Intraoperatively, several strategies may minimize the risk for cerebral hypoperfusion and embolic events: maintaining a higher mean arterial blood pressure, using cerebral oximetry monitoring and intervening with different techniques for low cerebral saturation, using TEE and epiaortic ultrasonography to identify and avoid manipulation of aortic atheromatous disease, using neuraxial anesthetic adjuncts, maintaining a hemoglobin concentration >60 g/L (or 6 g/dL) during CPB and at least 7 g/dL (or 70 g/L) postoperatively, and maintaining rationalized control of blood glucose (80–180 mg/dL).

Pharmacologically, aspirin initiated within 6 hours postoperatively continues to be the standard of care after CABG. This approach is associated with lower perioperative stroke rates if administered within 48 hours of revascularization. , Data from the LAAOS III trial indicate that intraoperative occlusion of the left atrial appendage in patients with preexisting AF and a CHA ₂ DS ₂ -VASc score of at least 2 was associated with significant short-term and long-term reduction in ischemic stroke and systemic embolism.

When the perioperative team identifies a new neurologic deficit indicative of a stroke after a cardiac procedure, a hospital-specific acute stroke protocol should be activated. , In addition to seeking neurologic consultation, the perioperative team should obtain brain and cerebrovascular imaging while maintaining adequate blood pressure and avoiding fever, hypoglycemia and hyperglycemia, and hypovolemia. Imaging typically consists of a noncontrast computed tomography (CT) of the head and CT angiography of the intracranial and extracranial (carotid and vertebral) vessels. Magnetic resonance imaging (MRI) of the brain and magnetic resonance angiography may be required. Of note, concerns persist about the presence of epicardial pacing wires, although investigations have shown the safety of performing an MRI in patients with retained materials after cardiac surgery. In this circumstance, the perioperative team should consult institutional policy before proceeding.

Treatment options for hyperacute stroke include thrombolysis and thrombectomy; however, only the latter is possible in the immediate postoperative patient due to the bleeding risk from IV alteplase. The timing of mechanical thrombectomy is an important consideration, with the current indication for patients with an acute ischemic stroke and large-vessel occlusion being within the first 24 hours, on the basis of recent trials , (but not yet updated in current guidelines). Thus, the perioperative team should examine and document last-known normal exam results to ensure that the patient has appropriate IV vascular access for the procedure. In addition, for patients with a pending mechanical thrombectomy , it is reasonable to maintain blood pressure at ≤185/110 mmHg before the procedure.

Patient-related factors being examined for their association with postoperative neurocognitive difficulties include silent cerebral infarction, cerebral autoregulation, alterations in blood-brain barrier function, and brain protein biomarkers such S100Beta, glial fibrillary acidic protein, tau, neurofilament light, and neuron-specific enolase. Rudolph and colleagues identified four risk factors associated with an increased risk for postoperative delirium after cardiac surgery: previous stroke or transient ischemic attack, low Mini Mental State Examination score, abnormal serum albumin, and a Geriatric Depression Scale score indicating depression. Other researchers have further identified elevated preoperative EuroSCORE, older age (≥70 years), number of comorbidities, history of delirium, alcohol use, type of surgery, blood transfusion, mechanical ventilation, postoperative low cardiac output, inadequate management or overtreatment of pain, sleep deprivation, use of benzodiazepines, and infection as additional risk factors. , ,

Postoperative delirium can be difficult to manage. Thus, prevention and management should begin in the preoperative phase of care by identifying key risk factors for its occurrence in the postoperative phase. Risk should be documented in the patient’s chart, and delirium risk should be communicated to the intraoperative team during the OR huddle and at handover to the ICU team postoperatively. ,

In the postoperative phase, a systematic delirium screening tool that assesses the full spectrum of delirium presentation (hypoactive, hyperactive, or mixed) should be incorporated as part of routine monitoring in the ICU at least once per nursing shift. When an active screening program in the ICU is lacking, delirium goes undiagnosed in more than 70% of cases. , A systematic methodology for delirium screening in the cardiothoracic ICU will use either a confusion assessment method , or an intensive care delirium screening checklist. Delirium screening should be undertaken in tandem with arousal assessment. The Riker Sedation Agitation Scale and the Richmond Agitation-Sedation Scale are some of the more widely used tools for determining levels of sedation and agitation.

Once delirium is detected, it is important to determine the underlying cause (e.g., pain, hypoxemia, low cardiac output, sepsis) and initiate appropriate treatment. , , , Nonpharmacologic strategies are the first-line component of management and should include the multicomponent Society of Critical Care Medicine ICU “ABCDEF Liberation Bundle” in the postoperative cardiac surgery ICU (see: https://www.sccm.org/ICULiberation/Home ). , A multicenter randomized trial examining a multicomponent family-support intervention found that the intervention was associated with higher ratings of the quality of communication and a shorter ICU LOS. Although these data were not specific to the cardiac surgery patient, it is reasonable for the perioperative team to consider incorporating this approach within their delirium management protocols.

Pharmacologic strategies using prophylactic medicines such as clonidine or antipsychotics have been found to reduce delirium rates, as shown in several RCTs. Current recommendations indicate that the use of antipsychotic medications should be limited to patients who have particularly distressing symptoms or who pose safety risks to themselves or the healthcare team. An exception may be dexmedetomidine, a selective α2-adrenergic agonist that may facilitate lighter sedation, shorten the duration of mechanical ventilation, reduce the risk for delirium, and provide analgesia in critically ill patient populations. , However, dexmedetomidine is also associated with bradycardia and lowered blood pressure. In a recent rapid practice guideline, the writing committee stated that the use of dexmedetomidine could be considered “…if the desirable effects including a reduction in delirium are valued over the undesirable effects, including an increase in hypotension and bradycardia….”

Importantly, McPherson and colleagues reported that the use of chemical restraints (such as benzodiazepines) or physical restraints and restraining devices predisposed patients to a greater risk for delirium, pointing to potential areas for quality improvement in the cardiac surgical ICU.

The absence of daily access to an empathetic individual who is not part of the surgical team, both before and after surgery, is an important risk factor for increased anxiety and depression postoperatively. Postoperative depression after cardiac surgery is an independent risk factor for subsequent cardiac events and is associated with increased risk for hospital readmission and death. , , Postoperative depression is probably influenced by a complex interaction between genetic and psychosocial predisposing factors, by neuroendocrine activation, and by the release of proinflammatory factors.

The most important methods for preventing mood dysfunction include daily access to family or caregivers and physician attention to the patient as a sensitive and threatened human being. When the mood state becomes severely abnormal, psychiatric assistance is required. Lastly, appropriate communication to the patient’s primary caregiver at the time of hospital discharge should include follow-up of both cardiac rehabilitation and mental health needs.

Visual-field defects are sometimes perceived by the patient and confirmed by testing; these usually regress or disappear spontaneously. Difficulty focusing, such as is required for reading, is sometimes experienced in the early postoperative period (up to 30 days). These issues can be classified as mild-delay neurocognitive recovery and often resolve without specific treatment.

Postoperative seizures are relatively uncommon after cardiac surgery but can have important consequences for the patient. , Incidence rates of clinically evident seizures range from 0.5% to 8.0%. Postoperative seizures usually appear within the first 2 days after surgery and are typically short (2–3 minutes), but they can deteriorate to status epilepticus. When seizures do occur, recurrence rates of 40% to 66% have been reported.

Goldstone and colleagues, in an analysis of approximately 2600 patients undergoing cardiac surgery, determined that 71% of observed seizures were of the generalized tonic-clonic type, 26% were simple/complex partial seizures, and 3% were status epilepticus. The lowest seizure rates were for isolated CABG, and higher rates were observed for combined valve/CABG and aortic procedures. Postoperative seizure was associated with a longer ICU and hospital LOS and nearly fivefold higher operative mortality (29% vs. 6% in patients with no seizure). Reported causes include ischemic strokes (the most common etiology in this cohort), DHCA, aortic calcification or atheroma, older age (>75 years old), previous strokes, critical preoperative state, cerebral air embolism, medication toxicity or withdrawal, substance (alcohol) withdrawal, and other operative drugs such as larger doses of tranexamic acid.

Immediate management should focus on the “ABCs” of acute resuscitation: airway, breathing (with an increase in Fi o ₂ ), and circulation. If the seizure has not self-terminated, IV benzodiazepine is the appropriate first-line agent. If this is unsuccessful, one of several antiseizure medications (e.g., valproic acid, levetiracetam, phenytoin) can be given. If seizures do not terminate with one antiseizure medication load, the perioperative team should next infuse propofol or midazolam and plan to secure the airway with endotracheal intubation.

Once the patient is stabilized, appropriate neurologic imaging (CT) should be undertaken to rule out acute ischemic stroke or hemorrhage. An electroencephalogram may be required to rule out nonclonic seizures in patients with remaining somnolence or delirium. The patient’s medical history should be queried for a previous history of seizures or other conditions that may predispose the patient to seizure activity (e.g., substance withdrawal). Decisions about the need for longer-term antiepileptics should be made in conjunction with a neurology consultation.