The Systemic Inflammatory Response to Cardiac Surgery: Implications for the Anesthesiologist

Inflammation is the body's response to tissue injury and is a rapid, highly amplified, controlled humoral and cellular response. 1While the term “sepsis” has classically been utilized to imply a clinical response to infection, a similar response may arise in the absence of infection. 2In fact, patients who appear to have sepsis but have negative microbial cultures have similar morbidity and mortality rates to the respective culture-positive populations. 3This has led to the understanding that this process is a generalized, nonspecific inflammatory response to injury and prompted a diagnostic reclassification of these events into a pathophysiologic continuum by the American College of Chest Physicians–Society of Critical Care Medicine Consensus Conference Committee in 1991 (table 1). 2 

The term “systemic inflammatory response syndrome” (SIRS) has been proposed to describe the entry point to this continuum, an entity that overlaps with normal postoperative physiology. 2SIRS is a nonspecific, generalized inflammatory process, independent of the causative factors, and is of importance for several reasons. It is a sensitive if nonspecific indicator of injury. The classification of severity of SIRS into uncomplicated SIRS, sepsis, severe sepsis, and septic shock based on the existence of documented infection or hypotension has prognostic significance. 3A frequent complication of SIRS is the development of organ dysfunction, including acute lung injury, shock, renal failure, and multiple organ dysfunction syndrome (MODS). Finally, long-term survival in patients developing SIRS may also be adversely affected. This is well documented in the context of sepsis, with the risk of death increased for up to 5 yr after the septic episode. 4 

Cardiac surgery provokes a vigorous inflammatory response, which has important clinical implications. In the report from the Society of Thoracic Surgeons National Database, 20% (22,000 patients) of “low-risk” patients developed postoperative complications. 5The incidence of MODS following cardiopulmonary bypass (CPB) was 11%, with a mortality rate of 41% in these patients in another study. 6Acquired multiple organ dysfunction is the best predictor of mortality in cardiac surgical patients who require prolonged postoperative mechanical ventilation. 6Many aspects of a patient's risk of serious perioperative complications are perceived as being relatively fixed (genotype, preoperative health status, surgical difficulty, etc. ), but the degree to which these may be improved (e.g. , hemodynamic optimization using pharmacologic or mechanical support) is still under assessment. The contribution of the inflammatory response to patient outcome is potentially remediable and therefore deserves attention.

Factors influencing incidence, severity, and clinical outcome of the inflammatory response, and in particular the reasons why certain patients develop life-threatening perioperative complications, are currently not well understood. Three separate perspectives contribute to our understanding of the link between the inflammatory response and adverse clinical sequelae. First, the complex interaction of humoral proinflammatory and antiinflammatory molecules may influence the clinical presentation and course of SIRS, with the balance of proinflammatory and antiinflammatory cytokines determining the clinical course following cardiac surgery. 7Alternatively, changes in the time course, magnitude, or patterns of cytokine release following CPB may contribute to abnormalities in the inflammatory response to cardiac surgery.

Second, a “multiple-hit” scenario may be seen, whereby serious sequelae develop after cardiac surgery as a result of adverse events, such as infection or ongoing organ hypoperfusion. 1,8The combination results in the conversion of an inherently self-limiting, tightly controlled homeostatic response to an uncontrolled destructive process resulting in organ dysfunction. 1,8,9Potential mechanistic insights into the pathophysiologic basis for multiple hits include the ability of CPB to “prime” neutrophils, causing pulmonary leukosequestration, 10and enhanced cytotoxin release following a subsequent insult. 10 

Third, it has been suggested that there may be a fundamental misconception about the inflammatory response. The proinflammatory state, SIRS, may be only one aspect of a multifaceted response. The converse has been termed the compensatory antiinflammatory response syndrome. 11CPB-induced generalized immunosuppression may play an important role in the development of infectious complications. 12Cumulatively, these responses represent the body's attempt to reestablish homeostasis and may clinically manifest as predominantly proinflammatory (SIRS), antiinflammatory (compensatory antiinflammatory response syndrome), or an intermediate mixed response. 11 

As perioperative physicians, anesthesiologists contribute to the management of the patient when adverse sequelae of CPB may pose a significant threat. Anesthesiologists are well positioned to minimize the risk of adverse sequelae resulting from the inflammatory response to CPB by reducing perioperative risk factors. Many drugs administered during the perioperative period, particularly for the purposes of anesthesia and sedation, possess potentially important immunomodulatory effects. Anesthesiologists may be best placed to properly evaluate and eventually implement therapeutic strategies, particularly as many potentially useful therapies seem poised to enter the clinical arena. Finally, a thorough knowledge of all aspects of the inflammatory response to CPB is mandatory if the anesthesiologist is to realize the goal of minimizing perioperative risk.

Nonspecific activators of the inflammatory response include surgical trauma, blood loss or transfusion, and hypothermia. In addition, CPB may specifically activate the inflammatory response via  at least three distinct mechanisms (fig. 1). One mechanism involves direct “contact activation” of the immune system following exposure of blood to the foreign surfaces of the CPB circuit. A second mechanism involves ischemia–reperfusion injury to the brain, 13,14heart, 14,15lungs, 16kidney 17,18and liver 19as a result of aortic cross-clamping. Restoration of perfusion on release of the aortic cross-clamp is associated with activation of key indices of the inflammatory response. 20,21 

Endotoxemia may indirectly activate the inflammatory cascade. Splanchnic hypoperfusion, a common finding during and following CPB, 22may damage the mucosal barrier, allowing gut translocation of endotoxin. 23,24Systemic endotoxin concentrations correlate closely with the degree of cardiovascular dysfunction following CPB, 25,26while low preoperative serum immunoglobulin M antiendotoxin core antibody concentrations predict poor outcome. 27However, the importance of endotoxin in stimulating the inflammatory response to cardiac surgery remains in doubt, with conflicting evidence regarding the importance of gut translocation. 26,28In fact, endotoxin may be a contaminant of fluids, such as cardioplegia and circuit priming fluid, routinely used during CPB. 29The sole study to examine the incidence and time sequence of splanchnic hypoperfusion (as measured by intramucosal pH), gut permeability, and endotoxemia during CPB found no association between mucosal acidosis and either endotoxemia or intestinal permeability. 30 

Complement is activated during extracorporeal circulation, reperfusion of ischemic tissues, and heparin neutralization with protamine. 31Exposure of blood to the foreign surface of the extracorporeal circuit results in direct “contact” activation of the complement cascade, predominantly via  the alternate pathway (fig. 2). 20,32,33The blood–gas interface of the CPB circuit may also play a role in complement activation. 34The formation of heparin–protamine complexes activates the complement cascade mainly via  the classic (C4a) pathway. 32,35In the first 5 days following cardiac surgery, a second delayed increase in complement activation products is seen. 36This delayed activation of complement appears to be mediated by C reactive protein in response to heparin–protamine complexes. 36 

The central role of complement in the inflammatory response to cardiac surgery has been demonstrated by the effects of complement-specific inhibitors. Soluble human complement receptor type 1 attenuates lung and myocardial injury, 37while blockade of C3a formation prevents activation of neutrophils, monocytes and platelets in models of CPB. 38Anti-C5a monoclonal antibodies attenuate CPB-mediated pulmonary, 39myocardial, 40mesenteric, 41and microvascular 39,40dysfunction. Specific blockade of the alternative pathway of complement activation by monoclonal antibodies to properdin causes near complete inhibition of C3a and C5b-9 formation and dramatically reduces platelet and neutrophil activation. 42Recombinant soluble complement receptor 1, 43C3-binding cyclic synthetic peptide, 44and antihuman C5 monoclonal antibody 45all prevent up-regulation of adhesion molecules necessary for neutrophils to bind to the CPB circuit and vascular endothelium, a necessary step in the injury process. Finally, increased plasma concentrations of C5b-9, a terminal complex of complement proteins C5 to C9, are seen during CPB (fig. 2). Selective blockade of C5b-9 formation by antihuman C8 monoclonal antibody inhibits platelet but not leukocyte activation in a model of simulated extracorporeal circulation. 46 

The degree of complement activation in patients undergoing CPB has clinical significance. The degree of postoperative pulmonary shunt correlates with activation of the classic pathway by protamine–heparin complexes. 47Postoperative levels of C4d–C-reactive protein, a specific marker for C-reactive protein–mediated activation of complement, correlate with the incidence of postoperative arrhythmia following coronary artery bypass graft (CABG). 35,36Postoperative C3a concentrations may predict the probability of cardiac, pulmonary, renal, and hemostatic dysfunction 48and the likelihood of developing MODS in children. 33Strategies that improve CPB circuit biocompatibility reduce indices of complement activation and may decrease postoperative morbidity, particularly in high-risk patients. 49Anti-C5a antibody, which reduces sC5b-9 formation, significantly reduces myocardial injury, blood loss, and postoperative cognitive deficits in patients undergoing CPB. 50 

Cytokines are soluble proteins and polypeptides that act as paracrine messengers of the immune system and are produced by a large variety of cell types, including activated monocytes, tissue macrophages, lymphocytes, and endothelial cells (fig. 3). Individual cytokines may exert either proinflammatory or antiinflammatory effects. Cytokines are essential for immunologic and physiologic homeostasis, are normally subject to tight homeostatic control, and are produced in response to a variety of physiologic and pathologic stimuli.

Proinflammatory cytokines play a pivotal role in stimulating the inflammatory process, with plasma concentrations of specific cytokines, such as interleukin-1β (IL-1β) and interleukin-6 (IL-6), predictive of outcome in certain critically ill patient subgroups (table 2). 51Tumor necrosis factor α (TNF-α) and IL-1β are elevated early following cardiac surgery, with IL-6 and IL-8 peaking later. 52,53While a direct cause-and-effect relation has not been demonstrated, elevations of proinflammatory cytokines have been strongly associated with adverse outcome following cardiac surgery. Patients who develop SIRS demonstrate significant elevations in cytokine concentrations compared to patients with an uncomplicated course following cardiac surgery. 54Within the subgroup of cardiac surgical patients who develop SIRS, nonsurvivors had dramatically higher IL-8 and IL-18 concentrations compared to survivors. 54In addition, serum concentrations of IL-6 correlate with mortality following pediatric cardiac surgery. 55 

The proinflammatory cytokine response to cardiac surgery is balanced by a phased antiinflammatory cytokine response, with soluble cytokine receptors, cytokine receptor antagonists, and antiinflammatory cytokines also produced in large quantities (table 3). 52Key antiinflammatory cytokines include interleukin-10 (IL-10), interleukin-1 receptor antagonist (IL-1ra), TNF soluble receptors 1 and 2 (TNFsr 1 and 2), and transforming growth factor β. IL-10 is a potent inhibitor of the production of TNF-α, IL-1β, IL-6, and IL-8. 56While the specific role of other antiinflammatory mediators remains undefined, it has been suggested the clinical prognosis following CPB may depend on the balance between proinflammatory and antiinflammatory cytokines. 7 

Nitric oxide (NO) is a ubiquitous biologic mediator that acts as a physiologic regulator and can be responsible for tissue damage. Physiologic regulatory functions include endothelial-mediated vasodilation in both the systemic and pulmonary circulations, potentially significant immunomodulatory functions, as well as protean roles in nociception, memory, and erectile function. 57NO may have several protective roles in the inflammatory response. NO-induced vasodilation may prevent accumulation of injurious mediators at the endothelium (fig. 3). NO may scavenge free radicals and prevent up-regulation of neutrophil CD11/CD18 adhesion molecules. 58Supplementation of cardioplegia and perfusate with an NO precursor (arginine) or NO donor (SPM-5185) has beneficial effects on postreperfusion neutrophil accumulation, endothelial function, and myocardial performance following experimental myocardial ischemia, 59possibly by inhibiting neutrophil adherence and cytotoxicity. The release of NO during CPB appears to be dependent on the type of bypass flow, with attenuation of basal NO release during nonpulsatile flow leading to end-organ functional capillary closure as a result of diminished vessel wall shear stress. 60,61This release of NO is considered to be physiologic, produced by endothelial constitutive NO synthase (cNOS), 60,62and its release appears to be a function of both the frequency and amplitude of the pulsatile flow. 63 

The role of NO in the inflammatory response is complex, however, and NO has several potentially deleterious actions. Cytokine-induced decrements of myocardial function appear to be related to increases in myocardial inducible NO synthase (iNOS), 64which is up-regulated by CPB. 65Prevention of myocardial iNOS up-regulation may reduce hemodynamic instability post-CPB, 65while NOS inhibition can reverse refractory hypotension in established shock. NO is a highly reactive free radical and combines with a variety of molecules in vivo . While the free radical scavenging role of NO is generally protective, 66NO may combine with the free radical superoxide to form peroxynitrite, a more reactive and injurious free radical. 66,67Finally, NO appears to be a powerful direct cellular toxin, inactivating enzymes involved in glycolysis, the Krebs cycle, and the electron transport chain 67and reducing intracellular adenosine triphosphate and antioxidant concentrations, hence predisposing to cell death.

The timing, source, and quantities of NO produced may be the key to dissecting these apparently paradoxical roles. NO is produced constitutively in small (picomolar) quantities by cNOS isoforms, such as the vascular endothelial isoform (ecNOS). ecNOS plays a pivotal role in the physiologic regulation of basal vascular tone, blood flow capillary integrity, and leukocyte and platelet adhesiveness to the endothelium. 57The activity of ecNOS appears to be inhibited in the earliest phases of the inflammatory response, allowing both unopposed vasoconstrictive influences and increased leukocyte and platelet adhesion to the endothelium. However, within 4–8 h, iNOS is produced in a wide variety of tissues, including vascular smooth muscle, hepatocytes, and Kupffer cells, 67,68and produces NO at much higher (nanomolar) quantities. Cytokines, particularly IL-1β, play a pivotal role in the process of NO-induced inflammatory dilatation. 69The proinflammatory response, once fully developed, represents a high-output NO state. 67In fact, exhaled NO following initiation of CPB may represent an index of severity of the inflammatory response, 70although this has been disputed. 71 

The coagulation–fibrinolytic cascades and the inflammatory response, while in many respects separate processes, are closely interconnected, with activation of coagulation a key component of the acute inflammatory response and vice versa . In this context, inflammation and coagulation should perhaps be considered different facets of the same host response to injury.

The coagulation system has traditionally been divided (for conceptual and practical purposes) into the intrinsic and extrinsic pathways, both of which lead to a final common pathway, and result, via  thrombin generation, in the formation of an insoluble fibrin clot. Activation of coagulation after CPB had been thought to be predominantly due to activation of the intrinsic pathway via  the contact system. In this paradigm, plasma factor XII was pivotal to this process, becoming adsorbed and activated on contact with the CPB circuit. However, patients with congenital deficiency of factor XII still generate thrombin following CPB. 72This suggests that the extrinsic pathway, which requires expression and activation of tissue factor, perhaps in response to inflammatory stimuli and oxidative or shear stress, may also be involved. 73 

Hemostasis is mediated by a balance of procoagulant and anticoagulant forces, which normally coexist in a delicate balance. The coagulation cascade consists of inactive circulating precursors, which are sequentially activated via  enzymatic cleavage, yielding an active serine protease that hydrolyses the next factor in the cascade. Thrombin, the end product, catalyzes the formation of insoluble fibrin from fibrinogen, which form the strands that bind the platelet plug. This process is normally controlled and limited to discrete sites of injury by modulators, including plasminogen activators, thrombomodulin, proteins C and S, and serine protease inhibitors such as antithrombin III. 74The fibrinolytic cascade, activated during coagulation, results in the formation of plasmin, which splits fibrinogen and fibrin, remodelling the formed clot and later removing the thrombus when the endothelium heals.

Several specific points regarding the complex interrelation between coagulation and inflammation, in the context of CPB, deserve attention. The endothelium is intricately involved in both processes (see Endothelium). Proinflammatory cytokines play a key role in initiating the coagulation process locally at sites of inflammation, by activation of the endothelium, induction of the expression of tissue factor, eliciting the expression of leukocyte adhesion molecules on the intravascular cell surfaces, and stimulating production of platelet-activating factors. 75,76This, combined with down-regulation of thrombomodulin expression and of the fibrinolytic and protein C anticoagulant pathways, alters the balance between procoagulant and anticoagulant activities, resulting in a markedly procoagulant state. 75,76 

The coagulation system in turn impacts on the inflammatory response. Platelet activation at sites of tissue injury results in the release of multiple mediators that alter tissue integrity. Several key coagulation proteins, such as thrombin and factor Xa, have proinflammatory properties. Thrombin, formed following activation of the coagulation cascade, stimulates several cell chemotaxins and mitogens, which are responsible for the spreading of the lesion and the tissue repair process. 75,76 

Heparin and protamine, which are used to modulate coagulation in almost all patients undergoing cardiac surgery, may have important immunomodulatory effects. 32,77Heparin appears to possess important antiinflammatory effects. 77Protamine neutralization of heparin may result in multiple cardiovascular effects, including increased pulmonary artery pressures and decreased systolic and diastolic blood pressure, myocardial oxygen consumption, cardiac output, heart rate, and systemic vascular resistance. 78Although protamine by itself has adverse effects, the heparin–protamine complex is particularly deleterious. The heparin–protamine interaction activates the inflammatory response by several mechanisms, including complement activation, histamine release, thromboxane and nitric oxide production, and antibody formation. 78The release of thromboxane may result in severe pulmonary hypertension. 78In a minority of patients, severe anaphylactoid reactions may result from the heparin– protamine interaction.

The balance of procoagulants and anticoagulants is profoundly disturbed in CPB patients. Activation of procoagulants such as thrombin mandates administration of anticoagulant drugs prior to CPB to prevent blood from clotting instantly on contact with the extracorporeal circuit. In addition, the stimulation of fibrinolysis during CPB appears to contribute to the postoperative coagulopathy commonly seen in these patients. 79Widespread vascular injury following CPB may result in uncontrolled platelet activation, thrombin generation, and disseminated intravascular coagulation. 80The resulting widespread fibrin deposition in the microvasculature may occlude microcirculatory flow and cause end-organ damage, which may progress to MODS and death. 80 

The vascular endothelium is a dynamic participant in cellular and organ function rather than a static barrier, as was once believed. It is intimately involved in a variety of physiologic and pathologic processes and has emerged as the central focus of many of the biologic events that affect the perioperative course of the cardiac surgical patient. The endothelium controls vascular tone and permeability, regulates coagulation and thrombosis, and directs the passage of leukocytes into areas of inflammation, through the expression of surface proteins and secretion of soluble mediators.

The inflammatory response to CPB is characterized by a state of widespread endothelial activation and diffuse endothelial dysfunction. 70Inflammatory mediators, including TNF-α and IL-1β, bind to specific receptors on the endothelium, initiating diverse signal transduction pathways, which in turn activate a specific set of genes within the nucleus of the endothelial cell, termed activation genes. The transcription factor NF-κB plays a pivotal role in the signal transduction process. When activated, it dissociates from the cytosolic inhibitory protein IκB, translocates to the endothelial cell nucleus, binds with specific DNA sequences, and alters the conformation of the basal transcriptional apparatus, resulting in the transcription of the activation genes. 81This process results in the translation of proteins, including adhesion molecules (e.g. , E-selectin, intercellular adhesion molecule-1) and cytokines (e.g. , IL-8), required for endothelial cell activation, a process that takes approximately 4 h and peaks at 8–24 h depending on the gene. 73 

The activated endothelial cell plays a pivotal role in linking the inflammatory and coagulation systems, by expressing proteins central to the activation of coagulation and inflammation. 73,76Endothelial cell adhesion molecule expression mediates the interaction between the neutrophil and the endothelial cell (see The Cellular Immune Response), resulting in neutrophil adhesion, activation, and degranulation. This further damages the endothelium, causing diffuse capillary leak and edema formation. 73,82Endothelial injury results in the expression of tissue factor, augmented by IL-1β and TNF-α, which activates the extrinsic pathway of coagulation 76and may result in disseminated intravascular coagulation. 75In addition, protein C, a key inhibitory regulator of hemostasis, is antagonized in inflammatory states, most probably by TNF-α, further shifting the balance toward a procoagulant state. 76 

Vascular endothelium plays a central role in the pathogenesis of microcirculatory derangement following CPB. Endothelial regulation of local vascular tone (fig. 3) is mediated via  a variety of endothelium-derived relaxing and contracting factors such as NO, prostacyclin, endothelium-derived hyperpolarizing factor, endothelin, and thromboxane A2. 83The increase in pulmonary vascular resistance following CPB is attributed to reduced NO release from dysfunctional pulmonary endothelium 70and is reversed by NO supplementation. 84 

A complex interplay exists between impaired endothelial function, inflammation, and atherosclerosis in the pathogenesis of adverse cardiovascular events. 85,86Alterations in NO generation appear to underpin these interrelationships. 86Of particular concern, the inflammatory response to cardiac surgery may increase the risk of a postoperative cardiac event. Proinflammatory cytokines and endotoxin can impair endothelium-dependent dilatation, and the endothelium may lose its ability to respond to circulating hormones or autacoids. 69,86Studies of forearm vasoregulation demonstrate that IL-1β, TNF-α, and endotoxin 87,88cause prolonged but reversible impairment of endothelial relaxation, termed “endothelial stunning.”88,89Proinflammatory cytokines inhibit production of NO and a vasodilator antiplatelet prostanoid. 87Loss of the vasodilator and antithrombotic effects of NO may alter myocardial perfusion 88and expose preexisting atheroma to unopposed vasospastic and prothrombotic influences. 86This may explain the association between an acute inflammatory episode and a transient increase in the risk of a cardiovascular event. 89 

In addition, endothelial dysfunction may limit the long-term success of cardiac surgery, particularly CABG, by contributing to the development of narrowing at graft anastomotic sites as a result of medial hyperplasia 85and by accelerating the progression of atherosclerosis. Abnormalities of endothelium-dependent vasodilation, including paradoxical vasoconstrictor responses to, e.g. , exercise, are often observed in the earliest stages of coronary artery disease. 85Long-term follow-up clinical studies demonstrate that this is associated with an increased rate of cardiovascular events. 90Endothelial dysfunction activates the inflammatory response, recruiting leukocytes and platelets to the arterial wall, which may initiate the formation of atherosclerotic plaque. 85,89This process is particularly likely at sites of disturbed blood flow such as occur at graft anastomoses 91and is markedly potentiated by the presence of hypercholesterolemia. Endothelial dysfunction in hypercholesterolemic patients is in large part due to a reduced bioavailability of NO. 92In this regard, statins, which lower cholesterol, have been demonstrated to rapidly restore endothelial function, in part by directly up-regulating eNOS. 85This restoration of endothelial function results in improved myocardial perfusion, reduced myocardial ischemia, and reversal of atherosclerosis. 85 

The process of neutrophil–endothelial adhesion is an essential component of the inflammatory response leading to widespread endothelial damage and is now well understood, involving distinct phases of primary and secondary adhesion (fig. 4). 82In the noninflamed state the leukocyte travels at around 1,000 μm/s, along with the erythrocytes, in the center of the postcapillary venule. In the first phase, primary adhesion, the freely moving neutrophil is converted to the “rolling” state, in which it tumbles slowly (around 30 μm/s) along the endothelium. 82This is mediated by the expression of a family of adhesion molecules known as selectins. P- and E-selectin are expressed on the endothelium, and L-selectin is expressed on the neutrophils. These are involved in the formation of loose bonds between the endothelium and the neutrophil, which slows down the passage of the leukocyte along the blood vessel wall. C5a, released on activation of the complement cascade following contact with the extracorporeal circulation, is a potent stimulant of P-selectin expression. 93P-selectin is stored preformed in cytoplasmic vacuoles (the Weibel Palade Bodies) and rapidly reaches the plasma membrane by exocytosis after endothelial cell activation. 94This may underlie the sudden leukosequestration in the pulmonary circulation, which occurs following initiation of CPB. 93This process of primary adhesion is later maintained by E-selectin, which must be produced de novo  by the activated endothelium. 73 

Secondary or tight adhesion of the neutrophil to the endothelium requires the action of the integrin family of cell surface receptor molecules. These macromolecules consist of two different protein chains (α and β) that are noncovalently associated and are expressed on the cell surface. One subfamily of integrins, which share the same β chain (CD18), is expressed only on leukocytes. CD11a/CD18 and CD11b/CD18 are expressed abundantly by neutrophils, the former at relatively constant concentrations, while expression of the latter is greatly enhanced by cytokines, such as IL-8 and C5a. 73,82,95The activated integrins bind to adhesive molecules expressed on the endothelial surface, in particular, intercellular adhesion molecule-1. CD11a/CD18 may also bind to a related adhesion molecule, intercellular adhesion molecule-2, and CD11b/CD18 may adhere to elements of the extracellular matrix, such as fibrinogen. 82The neutrophil becomes tightly adherent to the endothelium, flattens out, and eventually transmigrates out of the circulation. 82,96Adherence via  CD11b/CD18 appears to prime the neutrophil to degranulate and undergo the respiratory burst. 82The primed neutrophils release cytotoxic proteases, such as elastase and myeloperoxidase, and reactive oxygen species, 96,97resulting in damage to the vascular endothelium and surrounding tissues. This complex process of coordination of the function of adhesive molecules and regulation of the migration of neutrophils is modulated by chemokines, in particular, platelet activating factor, IL-8, and C5a. 82,98Therefore, in the absence of secondary insults, such as shock or infection, endothelial cells return to their resting state and lose their adhesive properties following return of cytokine and C5a concentrations to normal. This limits the inflammatory response in most patients undergoing cardiac surgery. 73 

The cellular immune system is central to the inflammatory response following cardiac surgery. Increased spontaneous activation of both granulocytes and macrophages is observed following CPB. 99Hyperstimulation of naïve monocytes and granulocytes occurs following exposure to plasma from CPB patients. 100CPB activates monocytes and macrophages via  increased monocyte chemoattractant factor production, 101and up-regulation of macrophage adhesion molecule expression and cytokine production. 102Increased concentrations of leukocyte adhesion molecules (selectins and integrins) have been demonstrated following CPB. 95During CPB, the neutrophils are primed for enhanced free radical generation, as evidenced by an enhanced oxidative burst response seen up to 24 h after CPB 97and demonstration of the release of intracellular lysosomal granular contents post-CPB. 96Elevated circulating neutrophil counts and increased neutrophil aggregability, superoxide generation, and elastase release occur up to 24 h after surgery, 103with impaired oxidative function at 48 h postoperatively. 104This suggests a biphasic abnormality with potential early tissue cytotoxicity followed later by neutrophil dysfunction. In addition, reticuloendothelial Kupffer cells demonstrate greatly increased phagosomes following CPB, 105which may cause a relative “overuse” type of impairment of phagocytic activity. 105,106 

The clinical importance of leukocytes in the inflammatory response to CPB is underscored by several findings. Pulmonary neutrophil sequestration has been demonstrated following CPB and is associated with evidence of severe histologic lung injury. 107,108Inhibition of neutrophil CD11/CD18 expression 109or function 110improves myocardial function following cardiac surgery. Blockade of neutrophil adhesion 111decreases pulmonary injury after CPB. Finally, strategies that deplete circulating leukocytes can attenuate organ injury and may improve patient outcome following CPB (see Therapeutic Strategies). 112 

Preexisting disorders may influence the inflammatory response to cardiac surgery. Disordered cytokine balance may be a pathophysiologic feature of ischemic heart disease. 113Patients with preoperative left ventricular dysfunction undergoing CPB appear to have a greater degree of proinflammatory cytokine release, which is associated with impaired hemodynamics and a higher incidence of perioperative complications. 114CPB appears to cause greater oxidative stress in patients with diabetes, and there are qualitative differences in the inflammatory response in diabetic patients. 115Poorly controlled diabetes impairs the inflammatory response. 116 

The perioperative course of proinflammatory and antiinflammatory cytokine release in adult patients undergoing CABG does not differ with age. 117 

Perioperative hemodynamic instability predicts morbidity and mortality following CPB, 118–120with low cardiac output syndrome the most common event leading to patient death. 121Strong clinical and experimental evidence links postoperative splanchnic hypoperfusion to the development of complications such as acute respiratory distress syndrome. 16Experimental splanchnic ischemia–reperfusion induces lung recruitment and sequestration of neutrophils, 122increased pulmonary iNOS expression, 123and free radical injury. 124Cytokines, particularly TNF-α, also appear central to this process. 125,126In addition, translocation of toxins and microbial flora may occur following ischemic compromise of the integrity of the mucosal barrier.

Gastric mucosal pH, as measured by gastric tonometry, provides an index of splanchnic perfusion. 127Gastric mucosal acidosis indicates inadequate oxygenation of splanchnic tissue and is common following weaning from CPB, even in the absence of gross hemodynamic instability. 128It is a highly sensitive predictor for complications, 129independent of clinical risk stratification, even during uneventful cardiac surgery with low expected mortality. 130Therapeutic strategies to maintain hemodynamic stability and minimize postoperative splanchnic ischemia may reduce postoperative complications.

Thoracic epidural anesthesia combined with general anesthesia for CABG decreased the perioperative stress response, as measured by plasma epinephrine, and may decrease postoperative myocardial and pulmonary injury. 131However, thoracic epidural anesthesia does not significantly alter the cytokine response to CPB. 132 

Apnea during CPB may lead to activation of lysosomal enzymes in the pulmonary circulation, which in turn are correlated with the degree of postoperative acute lung injury (ALI). 133This may be attenuated by a vital capacity maneuver performed before termination of the bypass 134; however, continuous positive airway pressure applied during CPB appears ineffective. 135 

Many of the drugs used to produce anesthesia and maintain postoperative sedation and analgesia possess immunomodulatory effects (table 4). The clinical implications of such effects, particularly in the context of CPB, remain unknown, with most data in this area to date confined to in vitro  experiments. Nevertheless, developments in this emerging field are worthy of consideration in the light of their future therapeutic potential.

Propofol may enhance the antiinflammatory response to surgery by several mechanisms. Propofol may preserve hepatosplanchnic blood flow during CPB, thereby aiding maintenance of the mucosal barrier. 136It alters the balance between proinflammatory and antiinflammatory cytokines, increasing production of the antiinflammatory cytokine IL-10 and IL-1ra, 137while decreasing neutrophil IL-8 secretion, 138and scavenges reactive oxygen species. 139Low concentrations of propofol reduce neutrophil uptake in the coronary circulation following myocardial ischemia and reperfusion. 140However, this effect is abolished at higher propofol concentrations; this may be due to the propofol solvent Intralipid (Kabi Pharmacia, Uppsala, Sweden). 140Propofol impairs several aspects of monocyte and neutrophil function, including the respiratory burst, 141,142polarization, 143chemotaxis, 144–146phagocytosis, 141and oxygen radical generation. 147Certain immunomodulatory effects of propofol, such as suppression of respiratory burst of neutrophils by propofol, may be caused by the solvent Intralipid, 148while other actions, such as its ability toscavenge free radicals, appears to be a property of propofol itself. 137 

Sodium thiopental impairs the neutrophil respiratory burst, 142polarization, chemotaxis, adherence, phagocytosis and killing, 144–146and coronary uptake of neutrophils following myocardial ischemia and reperfusion. 140At therapeutic concentrations, thiopental also inhibits the monocyte respiratory burst. 141In high concentrations thiopental affects Escherichia coli  clearance in vitro  and neutrophil and monocyte phagocytosis. 141The effect of thiopental on the respiratory burst of neutrophils appears less pronounced compared to propofol. 148 

Ketamine attenuates the increase of serum IL-6 concentrations during and following CPB 149and reduces coronary uptake of neutrophils following myocardial ischemia and reperfusion. 140Ketamine affects E. coli  clearance and neutrophil and monocyte phagocytosis in vitro , although only in high concentrations. 141Methohexitone has only minimal effects on the respiratory burst of neutrophils in vitro . 142,148 

Opioids have multiple effects on the immune system, mediated indirectly via  the central nervous system or through direct interactions with the cellular immune system. 150,151While the precise cellular mechanisms underlying the immunomodulatory effects of opioids are largely unknown, emerging evidence indicates that opioids share many properties with cytokines. 151,152Ultrahigh affinity novel δ, μ, and κ opioid receptors have been demonstrated on inflammatory cells. 151Granulocytes contain both opioid peptide selective δ 2 receptors (which stimulate chemotaxis) and opiate alkaloid-selective, opioid peptide-insensitive receptors (which inhibit cytokine-induced activation and chemotaxis). 151 

Morphine down-regulates the activity of lymphocytes, granulocytes, and macrophages, and suppresses the antibody response. 153Microinjection of morphine into the lateral ventricle of the rat induces pronounced, dose-dependent reductions in lymphocyte proliferation to T- and B-cell mitogens, natural killer cell cytotoxicity, and the production of IL-2 and interferon-γ. 150Morphine also increases the secretion of CRH, ACTH, and glucocorticoids, i.e. , substances with inhibitory effects on the immune system. Certain immunomodulatory actions of morphine, including NO release and inhibition of cell adhesion, appear to be mediated specifically via  the μ3 receptor. 154 

Fentanyl increases concentrations of IL-1ra in in vitro  monocyte cultures. 155In an isolated blood primed CPB circuit, fentanyl increased CD11b, augmented the reduction in lymphocyte HLA-DR expression, and attenuated the increase seen in monocyte HLA-DR expression. 52However, fentanyl, unlike morphine, appears to lack the ability to bind to the μ3 receptor, diminishing its ability to down-regulate the inflammatory response to CPB. 154 

Midazolam, the best studied benzodiazepine, has little influence on host defense mechanisms. Midazolam decreases neutrophil IL-8 secretion in response to lipopolysaccharide but does not alter IL-8 production. 138Midazolam reduces postischemic uptake of neutrophils in the coronary circulation following myocardial ischemia and reperfusion. 140Midazolam, at clinically relevant concentrations in vitro , does not attenuate neutrophil polarization 143and has minimal effects on the neutrophil respiratory burst, 148neutrophil phagocytosis, 148and clearance of E. coli . 148 

Sevoflurane, isoflurane, and enflurane decrease proinflammatory cytokine (IL-1β, TNF-α) release by human peripheral mononuclear cells in vitro . 156Isoflurane decreases alveolar macrophage phagocytosis and microbicidal function to a greater extent compared with propofol. 157Halothane, isoflurane, and enflurane attenuate free radical–mediated myocardial injury. 158,159Isoflurane and halothane (but not sevoflurane) appear to attenuate hydroxyl radical production in the ischemic rat heart. 160Sevoflurane and isoflurane and halothane reduce neutrophil 161and platelet 162,163uptake in the coronary circulation and preserve cardiac function following myocardial ischemia and reperfusion. 162–164This effect is mediated at least in part via  reduced neutrophil expression of the adhesion molecule CD11b. 164 

Clonidine appears to exert antiinflammatory actions in such diverse areas as acute pain models, 165in extrinsic asthma, 166and angiotensin-converting enzyme inhibitor–induced inflammation. 167It appears that the antiinflammatory action of clonidine is a property of α²adrenoceptor activation. 168Furthermore, α²adrenoceptor agonists may regulate cytokine production via  stimulation of α²receptors on macrophages to augment TNF-α release in response to endotoxin. 169While use of clonidine during CABG does not appear to influence the perioperative stress response, 131its immunomodulatory effects in the context of CPB remain to be characterized.

Proinflammatory cytokine concentrations in patients undergoing heart transplantation are greater than in CABG patients. 12Possible explanations include the fact that these patients have severe ventricular dysfunction, a condition known to increase cytokine concentrations after cardiac surgery, 114or to the longer ischemia time required to perform this procedure. 12The later course of cytokine concentrations after heart transplantation may be further influenced by immunosuppressive therapy. 12Patients undergoing valve surgery appear to have similar immunologic response profiles to CABG patients. 170In general, indices of inflammation appear to correlate with overall severity of illness rather than specific surgical procedure. 171 

The composition of the priming solution, cardioplegia, presence of pulsatile or nonpulsatile perfusion, type of oxygenator and pump, type of extracorporeal circuit, and temperature during CPB are all important factors influencing the inflammatory response. In an isolated circuit, CPB duration correlated with IL-8 concentrations and neutrophil adhesion molecule expression. 52The release of inflammatory mediators appears to be temperature dependent, with warm CPB associated with an increased inflammatory response compared to hypothermic CPB. 172,173However, there is evidence to suggest that hypothermic CPB may simply delay cytokine release and neutrophil activation, and that tepid CPB, in which the core temperature is simply allowed to drift to 32–34°C, may most effectively attenuate the inflammatory response. 173,174Warm blood cardioplegia reduces the inflammatory response as compared to cold crystalloid cardioplegia. 175Heparin-coated CPB circuits reduce complement and proinflammatory cytokine release. 176The type of CPB prime, i.e. , colloid (gelofusine) versus  crystalloid, does not appear to influence cytokine release. 177Membrane oxygenators may result in less activation of the inflammatory system and improve postoperative cardiac, respiratory, and renal function. 178However, a later study demonstrated no sustained advantage over bubble oxygenators in terms of inflammatory activation or postoperative respiratory function. 179In pediatric patients undergoing CPB, use of centrifugal pumps may result in less activation of the inflammatory response 180; however, another study demonstrated no advantage over conventional roller pumps. 181In adult CPB patients, centrifugal pumps may induce a greater inflammatory response than roller pumps. 182Pulsatile CPB flow is associated with less endotoxemia, 23lower cytokine and endothelin-1 concentrations, 23,183and increased NO production. 60,62 

During normal circulation within the vasculature, blood is exposed to various physical or mechanical forces. This is termed “shear stress” and may have a physiologic function, such as releasing NO from the endothelium. 60However, excessive shear stress may develop during CPB as a result of large pressure changes across the CPB circuit, causing damage to blood constituents and activating the inflammatory response. Shear stress appears to be especially pronounced along the arterial cannula. 184Shear stress decreases erythrocyte deformability and increases hemolysis. 185Leukocyte adhesiveness is increased, and mechanical disruption, with neutrophil degranulation and release of cytotoxic products, may be seen at high levels of shear stress. 186Excess shear stress also increases platelet activation 187and may contribute to endothelial injury. 188Strategies that minimize shear stress, by decreasing the pressure decrease across the CPB circuit, such as the use of hollow fiber membrane oxygenators, reduce leukocyte activation. 189 

An estimated 20% of allogeneic blood transfusions in the United States are associated with cardiac surgery. 190The immunomodulatory effects of allogeneic transfusion in cardiac surgical patients are increasingly recognized. Allogeneic blood transfusions appear to exacerbate the proinflammatory response to cardiac surgery. Intraoperative packed erythrocyte transfusions directly increase the concentrations of inflammatory mediators and indirectly stimulate the inflammatory response. 191This may explain, in part, the greater morbidity in patients who receive intraoperative allogeneic packed cells. 191In addition, transfusion-associated graft-versus-host disease is a recognized complication of fresh nonirradiated homologous whole blood. 192 

Allogeneic erythrocytes are usually given to augment systemic oxygen delivery. However, a controlled clinical trial failed to demonstrate a beneficial effect of erythrocyte transfusion on systemic oxygen uptake in critically ill patients with sepsis. In fact, patients receiving old transfused erythrocytes developed evidence of splanchnic ischemia. The investigators postulated that the poorly deformable transfused erythrocytes cause microcirculatory occlusion and tissue ischemia in some organs. 193These concerns have been further highlighted by a recent large multicenter study of critically ill intensive care unit (ICU) patients that demonstrated that a restrictive strategy of erythrocyte transfusion improved mortality in the less severely ill patients and overall was at least as effective as and possibly superior to a liberal transfusion strategy. 194 

Autotransfusion of shed blood from the thoracic cavities during and following CPB is a common clinical practice used to reduce the need for allogeneic blood transfusion. 195,196However, concerns exist regarding the efficacy and safety of autotransfusion. Extensive clotting and fibrinolysis have been demonstrated in blood from the thoracic cavities during CPB. 196Studies to date on the efficacy of autotransfusion in reducing allogeneic blood transfusion have yielded conflicting results. One recent randomized clinical trial demonstrated that autotransfusion reduced allogeneic transfusion requirements but led to increased postoperative bleeding. 197Conversely, in another randomized controlled trial, autotransfusion of blood suctioned from the thoracic cavity during the surgical procedure increased circulating concentrations of thrombin–antithrombin III complex, tissue-type plasminogen activator, fibrin degradation products, and free plasma hemoglobin. Autotransfusion impaired hemostasis, resulting in increased postoperative blood loss and similar if not increased consumption of blood products when compared to patients who were not autotransfused. 196A metaanalysis of randomized trials assessing the effectiveness of cell salvage to minimize perioperative allogeneic blood transfusion in cardiac surgical patients concluded that it was only of marginal benefit. 198Of further concern, the potential exists for autotransfused blood to stimulate the inflammatory response. Shed mediastinal blood contains a high proportion of activated leukocytes with a significantly increased production of TNF-α and IL-6 seen in in vitro  studies. 195However, in one study, while shed mediastinal blood did contain high concentrations of IL-6 and activated leukocytes, autotransfusion in the postoperative period did not cause measurable elevations in cytokines. 195 

Continuous renal replacement therapies such as hemofiltration appear to remove both mediators, including TNF-α and IL-1β, and their inhibitors, such as TNFsr1, TNFsr2, and IL-1ra, from the plasma of patients with SIRS. 199This offers the possibility that these therapies may alter the course of SIRS in critically ill patients. However, the use of continuous renal replacement therapy in the absence of conventional indications for dialytic support remains unproven. No significant survival advantage has been demonstrated for critically ill patients with SIRS–MODS when treated with continuous renal replacement therapy as an adjunct to conventional therapy. 200 

Ventilatory strategies involving stretch of unrecruited lung units (i.e. , high tidal volume, low positive end-expiratory pressure), appear to cause—or potentiate— ALI. 201Furthermore, such injury may contribute to the inflammatory response by several mechanisms, such as facilitating translocation of intrapulmonary bacteria across the alveolar capillary barrier, 202accumulation of intrapulmonary cytokines, 203and release of cytokines into the systemic circulation. 204The role of ventilatory strategies in the inflammatory response following CPB is not yet known. However, in patients with acute respiratory distress syndrome, ventilation strategies that minimize overdistention and recruitment– derecruitment of the lung attenuate the inflammatory response. 205 

A controlled, self-limiting inflammatory response to cardiac surgery has important beneficial effects such as immune system priming, which may aid in preventing perioperative infection and promoting wound healing (fig. 5). CPB priming of neutrophils may be beneficial in preparing the host to mount a robust response to the physiologic stresses of the perioperative period. In this context, experimental pretreatment with inhaled endotoxin induces pulmonary leukocyte recruitment, which protects against subse-quent bacterial lung infections. 206Gut translocation of endotoxin in minute amounts may be a physiologically important phenomenon to boost the reticuloendothelial system, especially Kupffer cells. 207TNF-α and IL-1β have been shown to be necessary for wound healing, 208and TNF-α may even function to down-regulate certain aspects of the inflammatory response. 209Finally, IL-1β pretreatment may reduce the severity of ischemia and reperfusion injury. 210 

An uncontrolled inflammatory response appears to play a significant role in the morbidity or mortality observed in patients undergoing CPB (fig. 5). The inflammatory response contributes to the pathogenesis of acute pulmonary, cardiovascular, neurologic, splanchnic, hematologic, and immune system dysfunction following cardiac surgery. In addition, although less well documented, there is evidence to suggest that the inflammatory response contributes to the pathogenesis of subacute sequelae, such as postoperative fever, 211–215postcardiotomy pericarditis, 211,216and pleural effusions after CPB. 217 

Acute lung injury, defined as the triad of hypoxemia (Pao²/Fio²≥ 300), bilateral pulmonary infiltrates, and normal pulmonary capillary wedge pressure, 218is a serious and poorly understood complication following cardiac surgery. The risk 219,220and severity 221of ALI has been consistently linked to duration of CPB. Other issues related to CPB, such as the type of oxygenator used, may also contribute to the risk of ALI. 178,222Severe ALI following CPB, while relatively uncommon (1–3%), 220,223has been associated with a 50% mortality. 220Lesser degrees of lung injury, such as reduced oxygenation index, increased V̇/Q̇ mismatch, decreased lung compliance, are seen in up to 12% of patients. 96,219Pulmonary injury is detectable even following uncomplicated CPB using sensitive measures of ALI, such as protein accumulation index and bronchoalveolar lavage neutrophil and myeloperoxidase concentrations. 221Early pulmonary dysfunction after cardiac surgery increases morbidity, including renal, neurologic, and infectious complications, duration of mechanical ventilation, ICU and hospital stays, and risk of mortality. 219 

The inflammatory response has deleterious effects on the pulmonary circulation 107,108,224and the lung parenchyma. 107Several lines of evidence implicate the inflammatory response in the pathogenesis of postoperative pulmonary dysfunction. Histologic evidence demonstrates that CPB can cause profound pulmonary endothelial, epithelial, and interstitial damage. 107,225Up to 50% of circulating neutrophils are sequestered in pulmonary capillaries during rewarming, with subsequent degranulation contributing to pulmonary vascular endothelial damage. 108Postoperative elevations of granulocyte elastase, an index of neutrophil lysosomal degranulation, correlate with deteriorations in the oxygenation index, alveolar–arterial gradient, and intrapulmonary shunt. 96,226Increases in pulmonary vascular permeability following CPB correlate with markers of lipid peroxidation and a reduction of plasma antioxidants, both indicators of an oxygen-radical–induced injury process. 224 

Strategies that attenuate the inflammatory response may reduce post-CPB ALI. While serum and alveolar proinflammatory mediator concentrations correlate with post-CPB pulmonary dysfunction, 55,226mechanical mediator removal may decrease lung injury. 56,227Leukocyte depletion during CPB, 228or blockade of neutrophil adhesion, 111,229attenuates postoperative lung dysfunction. Finally, inhibition of complement activation results in significant reduction of post-CPB ALI. 230 

Major perioperative cardiovascular complications (cardiac death, myocardial infarction, heart failure) occur in at least 10% of CABG patients. 231The Multicenter Study of Perioperative Ischemia Research Group, in a multicenter study involving 566 patients after CABG, found that up to 25% fulfilled either electrocardiogram (i.e. , presence of Q waves), CK-MB fraction, or autopsy criteria for myocardial infarction. 232Nearly 10% of patients had a Q-wave infarct, while 20.5% fulfilled CK-MB criteria for myocardial infarction. 232 

The inflammatory response may be involved in the pathogenesis of post-CPB cardiovascular dysfunction. Increased hemodynamic instability during cardiac surgery may result from systemic spillover of proinflammatory cytokines, such as IL-6. 149Furthermore, there appears to be a clear link between CPB-induced inflammatory mediators and postbypass myocardial stunning, ischemia, and dysfunction, and β-adrenergic desensitization. 233,234TNF-α, IL-1β, and IL-6 have been implicated in cardiac myocyte refractoriness to adrenergic stimulation following CPB. 234Myocardial performance in elderly patients after CPB is suppressed by TNF-α, 25while anti-TNF antibodies reverse the myocardial depression seen in sepsis. 235Left ventricular wall motion abnormalities and myocardial ischemic episodes after CPB correlate with IL-6 or IL-8 concentrations. 233 

Alterations in NO homeostasis may play an important role in the pathogenesis of cardiovascular events following cardiac surgery. The human myocardium constitutively produces NO via  cNOS, and this is regulated by the contractile state of the heart, 236possibly by a β-adrenergic signaling pathway. 237This constitutively produced NO has important cardioprotective effects, including regulation of vascular tone (fig. 3), myocardial contractility, 238,239coagulation, 240and pulmonary endothelial function. 84However, up-regulation of iNOS by proinflammatory cytokines following CPB may greatly increase myocardial NO production. 241In this context, NO may have deleterious effects and contribute to postoperative myocardial dysfunction, particularly myocardial stunning. 241IL-1β up-regulates iNOS-mediated NO production 242and has been demonstrated to depress myocardial contractility through a NO-dependent mechanism. 64Mechanisms underlying NO-mediated myocardial dysfunction may include inhibition of mitochondrial activity, 243induction of adrenergic refractoriness, 234increased platelet adhesiveness, 240and peroxynitrite-induced cellular damage. 66 

Neurologic complications increase perioperative morbidity following cardiac surgery, resulting in longer hospitalization, and may increase mortality up to 10-fold. 13Focal cerebral deficits, such as transient ischemic attacks and stroke, occur in 1–3% of patients, while less severe clinical abnormalities, such as seizures, are observed in 5–10%. 244,245Postoperative neurologic dysfunction, including cognitive dysfunction and disability, unrelated to focal injury, is seen in 69% of patients immediately postoperatively and persists for at least 1 month in 36%. 246Patients with the apolipoprotein E-ε 4 allelle 247or previous neurologic injury 248may be predisposed to neurologic complications after CPB. In the largest series to date, the mortality rate was 29% in patients with adverse cerebral complications. 245 

The inflammatory response plays a pivotal role in the pathogenesis of post-CPB neurologic injury. The role of endothelial dysfunction and the endothelial–neutrophil interaction in neurologic injury after CPB is well documented. Loss of cerebrovascular endothelial-mediated vasodilation may contribute to the pathogenesis of postoperative perfusion deficits. 249NO formed by up-regulation of neuronal NOS, e.g. , in response to glutamate, has been implicated as a potent neurotoxin in animal models of post-CPB neurologic injury. 250,251 

Strategies that attenuate the inflammatory response may decrease neurologic injury. Experimental blockade of selectin adhesion molecules reduces cerebral injury in an animal model. 252Aprotinin, a serine protease inhibitor with important hemostatic and antiinflammatory properties, decreases the incidence of stroke after CPB. 253Finally, heparin coating of CPB circuits, which improves its biocompatibility, thereby reducing contact activation of the inflammatory response, may reduce the incidence of neurologic dysfunction in humans. 254 

Perioperative renal dysfunction occurs in 7–13% of patients, with 1–1.5% requiring some form of dialytic therapy. 18,255Renal insufficiency following cardiac surgery increases ICU and hospital stay and greatly increases mortality (27%vs.  0.9%). 18The mortality rate for patients requiring postoperative dialysis ranges from 28–64%. 18,255The incidence of renal dysfunction is directly related to the duration of CPB. 18 

The role of the inflammatory response in the pathogenesis of renal dysfunction continues to be defined. Renal ischemic–reperfusion injury combined with the inflammatory response to CPB may be important causes of renal dysfunction after CPB. Hypoperfusion of the superficial renal cortex has been demonstrated during the rewarming phase of CPB in an animal model. 256The inflammatory response may exacerbate renal hypoperfusion, both indirectly as a result of hemodynamic instability, and directly via  renal arteriolar vasoconstriction and altered intrarenal distribution of perfusion due to alterations in catecholamines and NO concentrations. TNF-α released during CPB induces glomerular fibrin deposition, cellular infiltration, renal cell apoptosis, and vasoconstriction, leading to a reduction in the glomerular filtration rate. 17Furthermore, anti–TNF-α strategies have been proposed to reduce renal insufficiency after CPB. 17 

Hepatic dysfunction following CPB is common, with up to 47% of patients developing at least transient postoperative dysfunction. 257Postoperative hepatic dysfunction is related to the duration of CPB and may greatly increase mortality. 258 

The mechanism of hepatic injury after CPB and the role of the inflammatory response is not well characterized. Nevertheless, there is some evidence that the inflammatory response to CPB may play a role in hepatic injury. Hepatic ischemia–reperfusion injury may result from hepatic venous congestion during weaning from CPB. 19High concentrations of C3a and C4a have been associated with post-CPB liver dysfunction in humans. 257During hypothermic bypass, portal concentrations of endothelin-1, a potent vasoconstrictor, have been associated with decreased hepatic blood flow and post-CPB liver dysfunction in an animal model. 259In patients following CPB, hepatic perfusion may be decreased for up to 24 h. 22Finally, there is increasing evidence that cytokines TNF-α and IL-6 may contribute to the pathogenesis of hepatocellular dysfunction. 260 

Cardiopulmonary bypass–induced hemostatic defects may contribute significantly to perioperative morbidity. Potential mechanisms include direct contact activation of the coagulation and fibrinolytic cascades by the bypass circuit, platelet dysfunction, and capillary leak due to endothelial damage. The inflammatory response may be central to the development of these hemostatic defects. The amount of postoperative blood loss has been correlated with degree of activation of the complement cascade. 261Mechanical removal of cytokines, such as by hemofiltration, has been associated with a reduction in postoperative blood loss following cardiac surgery. 56 

Cardiopulmonary bypass–associated impairment of platelet function may be cyokine mediated. 262Endotoxin and IL-1β stimulate the release of von Willebrand factor from the vascular endothelium, which may promote activation and localization of platelets and enhance thrombogenicity at inflammatory foci. 263Platelet activation during CPB may contribute, through IL-1β release, to endothelial cell activation. 262Preoperative harvest of platelet-rich plasma with reinfusion following CPB may result in significantly less pulmonary dysfunction and shorter ICU stays, in addition to improved hemostatic function. 264In addition, nonsurgical postoperative bleeding is correlated with both leukocyte count and the percentage change in leukocyte count over the course of CPB. 265Finally, CPB with more biocompatible circuits may decrease platelet activation, fibrinolysis, and thrombin generation and increase platelet preservation. 266 

Cardiopulmonary bypass–associated immunosuppression may play an important role in the development of postoperative infectious complications. Immunosuppression may result from a predominance of antiinflammatory cytokine production. 7Cell-mediated immunity is also altered with decreases in CD3+ T-lymphocyte and CD4+ T-helper cell counts, 267increases in CD8+ suppressor–cytotoxic T cells and monocyte counts, 268decreased lymphocyte responsiveness to mitogens, and suppression of the T-helper cell–induced cell-mediated inflammatory response after CPB. 267,269Pretreatment with indomethacin and thymopentin, which stimulates activation and differentiation of the T-lymphocytes, appears to restore certain aspects of this response. 269IL-10 may play a role in postoperative immunosuppression, with IL-10 gene expression correlating with decreased monocyte HLA-DR antigen expression in a study of patients undergoing major abdominal surgery. 270These alterations in the inflammatory response highlight the need to consider both the potential beneficial and adverse aspects of this response and further underline the clinical benefit of a balanced, controlled inflammatory response to CPB.

The development of strategies to control the inflammatory response following cardiac surgery is currently the focus of considerable research efforts. Diverse techniques, including maintainence of hemodynamic stability, minimization of exposure to CPB circuitry, and pharmacologic and immunomodulatory agents have been examined in clinical studies. Table 5summarizes the findings of recent randomized clinical trials in this field.

Accurate risk stratification would permit selective delivery of potentially useful therapies to those patients who might be expected to derive most benefit. Clinical risk assessment has taken the form of complex multivariate modelling. 271,272An approach that combines assessment of “fixed” baseline risk together with “dynamic” event-related risk related to adverse perioperative events probably holds most promise (fig. 6). In this regard, several key indices of the inflammatory response appear to predict postoperative morbidity following cardiac surgery and may help to further stratify patient risk. In the preoperative phase, elevation of C-reactive protein concentrations predicts a greater likelihood of septic complications and need for catecholamine therapy, longer duration of respiratory support, and increased duration of ICU stay, 273while low preoperative immunoglobulin M antiendotoxin core antibody concentrations independently predict poor postoperative outcome. 27Postoperatively, early elevations in serum soluble TNF receptor-p55 concentrations in high-risk cardiac surgical patients predicts increased mortality. 274Finally, the characterization of particular genotypes, such as apolipoprotein E-ε 4 allele, 247which predicts risk of post-CPB neurologic injury, is a novel approach that holds promise.

Concerns regarding complications and cost surrounding the use of CPB has led to renewed interest in off-pump coronary artery bypass grafting (OPCAB) techniques. Extensive observational data suggest a relation between poor outcome and the use of (1) aortic cross-clamping and concomitant global ischemia and reperfusion injury and (2) extracorporeal circulation. Avoidance of aortic cross-clamping and CPB may decrease the inflammatory response and improve postoperative organ function and patient outcome, particularly in high-risk patients.

Off-pump coronary artery bypass grafting does reduce the elaboration of key mediators of the systemic inflammatory response. OPCAB decreases concentrations of cytokines such as TNF-α, 275,276IL-6, 277IL-8, 276–280IL-10, 279,280and TNFsr1 and 2. 277,280OPCAB attenuates the cellular inflammatory response, decreasing neutrophil and monocyte counts, 278neutrophil elastase, 276,278,281and E-selectin 276concentrations. Indices of complement activation, such as C3a 276,281and C5a, 280are decreased. In addition, OPCAB attenuates other indices, including platelet β -thromboglobulin 281and procalcitonin. 282Finally, OPCAB decreases reactive oxygen species–induced injury. 276 

The clinical trials reported to date suggest that OPCAB may attenuate indices of postoperative organ dysfunction but does not eliminate SIRS following cardiac surgery (table 5). 276,278,281,283,284Initial clinical experience with selected patients suggests reduced cardiovascular, 275,279pulmonary, 275neurologic, 285and hemostatic 275dysfunction, with decreased morbidity and shorter ICU and hospital stay. 275One retrospective study comparing elderly patients undergoing OPCAB to matched conventional controls indicated that OPCAB decreased postoperative cardiovascular complications and expedited ICU and hospital discharge. 286However, these trials were small, with enrollment restricted to low-risk patients, often with specific single or double vessel coronary lesions. There are no reported randomized trials of OPCAB in high-risk patients. A retrospective comparison of the first 55 patients who underwent OPCAB at Duke University Medical Center to a larger cohort who underwent conventional CABG failed to demonstrate a reduction in renal risk. 287Therefore, the short-term benefits of OPCAB remain to be convincingly demonstrated.

Advances in minimally invasive surgery in other areas, such as laparoscopic and thoracoscopic procedures, have prompted interest in approaches that avoid full median sternotomy. 288Minimally invasive cardiac surgery techniques may be utilized in the presence or absence of CPB. The surgical incision may affect the inflammatory response generated, with reduced complement activation following limited anterolateral thoracotomy as opposed to median sternotomy, 289although this is disputed. 280One nonrandomized study of patients undergoing mitral valve surgery via  limited right anterolateral thoracotomy and port access suggested that this group resumed work or normal activity earlier than comparable patients undergoing conventional median sternotomy. 290However, the efficacy of minimally invasive cardiac surgery remains to be clearly demonstrated. At present, operating times are considerably longer with minimal access techniques. 290 

An intermediate, less invasive approach may be to use the patient's own lungs as an oxygenator via  bilateral extracorporeal circulation. This technique, introduced by Drew and Anderson, allows good surgical access yet avoids the need for an artificial oxygenator. A recent randomized clinical trial, focusing on this method as a means of attenuating the inflammatory response, reported decreased concentrations of IL-6 and IL-8 and an attenuation of postoperative hemostatic and pulmonary dysfunction (table 5). 291 

Since aortic cross-clamping and CPB were universally used for all forms of cardiac surgery, their relative role in the pathogenesis ofpostoperative complications has been unclear. OPCAB will help to delineate the intrinsic effects of bypass on the inflammatory response to cardiac surgery. OPCAB and minimally invasive cardiac surgery have demonstrated encouraging short-term benefits in the small-scale clinical studies of low-risk patients carried out to date, but clinical studies have not been carried out in the high-risk groups most likely to benefit. Finally, concerns regarding the quality of graft anastomoses achieved with OPCAB or minimal access techniques remain to be addressed.

Improving the biocompatibility of the CPB circuit in order to reduce contact activation of the immune system, particularly the complement cascade, may be a useful strategy to limit the inflammatory response. Potential approaches include use of more biocompatible materials in the circuit or modifications of the surface of the circuit by coating with compounds that are less immunogenic.

Heparin-coated CPB circuits (HCCs) enhance biocompatibility, reduce contact activation, and may decrease postoperative cardiovascular, respiratory, hemostatic, and neurologic dysfunction. 254The method by which the circuit is coated and the type of heparin used may have implications for its effects on the coagulation and complement systems. The Duraflo II HCC, which uses ionically bonded unfractionated heparin (Duraflo II surface; Baxter Healthcare Corp., Irvine, CA), reduces kallikrein and complement activation 49but is less effective in attenuating coagulation or fibrinolysis. 292The Carmeda Bioactive Surface system (CBAS; Medtronic Inc., Minneapolis, MN) uses end-attached covalently bonded heparin that has been fragmented by treatment with nitric acid. The Carmeda circuit appears superior to the Duraflo II in reducing complement, 293neutrophil 294activation, and endothelin-1 concentrations. 294 

Clinical studies to date suggest that the beneficial effects of HCCs are confined to high-risk patients (table 5). In low-risk patients, HCCs decrease neutrophil activation, 295decrease myocardial injury, 295and reduce complement activation. 296The largest study to date in low-risk patients revealed no overall clinical benefit, 297although subgroup analysis suggested that women and patients with prolonged aortic cross-clamp times may benefit from HCC. 297Duraflo II HCCs decreased the duration of ventilatory support and ICU stay and reduced the incidence of poor outcome (death or prolonged ICU stay) in a large study of high-risk patients. 49However, a similar but smaller study failed to confirm these benefits. 298The benefits of HCC-based technology may be more apparent in patients with preexisting organ dysfunction. 49Outcome may be enhanced when Duraflo HCCs are combined with the use of silicone-coated oxygenators. 299HCC may attenuate the proinflammatory response more markedly and may have greater myocardial protective effects where perfusion times are prolonged, especially in heart and heart–lung transplantation. 300 

Other strategies to improve biocompatibility having therapeutic potential include coating of circuitry with phosphatidylcholine, 301silicone, 302synthetic proteins 303and polymers, 304or surface–modifying additives. 305,306Decreasing oxygenator surface area may also decrease activation of the inflammatory response. 307 

Selective digestive decontamination (SDD) is a technique to reduce the gut content of enterobacteria. This is achieved by preoperative administration of oral nonabsorbable antibiotics such as polymyxin E, tobramycin, and amphotericin B, and has been demonstrated to reduce plasma concentrations of endotoxin, TNF-α, and IL-6 in patients undergoing CPB. 23A recent metaanalysis of SDD suggests that it reduces rates of postoperative infection, but not mortality, in patients undergoing cardiac surgery. 308Since mortality reduction with SDD in critically ill patients appears to be related to baseline mortality risk, 309trials of SDD in cardiac surgery thus far contain too many low-risk patients, resulting in inadequate study power. SDD may prove worthwhile in high-risk cardiac surgical patients, 308but since its use raises both practical issues (notably the logistics of performing it) and theoretical concerns (changes in bacterial flora, emergence of resistance), its adoption is unlikely pending further studies.

Hypoalbuminemia and low body mass index independently predict increased morbidity and mortality after cardiac operations. 310In an early study, well-nourished patients undergoing valve surgery had a much shorter hospital stay compared to those with preoperative malnutrition. 311Laboratory evidence in animals suggests that protein-calorie malnutrition decreases left ventricular function 312and that myocardial glycogen concentration correlates with left ventricular function following CPB. 313 

The beneficial role of early institution of enteral nutrition, particularly “immunonutrition,” which contains supplements such as arginine, purine nucleotides, and ω-3 fatty acids, which are considered to enhance immune function, has been established in other groups of postoperative and critically ill patients. In critically ill patients, immunonutrition reduced the duration of ICU and hospital stay, infectious complications, duration of SIRS, and mechanical ventilation compared to patients receiving conventional nutrition. 314In patients for elective gastrointestinal surgery, preoperative and postoperative immunonutrition has beneficial effects on immune function, complication rate, and duration of hospital stay. 315,316The use of glutamine supplementation may improve survival of patients with organ failure who require parenteral nutrition. 317 

There is no information available concerning the effect of nutritional support in patients undergoing cardiac surgery who have a complicated postoperative course.

Perioperative hemodynamic instability, especially post-CPB low cardiac output syndrome, is a marker for later adverse outcome and death. Splanchnic perfusion abnormalities, as measured by tonometrically derived intramucosal pH and Pco², strongly predict postoperative morbidity in cardiac surgical patients. 130Covert hypovolemia may exacerbate splanchnic hypoperfusion during and after CPB, and this may be treatable by increasing intravascular volume. Other therapeutic options include pharmacologic and mechanical circulatory support.

Optimization of intravascular volume status may be beneficial even in apparently stable patients. 318In low-risk cardiac surgical patients, intraoperative intravascular volume expansion with the objective of maximizing stroke volume resulted in improved gastric intramucosal pH values, less hemodynamic interventions, lower complication rates, and shorter ICU and hospital stay. 318Furthermore, low intraoperative filling pressures and high pressor usage may contribute to increased gut permeability and endotoxemia. 319Excessive hemodilution during CPB may play a role in the pathogenesis of post-CPB low cardiac output syndrome. In a multicenter observational study, excessively low hematocrit values during CPB independently predicted in-hospital mortality, need for intraaortic balloon pump usage, and return to bypass following failed separation. 320 

There are no outcome-based data to support the routine use of pharmacologic interventions to maximize splanchnic perfusion in patients undergoing cardiac surgery. However, patients failing to meet preset hemodynamic goals may benefit from such measures. In a recent randomized trial in which early postoperative cardiovascular function was optimized, using preload augmentation and dobutamine to normalize mixed venous oxygen saturation values and lactate concentration, optimized patients had reduced length of hospital stay and reduced organ dysfunction at the time of discharge compared with patients receiving standard care. 321Pharmacologic interventions to maximize splanchnic perfusion may attenuate the inflammatory response. Immune cells contain type IV and type III phosphodiesterase, 322and phosphodiesterase inhibitors appear to directly limit inflammatory activation and organ dysfunction in sepsis models. 322,323Milrinone attenuates the reduction in gastric intramucosal pH, reduces both venous and hepatic endotoxin concentrations, and may decrease postoperative IL-6 concentrations in healthy patients undergoing cardiac surgery, 24although this has been disputed. 324Dopexamine attenuates the postoperative increase in IL-6 concentrations 325and reduces gastrointestinal permeability, 326but does not improve splanchnic perfusion (as measured by intramucosal pH) or decrease plasma endotoxin concentrations, 325following CPB.

The elective use of mechanical circulatory support, such as preoperative intraaortic balloon pump, in selected high-risk patients, especially the elderly 327,328and those with severe left ventricular dysfunction, 329may reduce the incidence of postoperative low cardiac output syndrome, mortality, and ICU stay without significant increases in morbidity or cost.

Hemofiltration is a process that uses ultrafiltration, i.e. , the processes of convection, osmosis under a hydrostatic pressure gradient, to remove fluid and low-molecular-weight substances from plasma. Initially introduced to treat patients with renal failure and to correct accumulation of extravascular water following CPB, hemofiltration appears to exert beneficial antiinflammatory effects, particularly in pediatric patients. 56Hemofiltration may remove proinflammatory mediators, with reductions in postoperative TNF-α, IL-1, IL-6, IL-8, C3a, and myeloperoxidase concentrations. 56,227,330This technique improves hemodynamic stability 227and early postoperative oxygenation, 56,227,331and reduces postoperative blood loss 56,227,331and duration of mechanical ventilation 227,331in pediatric cardiac surgery. Hemofiltration may reduce pulmonary hypertension after congenital heart surgery, possibly by facilitating removal of endothelin-1. 330Modified hemofiltration 332after CPB improves intrinsic left ventricular systolic function and diastolic compliance, increases blood pressure, and decreases inotropic drug use in the early postoperative period in infants. 333 

Hemofiltration appears to be less effective in adults, with no decrease observed in postoperative blood loss, average bank blood transfused, postoperative weight gain, length of stay, or overall patient costs. 334,335Hemofiltration in adults undergoing CPB is less effective in removing proinflammatory cytokines than in pediatric patients, 335,336perhaps explaining its apparent lack of efficacy in this population.

Leukocytes play a central role in the inflammatory response to cardiac surgery. Leukocyte depletion during cardiac surgery, by means of leukocyte-specific filters, decreases circulating leukocyte 228,337,338and platelet concentrations 339and attenuates indices of oxidative stress 340and inflammation. 337There is increasing evidence that leukocyte depletion may attenuate pulmonary and myocardial injury following CPB. Benefit appears most consistent in patients with risk factors such as left ventricular dysfunction, urgent surgery, or long CPB time. Leukodepletion has been shown to improve postoperative respiratory function in CPB patients, 338,339particularly in patients with a low preoperative oxygenation capacity or long CPB time. 339In addition, leukocyte depletion of the residual heart–lung machine blood, which contains high quantities of activated leukocytes, prior to retransfusion, improved lung function in patients undergoing elective CABG. 228Leukodepletion during CPB, combined with leukodepletion of transfused blood, decreased indices of myocardial cell injury in patients undergoing urgent CABG for unstable angina. 340Conversely, in low-risk patients, depletion of activated neutrophils during CPB did not confer clinical benefit. 341Limiting leukocyte depletion to the “reperfusion phase” of CPB (following aortic unclamping) does not appear to confer any clinical benefit in CABG patients. 342 

Leukocyte depletion of blood cardioplegia alone attenuated myocardial cell injury and improved early myocardial function in patients with left ventricular dysfunction undergoing CABG with CPB. 112,343Leukodepletion of terminal blood cardioplegia, (i.e. , blood cardioplegia administered for 10 min immediately prior to aortic unclamping as an adjunct to crystalloid cardioplegia) decreased myocardial injury and improved cardiac function in patients with left ventricular hypertrophy undergoing valve surgery. 344 

The immunomodulatory effects of leukocytes in allogeneic blood has focused attention on the potential benefits of leukodepleting stored blood. 345A large-scale clinical trial conducted in CPB patients demonstrated that leukocyte depletion of transfused blood significantly reduced the overall 60-day mortality. 346The difference in mortality was predominantly due to a dramatic reduction in noncardiac causes of death, particularly multiorgan failure. 346In addition, leukodepletion reduced the postoperative infection rate in patients who received more than 3 units of blood. 346 

Many effector proteins of the cytokine, complement, and hemostatic cascades are serine proteases, i.e. , when activated they catalyze the next step in the cascade by hydrolyzing and activating further proteins, a process termed “cascade amplification.” Control processes that limit inflammation to the sites of injury and reduce systemic inflammation include serine protease inhibitors. Aprotinin is the best known and studied of these inhibitors.

Aprotinin, a complex polypeptide and nonspecific serine protease inhibitor, has clearly been demonstrated to prevent excessive blood loss during cardiac surgery. 347In addition, aprotinin has multiple actions that may suppress the inflammatory response, particularly at higher dosages. Antiinflammatory effects include attenuation of platelet activation, maintenance of platelet function, 348decreased complement activation, 348inhibition of kallikrein production, 348decreased release of TNF-α, 349IL-6, and IL-8, 350inhibition of endogenous cytokine-induced iNOS induction, 351decreased CPB-induced leukocyte activation, 348,352and inhibition of up-regulation of monocyte and granulocyte adhesion molecules. 95,353 

In clinical studies, high-dose aprotinin reduces postbypass myocardial ischemia and myocyte damage 354and length of hospital stay in high-risk patients. 355However, a pump-prime-only dose of aprotinin may increase the risk of postoperative myocardial infarction. 356Levy et al.  found no such increase in the incidence of perioperative myocardial infarction in patients undergoing repeat CABG. 253Concerns over graft patency following aprotinin therapy have been reduced by the IMAGE trial, which found no difference in early (10-day) patency rates for internal mammary artery grafts 357or for saphenous vein grafts after controlling for confounding factors. 358 

Aprotinin may reduce pulmonary and cerebral injury following CPB. Aprotinin decreases experimental CPB-induced and cytokine-induced bronchial inflammation 359and was demonstrated to attenuate lung reperfusion injury following CPB in one small clinical study. 360An early report of use of aprotinin in high-risk cardiac surgery patients indicated an incidence of fatal stroke of 0.5%, compared to rates of 2–3% in contemporary studies that did not use aprotinin. 361,362A multicenter trial of repeat CABG patients found that the incidence of stroke was reduced with high- or low-dose aprotinin. 253Lemmer et al.  failed to show a significant decrease in the incidence of stroke with three different dosage regimens in their large-scale multicenter study. 356However, a pooled analysis of six trials, including the aforementioned trials, found that high-dose aprotinin significantly reduced the incidence of stroke. 363Initial concerns over the potential for adverse effects of aprotinin on renal function appear unfounded. 356 

No single study of aprotinin has clearly demonstrated improved patient outcome to date. 253,356However, a recent metaanalysis reported that aprotinin reduces surgical blood loss, allogeneic blood transfusion, and the need for rethoracotomy, and decreases perioperative mortality almost twofold, with no increase in the risk of myocardial infarction. 347These data provides strong support for the use of aprotinin in patients undergoing cardiac surgery.

Pentoxifylline is a nonspecific phosphodiesterase inhibitor with diverse antiinflammatory effects, many of which may be mediated by inhibition of phosphodiesterase IV. 364These include attenuation of TNF-α release in sepsis, 365,366decreased endotoxin and cytokine activation of neutrophils, 367reduction of indices of endothelial injury and permeability, 368decreased pulmonary leukocyte sequestration, 369and attenuation of increases in pulmonary vascular resistance. 369Clinical studies to date have been limited. In one study, pentoxifylline decreased the duration of ventilation and hemofiltration and the incidence of MODS when administered postoperatively to selected high-risk cardiac surgical patients 370and may be of therapeutic benefit in human sepsis. 371However, pentoxifylline treatment did not improve postoperative renal 372or pulmonary 373function in other small clinical studies. Most recently, in elderly cardiac surgical patients, pentoxifylline attenuated the increase in neutrophil elastase, C-reactive protein, and proinflammatory cytokines (IL-6, IL-8, and IL-10). These patients also had reduced requirements for vasoactive medication and a shorter time to tracheal extubation. 374In a parallel study, the same investigators reported improved splanchnic perfusion and hepatic–renal function with pentoxifylline. 375 

Generation of reactive oxygen species (ROS) (hydrogen peroxide and the superoxide and hydroxyl radicals) occurs upon reperfusion following bypass, 376and these may be important contributors to tissue injury. Leukocytes activated during bypass may also release substantial amounts of cytotoxic ROS. 96,97When present in equimolar concentrations, superoxide and NO may combine form peroxynitrite, a more reactive and injurious free radical. 66,67Myocardial antioxidant enzymes, including glutathione reductase, superoxide dismutase, and catalase, are activated in proportion to the degree of myocardial ischemia and reperfusion injury. 377Host antioxidants become depleted after CPB, 378,379presumably as a result of consumption by free radicals. When ROS production exceeds host defense scavenging capacity, cellular injury results. 376,380There is an inverse correlation between preoperative total plasma antioxidant status and lipid peroxidation, the latter of which is directly correlated with indices of myocardial cellular injury. 379Furthermore, post-CPB coronary endothelial dysfunction appears to be partially mediated by ROS. 381Free radical scavengers, such as enzymatic scavengers, antioxidants, and iron chelators, may be potentially useful therapeutic adjuncts to control the deleterious effects of the inflammatory response.

High-dose vitamin C (ascorbic acid) has been demonstrated to effectively scavenge free radicals, decreasing cell membrane lipid peroxidation 376,382and indices of myocardial injury, and improving hemodynamics with a shorter ICU and hospital stay. 382Vitamin E (α-tocopherol) reduces plasma concentrations of hydrogen peroxide, a marker of free radical concentrations, 383and decreases cell membrane lipid peroxidation 376following CPB. Preoperative supplementation with a combination of ascorbic acid, α -tocopherol, and allopurinol reduced cardiovascular dysfunction in both stable and unstable patients undergoing CABG. Unstable CABG patients sustained less myocardial injury and a decreased incidence of perioperative myocardial infarction. 384A more recent trial of combined α-tocopherol and ascorbic acid supplementation in CABG surgery revealed no detectable decrease in myocardial injury. 15 

High-dose N -acetylcysteine before or during bypass appears to act as a free radical scavenger 385and reduces the neutrophil oxidative burst response 385and elastase activity. 386In an early interventional trial in patients with established acute lung injury, N -acetylcysteine was shown to improve oxygenation and lung mechanics, although no impact on progression to acute respiratory distress syndrome was noted. 387 

Allopurinol is an inhibitor of the enzyme xanthine oxidase, a pivotal generator of free radicals during reperfusion injury. Allopurinol may decrease myocardial formation of cytotoxic free radicals, 380,388lower markers of myocardial cellular injury, 389and improve recovery of myocardial function following CPB. 390,391However, other studies have demonstrated no improvement in either myocardial function 392or myocardial cellular injury with allopurinol use, 388,393casting doubt on its therapeutic potential.

Pretreatment of patients with mannitol reduces myocardial formation of cytotoxic free radicals after CPB in humans. 380Other free radical scavengers–antioxidants that appear from animal studies to possess therapeutic potential include methionine, reduced glutathione, dimethylthiourea, mercaptopropionyl glycine, superoxide dismutase, catalase, and desferrioxamine. 381,394–397 

The use of corticosteroids in the context of CPB continues to be controversial because of their potential risks. Past negative experience, particularly with the use of corticosteroids in septic shock, 398has served to emphasize the need for caution when considering corticosteroid use, even in noninfective inflammatory conditions, where their potent antiinflammatory actions might be expected to be beneficial. However, there have been significant advances in our understanding of the molecular mechanisms by which corticosteroids might blunt the inflammatory response to cardiac surgery.

Corticosteroid pretreatment may blunt the inflammatory response in humans by several distinct mechanisms. Administration of glucocorticoids prior to CPB may attenuate endotoxin release 399and complement activation. 400,401Methylprednisolone lowers post-CPB concentrations of the proinflammatory cytokines TNF-α, 349IL-6, and IL-8, 402and increases concentrations of the antiinflammatory cytokines IL-10 and IL-1ra, 402but not IL-4. 403Corticosteroids also attenuate post-CPB leukocyte activation, 20neutrophil adhesion molecule up-regulation, 349and pulmonary neutrophil sequestration. 401 

Prebypass administration of methylprednisolone in aprotinin-treated patients improves early postoperative indices of pulmonary, cardiovascular, hemostatic, and renal function. 404Glucocorticoid pretreatment may improve cardiac performance 405and reduce evidence of bronchial inflammation following CPB. 406Low-dose methylprednisolone in the pump prime solution appears to attenuate myocardial cell damage. 407Dexamethasone 213,214and methylprednisolone 212,215decrease the incidence of postoperative fever. In animal studies, corticosteroid pretreatment improved several indices of lung injury, including pulmonary compliance, alveolar– arterial gradient, pulmonary vascular resistance, and extracellular fluid accumulation. 408However, the ability of corticosteroid pretreatment to attenuate post-CPB pulmonary inflammation, 409endotoxemia, 400and complement activation is disputed. 20,410 

The clinical implications of corticosteroid use are not yet fully elucidated, and clear benefit is not yet demonstrated. The dosage, formulation, and timing of administration of corticosteroids may be critical, and differences in dosage regimens may explain conflicting results. Preoperative combined with prebypass administration may be superior to prebypass administration alone. 408It is premature to advocate the use of corticosteroids in the absence of proven outcome benefit, determination of optimal dosage regimens, and characterization of harmful effects, e.g. , immunosuppression, which may result from their use.

Aspirin, the prototype nonsteroidal antiinflammatory drug (NSAID), is widely used in cardiac surgical patients for the purposes of pain relief and antiplatelet activity. However, the potential for NSAID's inhibitors to attenuate the inflammatory response to cardiac surgery has not been widely evaluated in clinical trials. Traditional NSAIDs, such as indomethacin, inhibit both the constitutive cyclooxygenase 1 (COX-1) as well as COX-2, the inducible isoform activated by inflammatory stimuli. Nonspecific COX inhibition attenuates the increase in pulmonary vascular resistance and ALI 411and reverses pulmonary microvascular dysfunction 412in CPB models. One older clinical study of indomethacin demonstrated that it decreased the duration of postoperative fever, chest pain, malaise, and myalgias following CPB. 211However, inhibition of COX-1 appears to increase free radical–generated isoprostane formation, which aggravates postischemic myocardial dysfunction. 413,414 

Specific COX-2 inhibitors exhibit considerable potential to attenuate the inflammatory response following cardiac surgery. COX-2 has been implicated in the pathogenesis of adverse events after cardiac surgery. 415,416COX-2 is up-regulated following CPB, 415in multiple tissues, including the brain, 416while COX-2 products, particularly thromboxanes 417and vasoconstrictor prostaglandins, are increased. 412COX-2 up-regulation following experimental CPB may contribute to postoperative coronary vasospasm 415and increased pulmonary vascular resistance. 418In addition, myocardial COX-2 is up-regulated during cardiac allograft rejection 419and myocardial infarction 420and contributes to endotoxin-induced myocardial depression. 421Inhibition of COX-2 attenuates the myocardial inflammatory response during cardiac allograft rejection, 419reduces endothelial dysfunction following myocardial ischemia and reperfusion, 422and improves cardiac function in experimental myocardial infarction. 420In addition, COX-2 inhibition decreases endotoxin-induced myocardial depression 421and lung ischemia and reperfusion injury. 423However, the clinical efficacy of specific COX-2 inhibitors in attenuating the inflammatory response to cardiac surgery remains to be determined.

Therapies that utilize endogenous soluble complement inhibitors may be a suitable approach to reduce contact activation and thereby control the inflammatory response. A recent two-stage randomized clinical trial of a monoclonal antibody specific for human C5 demonstrated its efficacy and safety in patients undergoing CPB. 50The generation of activated complement mediators and leukocyte adhesion molecule formation was inhibited in a dose-dependent manner. Furthermore, C5 inhibition resulted in a dose-dependent reduction in myocardial injury, postoperative cognitive deficits, and coagulation dysfunction. These data suggest that C5 inhibition may represent a promising therapeutic modality for preventing complement-mediated inflammation and tissue injury in patients undergoing CPB. 50Compstatin, a recently discovered peptide inhibitor of complement, may have the potential to prevent complement activation during and after cardiac surgery. In preliminary primate studies, compstatin completely inhibited in vivo  heparin–protamine-induced complement activation without adverse effects. 31 

Other promising strategies include the C1 inhibitor, recombinant soluble inhibitor-1, monoclonal antibodies to C3 and C5a, and strategies that attenuate complement receptor 3–mediated adhesion of inflammatory cells to the vascular endothelium. Utilization of membrane-bound complement regulators may also be feasible by means of transfection techniques. 424 

Direct antimediator therapies that focus upon the endotoxin molecule itself and the proinflammatory cytokine cascade following CPB offer new approaches. However, the complex pathway observed in patients with SIRS does not appear to readily respond to antimediator therapy. Multicenter clinical trials blocking endotoxin and proinflammatory mediators such as IL-1 or TNF-α conducted in SIRS patients have shown no benefit in reducing mortality secondary to sepsis. Reasons for the relative failure of immunomodulatory therapies to date may include the timing of intervention, the heterogeneous nature of the inflammatory response, and the reciprocating and redundant nature of the proinflammatory cascades. High circulating concentrations of antiinflammatory mediators, such as the cytokine antagonists IL-1ra, TNFsr1, and TNFsr2, may also limit the efficacy of therapies that aim to augment natural defenses against endotoxin or the proinflammatory cytokines. 425 

The experience with antimediator therapy in sepsis suggests a need for caution in considering the application of these therapies to control the inflammatory response to CPB. Nevertheless, antimediator therapies may be worthy of investigation for two reasons. These therapies may enhance our understanding of the inflammatory process following CPB. Their administration prior to bypass, in order to modulate the inflammatory pathways at their earliest stages, might constitute a more successful approach than that used in other clinical scenarios, such as in sepsis, where the inflammatory response may be already well developed before antimediator therapy is possible.

Current evidence suggests that therapeutic efforts in patients with SIRS should include modulation of endothelial cell function. Better definition of the molecular mechanisms of endothelial cell activation may facilitate development of therapies that allow selective inhibition of vascular endothelial activation. Adhesion molecule blockade may prevent neutrophil adherence during the first 24 h after CPB, thereby preventing the neutrophils from mediating widespread organ damage. In this regard, blockade of endothelial and neutrophil selectin adhesion molecules results in marked attenuation of cerebral injury in an animal model of CPB and deep hypothermic circulatory arrest. 252Inhibition of neutrophil adhesion markedly reduced pulmonary injury in a porcine model of CPB. 229However, there may be limits to this approach because adhesion molecule blockade increases susceptibility to infection. 426Finally, methods to prevent nuclear localization of the transcriptional activator NF-κB in order to prevent endothelial cell activation are also being studied in animal models. 98 

The development of organ dysfunction following CPB remains an indicator of poor outcome despite advances in resuscitation, drug, and adjunctive therapies.Organ support therapies, e.g. , mechanical ventilation, can sustain life but may impair both local and remote organ function, e.g. , ventilator-associated lung injury. 201A better understanding of the inflammatory response to cardiac surgery may be the key to development of successful strategies to minimize patient morbidity. With some exceptions, such as recent data on HCCs and inhibition of complement activity, the link between the inflammatory response and adverse clinical sequelae, while persuasive, is currently associative rather than causative. A primary research priority must be to establish direct causal links between (and mechanistic insights into) the inflammatory response to surgery or CPB and clinical outcome. Therapeutic interventions cannot be justified in the absence of clear cause-and-effect relations.

There remains a need to document clear clinical benefits from interventions designed to modify the inflammatory response. Modulation of the human inflammatory response has always been difficult, primarily as a result of our incomplete understanding of this response, and may lead to unexpected sequelae. The complexity of the inflammatory response is a significant obstacle to identification of the mechanism(s) by which alteration of a particular aspect of the response may affect clinical outcome. Indiscriminate inhibition– prevention of the inflammatory response to CPB may have detrimental effects, such as loss of appropriate wound healing and defenses against infection. Certain proinflammatory cytokines, such as IL-1β210and even endotoxin, 206appear to have poorly understood protective effects. In this regard, IL-1β pretreatment protects against subsequent myocardial ischemia and reperfusion injury. 210The emphasis of future strategies must therefore be on the control of, rather than the simple inhibition of, the inflammatory response to CPB.

Further research is required to determine why certain patients appear to be at increased risk of clinically important bypass-induced injury. Monitoring tools such as gastric tonometry and other indicators of systemic organ hyoperfusion in CPB patients have provided physiologic insights, but their role in therapy remains to be clarified. Preoperative or early perioperative identification of patients most likely to develop adverse clinical sequelae from SIRS would allow focusing of investigative, preventive, and therapeutic maneuvers, and should be a research priority.

The heterogeneity of the patient population undergoing CPB is well illustrated by the finding of a near 20-fold difference in mortality (1.8%vs.  34.3%) for first time elective versus  high-risk reoperative coronary artery bypass procedures. 5Many studies of therapeutic interventions have been carried out in relatively low-risk patients, in whom the risk of adverse postoperative events might be expected to be minimal. It is possible that beneficial effects of these therapeutic strategies may be diluted by studies that include low-risk patients. Therefore, future investigations of interventions to control the inflammatory response to CPB may be best focused on patients at high risk of postoperative organ dysfunction. Future studies should utilize direct clinical indicators of organ injury as outcome measures rather than simply measurement of mediator concentrations as surrogates of injury. Many of the previously conducted clinical trials do not contain adequate patient numbers to detect meaningful differences in outcome. There is a clear need for large-scale clinical trials of more promising strategies such as selective gut decontamination, maintenance of splanchnic perfusion, drug therapies including corticosteroid pretreatment, aprotinin, acadesine, and adjunctive measures, such as hemofiltration, and use of biocompatible materials in the extracorporeal circuit. These trials, restricted to patients at significant risk of perioperative morbidity, may have to be completed using multiple centers. Finally, more investigation is needed to determine the efficacy and safety of off-pump cardiac surgery, and new methods for myocardial protection in this context need to be developed.

Cardiac surgery evokes a generalized inflammatory response in all patients, with serious clinical consequences in a minority, despite advances in pharmacology, perfusion technology, cardiovascular monitoring, and anesthetic and surgical technique. This is most clearly evident with regard to postoperative pulmonary and cardiovascular dysfunction. The etiology of these events is probably a composite of unstable peribypass hemodynamics, global myocardial ischemia, suboptimal organ perfusion during CPB, and immune events related to exposure to extracorporeal circulation per se . A balanced, controlled inflammatory response is potentially beneficial, aiding host defenses against infection and facilitating wound healing, but loss of control of the inflammatory response may herald the onset of SIRS and single or multiple organ dysfunction.

The complex interaction of key proinflammatory and antiinflammatory components of the inflammatory response, the significance of alterations in the magnitude or time course of release, and their relation to the clinical sequelae seen following CPB remain to be fully elucidated. Further work to establish direct causal links between the inflammatory response to CPB and organ dysfunction may lead to mechanistic insights and in turn stimulate the development of successful preventive and therapeutic strategies. The increasing use of minimally invasive surgical techniques such as OPCAB will contribute to these insights, particularly the relative roles of CPB and peribypass events. The link between interventions and clinical benefit is gradually being established for several strategies, including improved circuit biocompatibility, maintenance of splanchnic perfusion, hemofiltration (in the pediatric population), aprotinin, and corticosteroid therapy. The need to better characterize and focus therapeutic interventions on the patient at risk of MODS following CPB is imperative.

The goal of modulation of the perioperative inflammatory response is to attenuate its deleterious effects while preserving the ability of the patient to mount an appropriate defense to the physiologic trespasses of the perioperative period. Although knowledge is growing about the role of altered immune function, the role of immunomodulatory therapies will remain investigational (especially in view of the failures of these therapies in recent sepsis trials) until the initiating events in postoperative SIRS become clearer.

The authors thank Wil Gamble, Medical Illustrator, Montreal, Quebec, Canada, for his expert art work in relation to figure 4.