1 Systemic Response to Injury and Metabolic Support

J. Martin Perez, Edward Lim, Steven E. Calvano, and Stephen F. Lowry

The inflammatory response to injury is designed to restore tissue function and eradicate invading microorganisms. Injuries of limited duration are usually followed by functional restoration with minimal intervention. By contrast, major insults to the host are associated with an overwhelming inflammatory response that, without appropriate and timely intervention, can lead to multiple-organ failure and adversely impact patient survival. Therefore, understanding how the inflammatory response is mobilized and controlled provides a functional framework on which interventions and therapeutics are formulated for the surgical patient.

This chapter addresses the hormonal, immunologic, and cellular responses to injury. Alterations of metabolism and nutrition in injury states are discussed in continuum because the utilization of fuel substrates during injury also is subject to the influences of hormonal and inflammatory mediators.

THE SYSTEMIC INFLAMMATORY RESPONSE SYNDROME

The systemic response to injury can be broadly compartmentalized into two phases: (1) a proinflammatory phase characterized by activation of cellular processes designed to restore tissue function and eradicate invading microorganisms, and (2) an antiinflammatory (counterregulatory phase) that is important for preventing excessive proinflammatory activities and restoring homeostasis in the individual (Table 1-1).

CENTRAL NERVOUS SYSTEM REGULATION OF INFLAMMATION

Reflex Inhibition of Inflammation

The central nervous system (CNS), via autonomic signaling, has an integral role in regulating the inflammatory response that is primarily involuntary. The autonomic system regulates heart rate, blood pressure, respiratory rate, gastrointestinal (GI) motility, and body temperature. The autonomic nervous system also regulates inflammation in a reflex manner, much like the patellar tendon reflex. The site of inflammation sends afferent signals to the hypothalamus, which in turn rapidly relays opposing antiinflammatory messages to reduce inflammatory mediator release by immunocytes.

Afferent Signals to the Brain

The CNS receives immunologic input from both the circulation and neural pathways. Areas of the CNS devoid of blood–brain barrier admit the passage of inflammatory mediators such as tumor necrosis factor (TNF)-α. Fevers, anorexia, and depression in illness are attributed to the humoral (circulatory) route of inflammatory signaling. Although the mechanism for vagal sensory input is not fully understood, it has been demonstrated that afferent stimuli

TABLE 1-1 Clinical Spectrum of Infection and Systemic Inflammatory Response Syndrome (SIRS)

Term Definition

Infection Identifiable source of microbial insult SIRS Two or more of following criteria

Temperature ≥38^◦Cor ≤36^◦C

Heart rate ≥90 beats/min

Respiratory rate ≥20 breaths/min or

Paco2 ≤32 mm Hg or mechanical

ventilation

White blood cell count ≥12,000/µLor ≤4000/µLor

≥10% band forms Sepsis Identifiable source of infection + SIRS Severe sepsis Sepsis + organ dysfunction Septic shock Sepsis + cardiovascular collapse (requiring

vasopressor support)

to the vagus nerve include cytokines (e.g., TNF-α and interleukin [IL]-1), baroreceptors, chemoreceptors, and thermoreceptors originating from the site of injury.

Cholinergic Antiinflammatory Pathways

Acetylcholine, the primary neurotransmitter of the parasympathetic system, reduces tissue macrophage activation. Furthermore, cholinergic stimulation directly reduces tissue macrophage release of the proinflammatory mediators TNF-α, IL-1, IL-18, and high mobility group protein (HMG-1), but not the antiinflammatory cytokine IL-10. The attenuated inflammatory response induced by cholinergic stimuli was further validated by the identification of acetylcholine (nicotinic) receptors on tissue macrophages.

In summary, vagal stimulation reduces heart rate, increases gut motility, dilates arterioles, and causes pupil constriction, and regulates inflammation. Unlike the humoral antiinflammatory mediators, signals discharged from the vagus nerve are targeted at the site of injury or infection. Moreover, this cholinergic signaling occurs rapidly in real time.

HORMONAL RESPONSE TO INJURY

Hormone Signaling Pathways

Hormones are chemically classified as polypeptides (e.g., cytokines, glucagon, and insulin), amino acids (e.g., epinephrine, serotonin, and histamine), or fatty acids (e.g., glucocorticoids, prostaglandins, and leukotrienes [LT]). Most hormone receptors generate signals by one of three major overlapping pathways:

(1)

receptor kinases such as insulin and insulin-like growth factor receptors,

(2)

guanine nucleotide-binding or G-protein receptors such as neurotransmitter and prostaglandin receptors, (3) ligand-gated ion channels, which permit ion transport when activated. Membrane receptor activation leads to amplification via secondary signaling pathways. Hormone signals are further mediated by intracellular receptors with binding affinities for both the hormone itself, and for the targeted gene sequence on the deoxyribonucleic acid (DNA). The classic example of a cytosolic hormonal receptor is the glucocorticoid (GC) receptor.

TABLE 1-2 Hormones Regulated by the Hypothalamus, Pituitary, and Autonomic System

Hypothalamic Regulation

Corticotropin-releasing hormone Thyrotropin-releasing hormone Growth hormone-releasing hormone Luteinizing hormone-releasing hormone

Anterior Pituitary Regulation

Adrenocorticotropic hormone Cortisol Thyroid-stimulating hormone Thyroxine Triiodothyronine Growth hormone Gonadotrophins Sex hormones Insulin-like growth factor Somatostatin Prolactin Endorphins

Posterior Pituitary Regulation

Vasopressin Oxytocin

Autonomic System

Norepinephrine Epinephrine Aldosterone Renin-angiotensin system Insulin Glucagon Enkephalins

Hormones of the hypothalamic-pituitary-adrenal (HPA) axis influences the physiologic response to injury and stress (Table 1-2), but some with direct influence on the inflammatory response or immediate clinical impact will be highlighted.

Adrenocorticotropic Hormone

Adrenocorticotropic hormone (ACTH) is synthesized and released by the anterior pituitary. In healthy humans, ACTH release is regulated by circadian signals with high levels of ACTH occurring late at night until the hours immediately before sunrise. During injury, this pattern is dramatically altered. Elevations in corticotropin-releasing hormone and ACTH are typically proportional to the severity of injury. Pain, anxiety, vasopressin, angiotensin II, cholecystokinin, vasoactive intestinal polypeptide (VIP), catecholamines, and proinflammatory cytokines are all prominent mediators of ACTH release in the injured patient.

Cortisol and Glucocorticoids

Cortisol is the major glucocorticoid in humans and is essential for survival during significant physiologic stress. Following injury, the degree of cortisol elevation is dependent on the degree of systemic stress. For example, burn patients have elevated circulating cortisol levels for up to 4 weeks, whereas lesser injuries may exhibit shorter periods of cortisol elevation.

Cortisol potentiates the actions of glucagon and epinephrine that manifest as hyperglycemia. Cortisol stimulates gluconeogenesis, but induces insulin resistance in muscles and adipose tissue. In skeletal muscle, cortisol induces protein degradation and the release of lactate that serve as substrates for hepatic gluconeogenesis. During injury, cortisol potentiates the release of free fatty acids, triglycerides, and glycerol from adipose tissue providing additional energy sources.

Acute adrenal insufficiency (AAI) secondary to exogenous glucocorticoid administration can be a life-threatening complication most commonly seen in acutely ill patients. These patients present with weakness, nausea, vomiting, fever, and hypotension. Objective findings include hypoglycemia from decreased gluconeogenesis, hyponatremia, and hyperkalemia. Insufficient mineralocorticoid (aldosterone) activity also contributes to hyponatremia and hyperkalemia.

Glucocorticoids have long been employed as immunosuppressive agents. Immunologic changes associated with glucocorticoid administration include thymic involution, depressed cell-mediated immune responses reflected by decreases in T-killer and natural killer cell functions, T-lymphocyte blastogenesis, mixed lymphocyte responsiveness, graft-versus-host reactions, and delayed hypersensitivity responses. With glucocorticoid administration, monocytes lose the capacity for intracellular killing but appear to maintain normal chemotactic and phagocytic properties. For neutrophils, glucocorticoids inhibit intracellular superoxide reactivity, suppress chemotaxis, and normalize apoptosis signaling mechanisms. However, neutrophil phagocytosis function remains unchanged. Clinically, glucocorticoids has been associated with modest reductions in proinflammatory response in septic shock, surgical trauma, and coronary artery bypass surgery. However, the appropriate dosing, timing, and duration of glucocorticoid administration have not been validated.

Macrophage Inhibitory Factor

Macrophage inhibitory factor (MIF) is a glucocorticoid antagonist produced by the anterior pituitary that potentially reverses the immunosuppressive effects of glucocorticoids. MIF can be secreted systemically from the anterior pituitary and by T lymphocytes situated at the sites of inflammation. MIF is a proinflammatory mediator that potentiates gram-negative and gram-positive septic shock.

Growth Hormones and Insulin-Like Growth Factors

During periods of stress, growth hormone (GH), mediated in part by the secondary release of insulin-like growth factor-1 (IGF-1), promotes protein synthesis and enhances the mobilization of fat stores. IGF, formerly called somatomedin C, circulates predominantly in bound form and promotes amino acid incorporation, cellular proliferation, skeletal growth, and attenuates proteolysis. In the liver, IGFs are mediators of protein synthesis and glycogenesis. In adipose tissue, IGF increases glucose uptake and fat utilization. In skeletal muscles, IGF increases glucose uptake and protein synthesis. The decrease in protein synthesis and observed negative nitrogen balance following injury is attributed in part to a reduction in IGF-1 levels. GH administration has improved the clinical course of pediatric burn patients. Its use in injured adult patients remains unproven.

Catecholamines

The hypermetabolic state observed following severe injury is attributed to activation of the adrenergic system. Norepinephrine (NE) and epinephrine (EPI) are increased 3-to 4-fold in plasma immediately following injury, with elevations lasting 24–48 hours before returning toward baseline levels.

In the liver, EPI promotes glycogenolysis, gluconeogenesis, lipolysis, and ketogenesis. It also causes decreased insulin release, but increases glucagon secretion. Peripherally, EPI increases lipolysis in adipose tissues and induces insulin resistance in skeletal muscle. These collectively manifest as stress-induced hyperglycemia, not unlike the effects of cortisol on blood sugar.

Like cortisol, EPI enhances leukocyte demargination with resultant neutrophilia and lymphocytosis. However, EPI occupation of β receptors present on leukocytes ultimately decreases lymphocyte responsiveness to mitogens.

In noncardiac surgical patients with heart disease, perioperative β-receptor blockade also reduced sympathetic activation and cardiac oxygen demand with significant reductions in cardiac-related deaths.

Aldosterone

The mineralocorticoid aldosterone is synthesized, stored, and released, via ACTH stimulation, in the adrenal zona glomerulosa. The major function of aldosterone is to maintain intravascular volume by conserving sodium and eliminating potassium and hydrogen ions in the early distal convoluted tubules of the nephrons.

Patients with aldosterone deficiency develop hypotension and hyperkalemia, whereas patients with aldosterone excess develop edema, hypertension, hypokalemia, and metabolic alkalosis.

Insulin

Hormones and inflammatory mediators associated with stress response inhibit insulin release. In conjunction with peripheral insulin resistance following injury, this results in stress-induced hyperglycemia and is in keeping with the general catabolic state immediately following major injury.

In the healthy individual, insulin exerts a global anabolic effect by promoting hepatic glycogenesis and glycolysis, glucose transport into cells, adipose tissue lipogenesis, and protein synthesis. During injury, insulin release is initially suppressed followed by normal or excessive insulin production despite hyperglycemia.

Activated lymphocytes express insulin receptors, and activation enhances T-cell proliferation and cytotoxicity. Tight control of glucose levels in the critically ill has been associated with significant reductions in morbidity and mortality.

Acute Phase Proteins

The acute phase proteins are nonspecific biochemical markers produced by hepatocytes in response to tissue injury, infection, or inflammation. IL-6 is a potent inducer of acute phase proteins that can include proteinase inhibitors, coagulation and complement proteins, and transport proteins. Only C-reactive protein (CRP) has been consistently used as a marker of injury response because of its dynamic reflection of inflammation. The accuracy of CRP appears to surpass that of the erythrocyte sedimentation rate.

MEDIATORS OF INFLAMMATION

Cytokines

Cytokines are the most potent mediators of the inflammatory response. When functioning locally at the site of injury or infection, cytokines eradicate invading microorganisms and promote wound healing. Overwhelming production of proinflammatory cytokines in response to injury can cause hemodynamic instability (e.g., septic shock) or metabolic derangements (e.g., muscle wasting). If uncontrolled, the outcome of these exaggerated responses is end-organ failure and death. The production of antiinflammatory cytokines serves to oppose the actions of proinflammatory cytokines. To view cytokines merely as proinflammatory or antiinflammatory oversimplifies their functions, and overlapping bioactivity is the rule (Table 1-3).

Heat Shock Proteins

Hypoxia, trauma, heavy metals, local trauma, and hemorrhage all induce the production of intracellular heat shock proteins (HSPs). HSPs are intracellular protein modifiers and transporters that are presumed to protect cells from the deleterious effects of traumatic stress. The formation of HSPs requires gene induction by the heat shock transcription factor.

Reactive Oxygen Metabolites

Reactive oxygen metabolites are short-lived, highly reactive molecular oxygen species with an unpaired outer orbit. Tissue injury is caused by oxidation of unsaturated fatty acids within cell membranes. Activated leukocytes are potent generators of reactive oxygen metabolites. Furthermore, ischemia with reperfusion also generates reactive oxygen metabolites.

Oxygen radicals are produced by complex processes that involve anaerobic glucose oxidation coupled with the reduction of oxygen to superoxide anion. Superoxide anion is an oxygen metabolite that is further metabolized to other reactive species such as hydrogen peroxide and hydroxyl radicals. Cells are generally protected by oxygen scavengers that include glutathione and catalases.

Eicosanoids

The eicosanoid class of mediators, which encompasses prostaglandins (PGs), thromboxanes (TXs), LTs, hydroxy-icosatetraenoic acids (HETEs), and lipoxins (LXs), are oxidation derivatives of the membrane phospholipid arachidonic acid (eicosatetraenoic acid). Eicosanoids are secreted by virtually all nucleated cells except lymphocytes. Products of the cyclooxygenase pathway include all of the prostaglandins and thromboxanes. The lipoxygenase pathway generates the LT and HETE.

Eicosanoids are synthesized rapidly on stimulation by hypoxic injury, direct tissue injury, endotoxin, NE, vasopressin, angiotensin II, bradykinin, serotonin, acetylcholine, cytokines, and histamine. COX-2, a second cyclooxygenase enzyme, converts arachidonate to prostaglandin E2 (PGE2). PGE2 increases fluid leakage from blood vessels, but a rising PGE2 level over several hours eventually feeds back to COX-2 and induces the formation of the antiinflammatory lipoxin from neutrophils. Nonsteroidal antiinflammatory drugs (NSAIDs) reduce the PGE2 

levels by COX-2 acetylation and increases lipoxin production. COX-2 activity also can be inhibited by glucocorticoids.

Cytokine Source Comment

TABLE 1-3 Cytokines and Their Sources

TNF-α	Macrophages/monocytes Kupffer cells Neutrophils NK cells Astrocytes Endothelial cells	Among earliest responders following injury; half-life <20 min; activates TNF-receptor-1 and -2; induces significant shock and catabolism
	T lymphocytes Adrenal cortical cells
	Adipocytes Keratinocytes Osteoblasts
	Mast cells
	Dendritic cells
IL-1	Macrophages/monocytes B and T lymphocytes NK cells Endothelial cells Epithelial cells Keratinocytes Fibroblasts	Two forms (IL-α and IB-β); similar physiologic effects as TNF-α; induces fevers through prostaglandin activity in anterior hypothalamus; promotes β-endorphin release from pituitary; half-life <6 min
	Osteoblasts
	Dendritic cells
	Astrocytes Adrenal cortical cells
	Megakaryocytes Platelets
	Neutrophils Neuronal cells
IL-2	T lymphocytes	Promotes lymphocyte proliferation, immunoglobulin production, gut barrier integrity; half-life <10 min; attenuated production following major blood loss leads to immunocompromise; regulates lymphocyte apoptosis
IL-3	T lymphocytes Macrophages Eosinophils Mast cells
IL-4	T lymphocytes Mast cells Basophils Macrophages B lymphocytes Eosinophils Stromal cells	Induces B-lymphocyte production of IgG4 and IgE, mediators of allergic and anthelmintic response; downregulates TNF-α, IL-1, IL-6, IL-8
IL-5	T lymphocytes Eosinophils Mast cells	Promotes eosinophil proliferation and airway inflammation
	Basophils
IL-6	Macrophages B lymphocytes Neutrophils Basophils Mast cells	Elicited by virtually all immunogenic cells; long half-life; circulating levels proportional to injury severity; prolongs activated neutrophil survival
	Fibroblasts
		(continued)

Endothelial cells Astrocytes Synovial cells Adipocytes Osteoblasts Megakaryocytes Chromaffin cells Keratinocytes

IL-8 Macrophages/monocytes T lymphocytes Basophils Mast cells Epithelial cells Platelets

IL-10 T lymphocytes B lymphocytes Macrophages Basophils Mast cells Keratinocytes

IL-12 Macrophages/monocytes Neutrophils Keratinocytes Dendritic cells B lymphocytes

IL-13 T lymphocytes

IL-15 Macrophages/monocytes Epithelial cells

IL-18 Macrophages Kupffer cells Keratinocytes Adrenal cortical cells Osteoblasts

IFN-γ T lymphocytes NK cells Macrophages

GM-CSF T lymphocytes Fibroblasts Endothelial cells Stromal cells

IL-21 T lymphocytes

HMGB-I Monocytes/lymphocytes

Chemoattractant for neutrophils, basophils, eosinophils, lymphocytes

Prominent anti-inflammatory cytokine; reduces mortality in animal sepsis and ARDS models

Promotes TH1 differentiation; synergistic activity with IL-2

Promotes B-lymphocyte function; structurally similar to IL-4; inhibits nitric oxide and endothelial activation

Anti-inflammatory effect; promotes lymphocyte activation; promotes neutrophil phagocytosis in fungal infections

Similar to IL-12 in function; elevated in sepsis, particularly gram-positive infections; high levels found in cardiac deaths

Mediates IL-12 and IL-18 function; half-life, days; found in wounds 5–7 days after injury; promotes ARDS

Promotes wound healing and inflammation through activation of leukocytes

Preferentially secreted by TH2 cells; structurally similar to IL-2 and IL-15; activates NK cells, B and T lymphocytes; influences adaptive immunity

High mobility group box chromosomal protein; DNA transcription factor; late (downstream) mediator of inflammation (ARDS, gut barrier disruption); induces “sickness behavior”

ARDS = acute respiratory distress syndrome; GM-CSF = granulocytemacrophage colony-stimulating factor; IFN = interferon; IgE = immunoglobulin E; IgG = immunoglobulin G; IL = interleukin; NK = natural killer; TH1 = T helper subset cell 1; TH2 = T helper subset cell 2; TNF = tumor necrosis factor.

Eicosanoids have diverse effects systemically on endocrine and immune function, neurotransmission, and vasomotor regulation Eicosanoids are implicated in acute lung injury, pancreatitis, and renal failure. LT are effective promoters of leukocyte adherence, neutrophil activation, bronchoconstriction, and vasoconstriction.

Eicosanoids are involved in the regulation of glucose, with the products of the cyclooxygenase pathway inhibiting pancreatic β-cell release of insulin, whereas products of the lipoxygenase pathway promote β-cell activity. Hepatocyte PGE2 receptors, when activated, inhibit gluconeogenesis. PGE2 also can inhibit hormone-stimulated lipolysis.

Fatty Acid Metabolites

Fatty acid metabolism potentially has a role in the inflammatory response. Omega-6 fatty acids, most commonly the primary lipid source in enteral nutrition formulas, also serve as precursors of inflammatory mediators, such as the eicosanoids, and are associated with injury and the stress response. Animal studies substituting omega-3 for omega-6 fatty acids have demonstrated attenuated inflammatory response in hepatic Kupffer cells as measured by TNF and IL-1 release and PGE2 production.

Kallikrein-Kinin System

Bradykinins are potent vasodilators that are produced through kininogen degradation by the serine protease kallikrein. Kinins increase capillary permeability and tissue edema, evoke pain, inhibit gluconeogenesis, and increase bronchoconstriction. An increase in renin secondary to reduced renal perfusion promotes sodium and water retention via the renin-angiotensin system.

Increased bradykinin levels are observed following hypoxia, reperfusion, hemorrhage, sepsis, endotoxemia, and tissue injury. These elevations are proportional to the magnitude of injury and mortality. Clinically, bradykinin antagonists in septic shock studies have only demonstrated modest reversal in gram-negative sepsis, but no overall improvement in survival.

Serotonin

The neurotransmitter serotonin (5-hydroxytryptamine, 5-HT) is a tryptophan derivative that is found in chromaffin cells of the intestine and in platelets. Serotonin stimulates vasoconstriction, bronchoconstriction, and platelet aggregation. Although serotonin is clearly released at sites of injury, its role in the inflammatory response is unclear.

Histamine

Histamine is derived from histidine and stored in neurons, skin, gastric mucosa, mast cells, basophils, and platelets. There are two receptor types (H1 and H2) for histamine binding. H1 receptor binding stimulates bronchoconstriction, intestinal motility, and myocardial contractility. H2 receptor binding inhibits histamine release. Both H1 and H2 receptor activation induce hypotension, peripheral pooling of blood, increased capillary permeability, decreased venous return, and myocardial failure. The rise in histamine levels has been documented in hemorrhagic shock, trauma, thermal injury, endotoxemia, and sepsis.

CYTOKINE RESPONSE TO INJURY

Tumor Necrosis Factor

TNF-α is among the earliest and most potent mediators of the inflammatory host responses. TNF-α synthesis occurs in monocytes/macrophages and T cells, which are abundant in the peritoneum, splanchnic tissues, and liver (Kupffer cells). Although the half-life of TNF-α is less than 20 minutes, this is sufficient to evoke marked muscle catabolism and cachexia during stress, hemodynamic changes, and activate mediators distally in the cytokine cascade. TNF-α also promotes coagulation activation, the expression or release of adhesion molecules, prostaglandin E2, platelet-activating factor (PAF), glucocorticoids, and eicosanoids.

Soluble TNF receptors (sTNFRs) are proteolytically cleaved extracellular domains of membrane-associated TNFRs that are elevated and readily detectable in acute inflammation. sTNFRs retain their affinity for the binding of TNF-α and therefore compete with the cellular receptors for the binding of free TNF-α. This may represent a counterregulatory response to excessive systemic TNF-α activity or serve as a carrier of bioactive TNF-α.

Interleukin-1

IL-1 (IL-1α and IL-1β) is primarily released by activated macrophages and endothelial cells. IL-1α is predominantly cell membrane–associated and exerts its influence via cellular contacts. IL-1β is more readily detectable in the circulation and is similar in its effects to TNF-α. High doses of either IL-1 or TNF-α initiate a state of hemodynamic decompensation. At low doses, they can produce the same response only if administered simultaneously. TNF-α and IL-1 appear to have synergy in the inflammatory response. IL-1 is predominantly a local mediator with a half-life of approximately 6 minutes. IL-1 induces the febrile response to injury by stimulating local prostaglandin activity in the anterior hypothalamus.

Endogenous IL-1 receptor antagonists (IL-1ra) also are released during injury and serve as an endogenous auto-regulator of IL-1 activity.

Interleukin-2

IL-2 is a primary promoter of T-lymphocyte proliferation, immunoglobulin production, and gut barrier integrity. Partly because of its circulation half-life of less than 10 minutes, IL-2 has not been readily detectable following acute injury. Attenuated IL-2 expression associated with major injuries or perioperative blood transfusions potentially contribute to the transient immunocompromised state of the surgical patient.

Interleukin-4

IL-4 is produced by activated type 2 T-helper (TH2) lymphocytes and is particularly important in antibody-mediated immunity and in antigen presentation. IL-4 also induces class switching in differentiating B lymphocytes to produce predominantly IgG4 and IgE, which are important immunoglobulins in allergic and anthelmintic responses. IL-4 has potent antiinflammatory properties against activated macrophages by downregulating the effects of IL-1, TNF-α, IL-6, and IL-8, and oxygen radical production. IL-4 also appears to increase macrophage susceptibility to the antiinflammatory effects of glucocorticoids.

Interleukin-6

TNF-α and IL-1 are potent inducers of IL-6 production from virtually all cells and tissues, including the gut. Circulating IL-6 levels appear to be proportional to the extent of tissue injury during an operation, more so than the duration of the surgical procedure itself. Recent evidence has demonstrated both a proinflammatory role and an antiinflammatory role for IL-6. IL-6 is an important mediator of the hepatic acute phase response during injury and convalescence, induces neutrophils activation, and paradoxically delays the disposal of activated neutrophils.

Interleukin-8

IL-8 is a chemoattractant and a potent activator of neutrophils. Expression and activity is similar to that of IL-6 after injury and has been proposed as an additional biomarker for the risk of multiple-organ failure. IL-8 does not produce the hemodynamic instability characteristic of TNF-α and IL-1.

Interleukin-10

IL-10 has emerged as a modulator of TNF-α activity. Experimental evidence has demonstrated that neutralization of IL-10 during endotoxemia increases monocyte TNF-α production and mortality, but restitution of IL-10 reduces TNF-α levels and the associated deleterious effects.

Interferon-γ

Human T-helper lymphocytes activated by bacterial antigens and interleukins readily produce IFN-γ . When released into the circulation, IFN-γ is detectable in vivo by 6 hours and may be persistently elevated for as long as 8 days. Injured tissues, such as operative wounds, also demonstrate the presence of IFN-γ production 5–7 days after injury. IFN-γ has important roles in activating circulating and tissue macrophages. Alveolar macrophage activation mediated by IFN-γ may induce acute lung inflammation after major surgery or trauma.

Granulocyte-Macrophage Colony-Stimulating Factor

Granulocyte-macrophage colony-stimulating factor (GM-CSF) delays apoptosis (programmed cell death) of macrophages and neutrophils. This growth factor is effective in promoting the maturation and recruitment of functional leukocytes necessary for normal inflammatory cytokine response, and potentially in wound healing. The delay in apoptosis may contribute to organ injury such as that found in acute respiratory distress syndrome (ARDS).

High Mobility Group Box-1

DNA transcription factor, HMGB-1, is released 24–48 hours after the onset of sepsis. The appearance of HMGB-1 is contrast to the early appearances of TNF-α, IL-1, IL-6, and IL-8. Clinically, HMGB-1 peak levels are associated with ARDS and mortality. As a late mediator of the inflammatory response, anti-HMGB-1 strategies may be utilized to modify the inflammatory response.

CELLULAR RESPONSE TO INJURY

Gene Expression and Regulation

In the inflammatory response, cytokine production involves rapid ribonucleic acid (RNA) transcription and translation or protein synthesis. These proteins can further be modified in the cytosol for specific functions. In essence, these cytosolic modifications supplement the primary regulatory mechanisms within the nucleus.

How a particular gene is activated depends on the orderly assemblage of transcription factors to specific DNA sequences immediately upstream to the target gene, known as the promoter region. The DNA binding sites are the enhancer sequences, and proteins that inhibit the initiation of transcription are repressors. Transcription factors become important during the inflammatory response because the ability to control the pathways leading to their activation means the ability to regulate the manner and magnitude by which a cell can respond to an injury stimulus.

Cell Signaling Pathways

Heat Shock Proteins

HSPs, also known as stress proteins, are produced by cells in response to injury or tissue ischemia. HSPs are essential for the ability of cells to overcome stress. HSPs’ primary role are to attenuate the inflammatory response by reducing oxygen metabolites, promoting TH2 cell proliferation, and inhibiting nuclear factor (NF)-κB activation.

G-Protein Receptors

Guanosine-5-triphosphate (GTP)-binding proteins (G-proteins), with activation of an adjacent effector protein, are the largest family of signaling receptors for cells and include many of the pathways associated with the inflammatory response. The two major second messengers of the G-protein pathway are

(a) formation of cyclic adenosine monophosphate (cAMP), and (b) calcium, released from the endoplasmic reticulum. An increase in cellular cAMP can activate gene transcription.

G-protein/calcium activation requires activation of the effector phospholipase C and phosphoinositols. G-protein signaling activates protein kinase C (PKC), which can in turn activate NF-κB and other transcription factors.

Ligand-Gated Ion Channels

These receptor channels, when activated by a ligand, permit rapid flux of ions across the cell membrane. Neurotransmitters function by this pathway, and an example of such a receptor is the nicotinic acetylcholine receptor.

Receptor Tyrosine Kinases

Receptor tyrosine kinases have significant intracellular tyrosine kinase domains. Examples of these receptors include insulin and various hormone growth factors (e.g., platelet-derived growth factor [PDGF], IGF-1, epidermal growth factor [EGF], and vascular endothelial growth factor [VEGF]).

Activation of protein kinase receptors is important for gene transcription and cell proliferation.

Janus Kinase/Signal Transduction and Activator of Transcription (STAT) Signaling

Janus kinase (JAK) is the receptor for many cytokines. Ligand receptor interaction leads to receptor dimerization and enzymatic activation results in propagation through the JAK domains of the receptors. JAK mediated signaling via STAT mediated transcription can activate different T-cell responses during injury and inflammation.

Suppressors of Cytokine Signaling

Suppressors of cytokine signaling (SOCS) specifically block JAK and STAT activation and regulate the signaling of certain cytokines. A deficiency of SOCS activity may render a cell hypersensitive to certain stimuli such as inflammatory cytokines and growth hormones.

Mitogen-Activated Protein Kinases

The mitogen-activated protein kinase (MAPK) pathway is a major cellular inflammatory signaling pathway with regulatory roles over cell proliferation and cell death. The three major isoforms are the JNK (c-Jun NH2-terminal kinase), ERK (extracellular regulatory protein kinase), and p38 kinase. The JNK pathway has clear links, via TNF-α and IL-1, to the inflammatory response with a regulatory role in apoptosis. The p38 kinase is activated in response to endotoxin, viruses, interleukins, TNF-α, and transforming growth factor (TGF)-β. P38-kinase activation triggers the recruitment and activation of leukocytes. MAPK isoforms exhibit appreciable “cross-talk,” which can modulate the inflammatory response.

Nuclear Factor-κB

NF-κB activates genes important for the activation of proinflammatory cytokines and acute phase proteins. NF-κB is really a complex of smaller proteins, and the p50-p65 heterodimer complex is the most widely studied. Cytosolic NF-κB is maintained by binding to the inhibitor protein I-κB. When a cell is exposed to an inflammatory stimulus (TNF-α or IL-β), a series of phosphorylation events leads to I-κB degradation. Low intracellular I-κB appears to prolong the inflammatory response and enhanced activity of NF-κB appears to delay the apoptosis of activated immune cells.

Toll-Like Receptors and CD14

Lipopolysaccharide (LPS), an endotoxin, is an important mediator of gramnegative sepsis syndrome. LPS recognition and activation of the inflammatory response by immune cells occurs primarily by the toll-like receptor-4 (TLR4) mechanism. LPS-binding proteins (LBPs) carry LPS to the CD14/TLR4 complex, which sets into motion cellular mechanisms that activate MAPK, NF-κB, and cytokine gene promoters. TLR4 is primarily the receptor for gram-negative endotoxins and TLR2 is the counterpart for gram-positive sepsis. The fact that some patient populations are more susceptible to infectious complications than others recently has been associated with specific point mutations in the TLR gene.

Tumor Necrosis Factor and CD95-Induced Apoptosis

Apoptosis is the principal mechanism by which senescent or dysfunctional cells, including macrophages and neutrophils, are systematically disposed of without activating other immunocytes or inducing an inflammatory response. The cellular environment created by systemic inflammation disrupts the normal apoptotic machinery in activated immunocytes, consequently prolonging the inflammatory response.

Several proinflammatory cytokines (e.g., TNF-α, IL-1, IL-3, IL-6, GM-CSF, granulocyte colony-stimulating factor [G-CSF], and IFN-γ ) and bacterial products (e.g., endotoxin) have been shown to delay macrophage and neutrophil apoptosis in vitro, whereas IL-4 and IL-10 accelerate apoptosis in activated monocytes.

In acute inflammation, the response of the immunocyte to TNF-α is perhaps the most widely investigated. This cytokine exerts its biologic effects by binding to specific cellular receptors, tumor necrosis factor receptor (TNFR)-1 (55 kDa) and TNFR-2 (75 kDa). When TNFR-1 is exclusively activated, it precipitates circulatory shock reminiscent of severe sepsis. However, exclusive activation of TNFR-2 fails to induce any inflammatory responses or shock.

The activation of NF-κB and JNK is believed to be the major antiapoptotic, and therefore proinflammatory, factor; it is signal induced by TNFR-1 and TNFR-2. It is well known that TNF-α–induced NF-κB activation delays cell death and is associated with the activation of diverse genes that include proinflammatory mediators. Exaggerated peripheral blood monocyte NF-κB activation has been associated with higher mortality rates in patients with septic shock.

The CD95 (Fas) receptor shares much of its intracellular structure with TNFR-1. Unlike TNFR-1, the only known function of CD95 is to initiate programmed cell death. Neutrophils and macrophages express CD95, and this expression may have important implications in the cellular contribution to the inflammatory response. In fact, both clinical sepsis and experimental endotoxemia have demonstrated prolonged survival of neutrophils and diminished responsiveness to CD95 stimuli.

Cell-Mediated Inflammatory Response

Platelets

Clot formation at the site of injury releases inflammatory mediators and serves as the principal chemoattractant for neutrophils and monocytes. The migration of platelets and neutrophils through the vascular endothelium occurs within 3 hours of injury and is mediated by serotonin release, platelet-activating factor, and prostaglandin E2. Platelets can enhance or reduce neutrophil-mediated tissue injury by modulating neutrophil adherence to the endothelium and subsequent respiratory burst. Platelets are an important source of eicosanoids and vasoactive mediators. NSAIDs irreversibly inhibit thromboxane production.

Lymphocytes and T-Cell Immunity

Injury, surgical or traumatic, is associated with acute impairment of cellmediated immunity and macrophage function. T-helper lymphocytes are functionally divided into two subgroups, referred to as TH1 and TH2. In severe infections and injury, there appears to be a reduction in TH1 (cell-mediated immunity) cytokine production, with a lymphocyte population shift toward the TH2 response and its associated immunosuppressive effects. In patients with major burns, a shift to a TH2 cytokine response has been a predictor of infectious complications. However, studies in patients undergoing major surgery have demonstrated a postoperative reduction in TH1 cytokine production that is not necessarily associated with increased TH2 response. Nevertheless, depressed TH1 response and systemic immunosuppression following major insults to the host may be a useful paradigm in predicting the subset of patients who are prone to infectious complications and poor outcome.

Eosinophils

Eosinophils are characteristically similar to neutrophils in that they migrate to inflamed endothelium and release cytoplasmic granules that are cytotoxic. Eosinophils preferentially migrate to sites of parasitic infection and allergen challenge. Major activators of eosinophils include IL-3, GM-CSF, IL-5, platelet-activating factor, and complement anaphylatoxins C3a and C5a.

Mast Cells

Tissue resident mast cells are important as first-responders to sites of injury. Activated mast cells produce histamine, cytokines, eicosanoids, proteases, and chemokines. The immediate results are vasodilation, recruitment of other immunocytes, and capillary leakage. TNF-α is secreted rapidly by mast cells because of the abundant stores within granules. Mast cells also can synthesize a variety of cytokines and migration-inhibitory factor (MIF).

Monocytes

In clinical sepsis, nonsurviving patients with severe sepsis have an immediate reduction in monocyte surface TNFR expression with failure to recover, whereas surviving patients have normal or near-normal receptor levels from the onset of clinically defined sepsis.

Thus, TNFR expression potentially can be used as a prognostic indicator of outcome in patients with systemic inflammation. There is also decreased CD95 expression following experimental endotoxemia in humans, which correlates with diminished CD95-mediated apoptosis. Taken together, the reduced receptor expression and delayed apoptosis may be a mechanism for prolonging the inflammatory response during injury or infection.

Neutrophils

Neutrophils mediate important functions in every form of acute inflammation, including acute lung injury, ischemia/reperfusion injury, and inflammatory bowel disease. G-CSF is the primary stimulus for neutrophil maturation. Inflammatory mediators from a site of injury induce neutrophil adherence to the vascular endothelium and promote eventual cell migration into the injured tissue. Neutrophil function is mediated by a vast array of intracellular granules that are chemotactic or cytotoxic to local tissue and invading microorganisms.

ENDOTHELIUM-MEDIATED INJURY

Neutrophil-Endothelium Interaction

Increased vascular permeability during inflammation is intended to facilitate oxygen delivery and immunocyte migration to the sites of injury. However, neutrophils migration and activation at sites of injury may contribute to the cytotoxicity of vital tissues and result in organ dysfunction. Ischemia/reperfusion (I/R) injury potentiates this response by unleashing oxygen metabolites, lysosomal enzymes that degrade tissue basal membranes, cause microvascular thrombosis, and activate myeloperoxidases. The recruitment of circulating neutrophils to endothelial surfaces is mediated by concerted actions of adhesion molecules referred to as selectins that are elaborated on cell surfaces.

Nitric Oxide

Nitric oxide (NO) is derived from endothelial surfaces in response to acetylcholine stimulation, hypoxia, endotoxin, cellular injury, or mechanical shear stress from circulating blood. NO promotes vascular smooth muscle relaxation, reduces microthrombosis, and mediates protein synthesis in hepatocytes. NO is formed from oxidation of L-arginine, a process catalyzed by nitric oxide synthase (NOS). NO is a readily diffusible substance with a half-life of a few seconds. NO spontaneously decomposes into nitrate and nitrite.

In addition to the endothelium, NO formation also occurs in neutrophils, monocytes, renal cells, Kupffer cells, and cerebellar neurons.

Prostacyclin

Although it is an arachidonate product, prostacyclin (PGI2) is another important endothelium-derived vasodilator synthesized in response to vascular shear stress and hypoxia. PGI2 shares similar functions with NO, inducing vasorelaxation and platelet deactivation by increasing cAMP. Clinically, it has been used to reduce pulmonary hypertension, particularly in the pediatric population.

Endothelins

Endothelins (ETs) are elaborated by vascular endothelial cells in response to injury, thrombin, transforming growth factor-β (TGF-β), IL-1, angiotensin II, vasopressin, catecholamines, and anoxia. ET is a 21-amino-acid peptide with potent vasoconstricting properties. Of the peptides in this family (e.g., ET-1, ET-2, and ET-3), endothelial cells appear to exclusively produce ET-1. ET-1 appears to be the most biologically active and the most potent known vasoconstrictor. It is estimated to be 10 times more potent than angiotensin II. The maintenance of physiologic tone in vascular smooth muscle depends on the balance between NO and ET production. Increased serum levels of ETs correlate with the severity of injury following major trauma, major surgical procedures, and in cardiogenic or septic shock.

Platelet-Activating Factor

Platelet-activating factor (PAF), an endothelial-derived product, is a natural phospholipid constituent of cell membranes, which under normal physiologic conditions is minimally expressed. During acute inflammation, PAF is released by neutrophils, platelets, mast cells, and monocytes, and is expressed at the outer leaflet of endothelial cells. PAF can further activate neutrophils and platelets and increase vascular permeability. Human sepsis is associated with a reduction in PAF-acetylhydrolase levels, which is the endogenous inactivator of PAF. Indeed, PAF-acetylhydrolase administration in patients with severe sepsis has shown some reduction in multiple organ dysfunction and mortality.

Atrial Natriuretic Peptides

Atrial natriuretic peptides (ANPs) are a family of peptides released primarily by atrial tissue, but are also synthesized by the gut, kidney, brain, adrenal glands, and endothelium. They induce vasodilation and fluid and electrolyte excretion. ANPs are potent inhibitors of aldosterone secretion and prevent reabsorption of sodium.

SURGICAL METABOLISM

The initial hours following surgical or traumatic injury are metabolically associated with a reduced total body energy expenditure and urinary nitrogen wasting. Following resuscitation and stabilization of the injured patient, a reprioritization of substrate utilization ensues to preserve vital organ function and for the repair of injured tissue. This phase of recovery also is characterized by augmented metabolic rates and oxygen consumption, enzymatic preference for readily oxidizable substrates such as glucose, and stimulation of the immune system.

Understanding the collective alterations, as a consequence of the inflammatory response, in amino acid (protein), carbohydrate, and lipid metabolism lays the foundation on which clinical metabolic and nutritional support can be implemented.

Metabolism Following Injury

Injuries or infections induce unique neuroendocrine and immunologic responses that differentiate injury metabolism from that of unstressed fasting. The magnitude of metabolic expenditure appears to be directly proportional to the severity of insult, with thermal injuries and severe infections having the highest energy demands. The increase in energy expenditure is mediated in part by sympathetic activation and catecholamine release.

Lipid Metabolism Following Injury

Lipids serve as a nonprotein and noncarbohydrate fuel sources that minimize protein catabolism in the injured patient, and lipid metabolism potentially influences the structural integrity of cell membranes and the immune response during systemic inflammation. Fat mobilization (lipolysis) occurs mainly in response to catecholamine stimulus of the hormone-sensitive triglyceride lipase. Other hormonal influences on lipolysis include adrenocorticotropic hormone, catecholamines, thyroid hormone, cortisol, glucagon, growth hormone release, reduction in insulin levels, and increased sympathetic stimulus.

Lipid absorption. Adipose tissue provides fuel for the host in the form of free fatty acids and glycerol during critical illness and injury. Oxidation of 1 g of fat yields approximately 9 kcal of energy. Although the liver is capable of synthesizing triglycerides from carbohydrates and amino acids, dietary and exogenous sources provide the major source of triglycerides. Dietary lipids are not readily absorbable in the gut, but require pancreatic lipase and phospholipase within the duodenum to hydrolyze the triglycerides into free fatty acids and monoglycerides. The free fatty acids and monoglycerides are then readily absorbed by gut enterocytes, which resynthesize triglycerides by esterification of the monoglycerides with fatty acylcoenzyme A (acyl-CoA). Long-chain triglycerides (LCTs), defined as those with 12 carbons or more, undergo esterification and enter the circulation through the lymphatic system as chylomicrons. Shorter fatty acid chains directly enter the portal circulation and are transported to the liver by albumin carriers. Hepatocytes use free fatty acids as a fuel source during stress states, but can also synthesize phospholipids or triglycerides (very-low-density lipoproteins) during fed states. Systemic tissue (e.g., muscle and the heart) can use chylomicrons and triglycerides as fuel by hydrolysis with lipoprotein lipase at the luminal surface of capillary endothelium. Trauma or sepsis suppresses lipoprotein lipase activity in both adipose tissue and muscle, presumably mediated by TNF-α.

Lipolysis and fatty acid oxidation. Periods of energy demand are accompanied by free fatty acid mobilization from adipose stores. In adipose tissues, triglyceride lipase hydrolyzes triglycerides into free fatty acids and glycerol. Free fatty acids enter the capillary circulation and are transported by albumin to tissues requiring this fuel source (e.g., heart and skeletal muscle). Insulin inhibits lipolysis and favors triglyceride synthesis by augmenting lipoprotein lipase activity and intracellular levels of glycerol-3-phosphate. The use of glycerol for fuel depends on the availability of tissue glycerokinase, which is abundant in the liver and kidneys.

Free fatty acids absorbed by cells is conjugated with acyl-CoA within the cytoplasm and transported across membranes by the carnitine shuttle. Mediumchain triglycerides (MCTs), defined as those 6 to 12 carbons in length, bypass the carnitine shuttle and readily cross the mitochondrial membranes. This accounts in part for why MCTs are more efficiently oxidized than LCTs. Ideally, the rapid oxidation of MCTs makes them less prone to fat deposition, particularly within immune cells and the reticuloendothelial system—a common finding with lipid infusion in parenteral nutrition. However, in animal studies, exclusive use of MCTs as fuel is associated with higher metabolic demands and toxicity, as well as essential fatty acid deficiency.

Fatty acyl-CoA undergoes mitochondrial β-oxidation, which produces acetyl-CoA. Each acetyl-CoA molecule subsequently enters the tricarboxylic acid (TCA) cycle for further oxidation to yield 12 adenosine triphosphate (ATP) molecules, carbon dioxide, and water. Excess acetyl-CoA molecules serve as precursors for ketogenesis. Unlike glucose metabolism, oxidation of fatty acids requires proportionally less oxygen and produces less carbon dioxide. This is frequently quantified as the ratio of carbon dioxide produced to oxygen consumed for the reaction, and is known as the respiratory quotient (RQ). An RQ of 0.7 would imply greater fatty acid oxidation for fuel, whereas an RQ of 1 indicates greater carbohydrate oxidation (overfeeding). An RQ of 0.85 suggests the oxidation of equal amounts of fatty acids and glucose.

Ketogenesis. Carbohydrate depletion slows acetyl-CoA entry into the TCA cycle secondary to depleted TCA intermediates and enzyme activity. Increased lipolysis and reduced systemic carbohydrate availability during starvation diverts excess acetyl-CoA toward hepatic ketogenesis. A number of extrahepatic tissues, but not the liver itself, are capable of utilizing ketones for fuel. Ketosis represents a state in which hepatic ketone production exceeds extrahepatic ketone use.

The rate of ketogenesis appears to be inversely related to the severity of injury. Major trauma, severe shock, and sepsis attenuate ketogenesis by increasing insulin levels and by rapid tissue oxidation of free fatty acids.

Carbohydrate Metabolism

Ingested and enteral carbohydrates are primarily digested in the small intestine, in which pancreatic and intestinal enzymes reduce the complex carbohydrates to dimeric units. Disaccharidases are further broken down into hexose units such as glucose and galactose, which require energy dependent transport for primary absorption.

Carbohydrate metabolism primarily refers to the utilization of glucose. The oxidation of 1 g of carbohydrate yields 4 kcal, but administered sugar solutions such as that found in intravenous fluids or parenteral nutrition provides only

3.4 kcal/g of dextrose. The primary goal for maintenance glucose administration in surgical patients serves to minimize muscle wasting. The exogenous administration of small amounts of glucose (approximately 50 g/d) facilitates fat entry into the TCA cycle and reduces ketosis. Studies providing exogenous glucose to septic and trauma patients never have been shown to fully suppress amino acid degradation for gluconeogenesis. This suggests that during periods of stress, other hormonal and proinflammatory mediators have profound influence on the rate of protein degradation and that some degree of muscle wasting is inevitable. Insulin has been shown to reverse protein catabolism during severe stress by stimulating protein synthesis in skeletal muscles and inhibits hepatocyte protein degradation.

In cells, glucose is phosphorylated to form glucose-6-phosphate (G6P). G6P can be polymerized during glycogenesis or catabolized in glycogenolysis. Glucose catabolism occurs by cleavage to pyruvate or lactate (pyruvic acid pathway) or by decarboxylation to pentoses (pentose shunt).

Excess glucose from overfeeding, as reflected by RQs greater than 1.0, can result in conditions such as glucosuria, thermogenesis, and conversion to fat (lipogenesis). Excessive glucose administration results in elevated carbon dioxide production, which may be deleterious in patients with suboptimal pulmonary function.

Injury and severe infections acutely induce a state of peripheral glucose intolerance, despite ample insulin production several-fold above baseline. This may occur in part because of reduced skeletal muscle pyruvate dehydrogenase activity following injury, which diminishes the conversion of pyruvate to acetyl-CoA and subsequent entry into the TCA cycle. The consequent accumulation of 3-carbon structures (e.g., pyruvate and lactate) is shunted to the liver as substrate for gluconeogenesis. Unlike the nonstressed subject, the hepatic gluconeogenic response to injury or sepsis cannot be suppressed by exogenous or excess glucose administration, but rather persists in the hypermetabolic, critically ill patient. Hepatic gluconeogenesis, arising primarily from alanine and glutamine catabolism, provides a ready fuel source for tissues such as those of the nervous system, wounds, and erythrocytes, which do not require insulin for glucose transport. The elevated glucose concentrations also provide a necessary energy source for leukocytes in inflamed tissues and in sites of microbial invasions. Glycogen stores within skeletal muscles can be mobilized as a ready fuel source by epinephrine activation of β-adrenergic receptors.

PROTEIN AND AMINO ACID METABOLISM

The average protein intake in healthy, young adults ranges from 80 to 120 g/d, andevery6gof protein yields approximately 1 g of nitrogen. The degradation of 1 g of protein yields approximately 4 kcal of energy, almost the same as for carbohydrates.

Following injury the initial systemic proteolysis, mediated primarily by glucocorticoids, increases urinary nitrogen excretion to levels in excess of 30 g/day, which roughly corresponds to a loss in lean body mass of 1.5 percent per day. An injured individual who does not receive nutrition for 10 days can theoretically lose 15 percent lean body mass. Therefore amino acids cannot be considered a long-term fuel reserve, and indeed excessive protein depletion (25–30 percent of lean body weight) is not compatible with sustaining life.

Protein catabolism following injury provides substrates for gluconeogenesis and for the synthesis of acute phase proteins. Radiolabeled amino acid incorporation studies and protein analyses confirm that skeletal muscles are preferentially depleted acutely following injury, whereas visceral tissues (e.g., the liver and kidney) remain relatively preserved. The accelerated urea excretion following injury is also associated with the excretion of intracellular elements such as sulfur, phosphorus, potassium, magnesium, and creatinine. Conversely, the rapid use of elements such as potassium and magnesium during recovery from major injury may indicate a period of tissue healing.

The net changes in protein catabolism and synthesis correspond to the severity and duration of injury. The rise in urinary nitrogen and negative nitrogen balance can be detected early following injury and peak by 7 days. This state of protein catabolism may persist for as long as 3–7 weeks. The patient’s prior physical status and age appear to influence the degree of proteolysis following injury or sepsis.

NUTRITION IN THE SURGICAL PATIENT

The goal of nutritional support in the surgical patient is to prevent or reverse the catabolic effects of disease or injury. Although several important biologic parameters have been used to measure the efficacy of nutrition regimens, the ultimate validation for nutritional support should be improvement in clinical outcome and restoration of function.

Estimating Energy Requirements

Nutritional assessment determines the severity of nutrient deficiencies or excess and aids in predicting nutritional requirements. Elements of nutritional assessment include weight loss, chronic illnesses, or dietary habits that influence the quantity and quality of food intake. Social habits predisposing to malnutrition and the use of medications that may influence food intake or urination should also be investigated. Physical examination seeks to assess loss of muscle and adipose tissues, organ dysfunction, and subtle changes in skin, hair, or neuromuscular function reflecting frank or impending nutritional deficiency. Anthropometric data including weight change, skinfold thickness, and arm circumference muscle area and biochemical determinations (i.e., creatinine excretion, albumin, prealbumin, total lymphocyte count, and transferrin) may be used to substantiate the patient’s history and physical findings. Nutritional assessment remains an imprecise method for the detection of malnutrition and prediction of patient outcome. Appreciation for the stresses and natural history of the disease process, in combination with nutritional assessment, remains the basis for identifying patients in acute or anticipated need of nutritional support.

A fundamental goal of nutritional support is to meet the energy requirements for metabolic processes, core temperature maintenance, and tissue repair. The requirement for energy may be measured by indirect calorimetry or estimated from urinary nitrogen excretion, which is proportional to resting energy expenditure. However, the use of indirect calorimetry, particularly in the critically ill patient, is labor-intensive and often leads to overestimation of caloric requirements.

Basal energy expenditure (BEE) may also be estimated using the Harris-Benedict equations:

BEE (men) = 66.47 + 13.75 (W) + 5.0 (H) − 6.76 (A) kcal/d BEE (women) = 655.1 + 9.56 (W) + 1.85 (H) − 4.68 (A) kcal/d

where W = weight in kilograms, H = height in centimeters, and A = age in years.

These equations, adjusted for the type of surgical stress, are suitable for estimating energy requirements in over 80 percent of hospitalized patients. It has been demonstrated that the provision of 30 kcal/kg per day will adequately meet energy requirements in most postsurgical patients, with low risk of overfeeding. Following trauma or sepsis, energy substrate demands are increased, necessitating greater nonprotein calories beyond calculated energy expenditure. These additional nonprotein calories provided after injury are usually 1.2 to 2.0 times greater than calculated resting energy expenditure (REE), depending on the type of injury. It is seldom appropriate to exceed this level of nonprotein energy intake during the height of the catabolic phase.

The second objective of nutritional support is to meet the substrate requirements for protein synthesis. An appropriate nonprotein calorie-to-nitrogen ratio of 150:1 (e.g., 1 g N = 6.25 g protein) should be maintained, which is the basal calorie requirement provided to prevent use of protein as an energy source. There is now greater evidence suggesting that increased protein intake, and a lower calorie-to-nitrogen ratio of 80:1 to 100:1, may benefit healing in selected hypermetabolic or critically ill patients. In the absence of severe renal or hepatic dysfunction precluding the use of standard nutritional regimens, approximately 0.25 to 0.35 g of nitrogen per kilogram of body weight should be provided daily.

Vitamins and Minerals

The requirements for vitamins and essential trace minerals usually can be easily met in the average patient with an uncomplicated postoperative course. Therefore vitamins are usually not given in the absence of preoperative deficiencies. Patients maintained on elemental diets or parenteral hyperalimentation require complete vitamin and mineral supplementation. Essential fatty acid supplementation may also be necessary, especially in patients with depletion of adipose stores.

Overfeeding

Overfeeding usually results from overestimation of caloric needs, as occurs when actual body weight is used to calculate the BEE in such patient populations as the critically ill with significant fluid overload and the obese. Indirect calorimetry can be used to quantify energy requirements, but frequently overestimates BEE by 10–15 percent in stressed patients, particularly if they are on a ventilator. In these instances, estimated dry weight should be obtained from preinjury records or family members.

Adjusted lean body weight also can be calculated. Clinically, increased oxygen consumption, increased CO2 production, fatty liver, suppression of leukocyte function, and increased infectious risks have all been documented with overfeeding.

ENTERAL NUTRITION

Rationale for Enteral Nutrition

Enteral nutrition generally is preferred over parenteral nutrition based on reduced cost and associated risks of the intravenous route. Laboratory models have long demonstrated that luminal nutrient contact reduces intestinal mucosal atrophy when compared with parenteral or no nutritional support. Studies comparing postoperative enteral and parenteral nutrition in patients undergoing GI surgery have demonstrated reduced infection complications and acute phase protein production when fed by the enteral route. Yet, prospectively randomized studies for patients with adequate nutritional status (albumin ≥ 4 g/dL) undergoing GI surgery demonstrate no differences in outcome and complications when administered enteral nutrition compared to maintenance intravenous fluids alone in the initial days following surgery. At the other extreme, recent meta-analysis for critically ill patients demonstrates a 44 percent reduction in infectious complications in those receiving enteral nutritional support over those receiving parenteral nutrition. Most prospectively randomized studies for severe abdominal and thoracic trauma demonstrate significant reductions in infectious complications for patients given early enteral nutrition when compared with those who are unfed or receiving parenteral nutrition.

Recommendations for instituting early enteral nutrition to surgical patients with moderate malnutrition (albumin = 2.9 to 3.5 g/dL) can only be made by inferences because of a lack of data directly pertaining to this population. It is prudent to offer enteral nutrition based on measured energy expenditure of the recovering patient, or if complications arise that may alter the anticipated course of recovery (e.g., anastomotic leaks, return to surgery, sepsis, or failure to wean from the ventilator). Other clinical scenarios with substantiated benefits from enteral nutritional support include permanent neurologic impairment, oropharyngeal dysfunction, short-bowel syndrome, and bone marrow transplantation patients.

Data support the use of early enteral nutritional support following major trauma and in patients who are anticipated to have prolonged recovery after surgery. Healthy patients without malnutrition undergoing uncomplicated surgery can tolerate 10 days of partial starvation (with maintenance intravenous fluids only) before any significant protein catabolism occurs. Earlier intervention is likely indicated in patients with poorer preoperative nutritional status.

Initiation of enteral nutrition should occur immediately after adequate resuscitation, most readily determined by adequate urine output. Presence of bowel sounds and the passage of flatus or stool are not absolute requisites for initiating enteral nutrition, but feedings in the setting of gastroparesis should be administered distal to the pylorus. Enteral feeding should also be offered to patients with short-bowel syndrome or clinical malabsorption, but caloric needs, essential minerals, and vitamins should be supplemented with parenteral modalities.

Enteral Formulas

The functional status of the GI tract determines the type of enteral solutions to be used. Patients with an intact GI tract will tolerate complex enteral solutions. In patients with malabsorption such as in inflammatory bowel diseases, absorption may be improved by provision of dipeptides, tripeptides, and MCTs. However, MCTs are deficient in essential fatty acids, which necessitates supplementation with some LCT.

In general, factors that influence the choice of enteral formula include the extent of organ dysfunction (e.g., renal, pulmonary, hepatic, or GI), the nutrient needs to restore optimal function and healing, and the cost of specific products.

Immune-enhancing formulas. These formulas are fortified with special nutrients that are purported to enhance various aspects of immune or solid organ function. Such additives include glutamine, arginine, branched-chain amino acids, omega-3 fatty acids, nucleotides, and beta-carotene. Although several trials have proposed that one or more of these additives reduce surgical complications and improve outcome, these results have not been uniformly corroborated by other trials. The addition of amino acids to these formulas generally doubles the amount of protein (nitrogen) found in standard formula; however, their use can be cost-prohibitive.

High-protein formulas. High-protein formulas are available in isotonic and nonisotonic mixtures and are proposed for critically ill or trauma patients with high protein requirements. These formulas comprise nonprotein calorie-tonitrogen ratios between 80:1 and 120:1.

Elemental formulas. These formulas contain predigested nutrients and provide proteins in the form of small peptides. Complex carbohydrates are limited, and fat content, in the form of MCTs and LCTs, is minimal. The primary advantage of such a formula is ease of absorption, but the inherent scarcity of fat, associated vitamins, and trace elements limits its long-term use as a primary source of nutrients. Because of its high osmolarity, dilution or slow infusion rates are usually necessary, particularly in critically ill patients. These formulas have been used frequently in patients with malabsorption, gut impairment, and pancreatitis, but their cost is significantly higher than that of standard formulas.

Renal-failure formulas. The primary benefits of the renal formula are the lower fluid volume and concentrations of potassium, phosphorus, and magnesium needed to meet daily calorie requirements. This formulation almost exclusively contains essential amino acids and has a high nonprotein-to-calorie ratio; however, it does not contain trace elements or vitamins.

Pulmonary-failure formulas. In these formulas, fat content is usually increased to 50 percent of the total calories, with a corresponding reduction in carbohydrate content. The goal is to reduce CO2 production and alleviate ventilation burden for failing lungs.

Access for Enteral Nutritional Support

The available techniques and repertoire for enteral access have provided multiple options for feeding the gut. Table 1-4 summarizes the presently used methods and preferred indications.

TABLE 1-4 Options for Enteral Feeding Access

Access option Comments

Nasogastric tube Short-term use only; aspiration risks; nasopharyngeal trauma; frequent dislodgment

Nasoduodenal/nasojejunal Short-term use; lower aspiration risks in jejunum; placement challenges (radiographic assistance often necessary)

Percutaneous endoscopic Endoscopy skills required; may be used for

gastrostomy (PEG) gastric decompression or bolus feeds; aspiration risks; can last 12–24 months; slightly higher complication rates with placement and site leaks

Surgical gastrostomy Requires general anesthesia and small laparotomy; may allow placement of extended duodenal/jejunal feeding ports; laparoscopic placement possible

Fluoroscopic gastrostomy Blind placement using needle and T-prongs to anchor to stomach; can thread smaller catheter through gastrostomy into duodenum/jejunum under fluoroscopy

PEG-jejunal tube Jejunal placement with regular endoscope is operator dependent; jejunal tube often dislodges retrograde; two-stage procedure with PEG placement, followed by fluoroscopic conversion with jejunal feeding tube through PEG

Direct percutaneous Direct endoscopic placement with

endoscopic jejunostomy enteroscope; placement challenges; greater

(DPEJ) injury risks

Surgical jejunostomy Commonly applied during laparotomy; general anesthesia; laparoscopic placement usually requires assistant to thread catheter; laparoscopy offers direct visualization of catheter placement

Fluoroscopic jejunostomy Difficult approach with injury risks; not commonly done

Nasoenteric tubes. Nasogastric feeding should be reserved for those with intact mental status and protective laryngeal reflexes to minimize risks of aspiration. Even in intubated patients, nasogastric feedings can often be recovered from tracheal suction. Nasojejunal feedings are associated with fewer pulmonary complications, but access past the pylorus requires greater effort to accomplish. Blind insertion of nasogastric feeding tubes is fraught with misplacement, and air instillation with auscultation is inaccurate for ascertaining proper positioning. Radiographic confirmation is usually required to verify the position of the nasogastric feeding tube.

Small bowel feeding is more reliable for delivering nutrition than nasogastric feeding. Furthermore, the risks of aspiration pneumonia can be reduced by 25 percent with small bowel feeding when compared with nasogastric feeding. The disadvantages of nasoenteric feeding tubes are clogging, kinking, inadvertent displacement or removal, and nasopharyngeal complications. If nasoenteric feeding will be required for longer than 30 days, access should be converted to a percutaneous one.

Percutaneous endoscopic gastrostomy. The most common indications for percutaneous endoscopic gastrostomy (PEG) placement include impaired swallowing mechanisms, oropharyngeal or esophageal obstruction, and major facial trauma. It is frequently utilized for debilitated patients requiring caloric supplementation, hydration, or frequent medication dosing. It is also appropriate for patients requiring passive gastric decompression. Relative contraindications for PEG placement include ascites, coagulopathy, gastric varices, gastric neoplasm, and lack of a suitable abdominal site.

If endoscopy is not available or technical obstacles preclude PEG placement, the interventional radiologist can attempt the procedure percutaneously under fluoroscopic guidance. If this is also unsuccessful, surgical gastrostomy or small-bowel tube placement can be considered.

Although PEG tubes enhance nutritional delivery, facilitate nursing care, and are superior to nasogastric tubes, serious complications can occur in approximately 3 percent of patients. These complications include wound infection, necrotizing fasciitis, peritonitis, aspiration, leaks, dislodgment, bowel perforation, enteric fistulas, bleeding, and aspiration pneumonia. For patients with significant gastroparesis or gastric outlet obstruction, feedings through PEG tubes are hazardous. In this instance, the PEG tube can be used for decompression and to allow access for converting the PEG tube to a transpyloric feeding tube.

Surgical gastrostomy and jejunostomy. In a patient undergoing complex abdominal or trauma surgery, thought should be given during surgery to the possible routes for subsequent nutritional support, because laparotomy affords direct access to the stomach or small bowel. The only absolute contraindication to feeding jejunostomy is distal intestinal obstruction. Relative contraindications include severe edema of the intestinal wall, radiation enteritis, inflammatory bowel disease, ascites, severe immunodeficiency, and bowel ischemia. Needlecatheter jejunostomies can also be done with a minimal learning curve. The biggest drawback is usually related to clogging and knotting of the 6F catheter.

Abdominal distention and cramps are common adverse effects of early enteral nutrition and usually managed by temporarily discontinuing feeds and resuming at a lower infusion rate.

Pneumatosis intestinalis and small bowel necrosis are infrequent but significant problems associated with patients receiving jejunal tube feedings. Several contributing factors have been proposed, including the hyperosmolar consistency of enteral solutions, bacterial overgrowth, fermentation, and metabolic breakdown products. The common pathophysiology is believed to be bowel distention and consequent reduction in bowel wall perfusion. Risk factors for these complications include cardiogenic and circulatory shock, vasopressor use, diabetes mellitus, and chronic obstructive pulmonary disease. Therefore, enteral feedings in the critically ill patient should be delayed until adequate resuscitation has been achieved.

PARENTERAL NUTRITION

Parenteral nutrition involves the continuous infusion of a hyperosmolar solution containing carbohydrates, proteins, fat, and other necessary nutrients through an indwelling catheter inserted into the superior vena cava. To obtain the maximum benefit, the ratio of calories to nitrogen must be adequate (at least 100 to 150 kcal/g nitrogen), and both carbohydrates and proteins must be infused simultaneously. When the sources of calories and nitrogen are given at different times, there is a significant decrease in nitrogen utilization. These nutrients can be given in quantities considerably greater than the basic caloric and nitrogen requirements. Clinical trials and meta-analysis of parenteral feeding in the perioperative period have suggested that preoperative nutritional support may benefit some surgical patients, particularly those with extensive malnutrition. Short-term use of parenteral nutrition in critically ill patients (i.e., duration <7 days) when enteral nutrition may have been instituted is associated with higher rates of infectious complications. Following severe injury, parenteral nutrition is associated with higher rates of infectious risks when compared with enteral feeding. However, parenteral feeding still has fewer infectious complications compared with no feeding at all.

Rationale for Parenteral Nutrition

The principal indications for parenteral nutrition are found in seriously ill patients suffering from malnutrition, sepsis, or surgical or accidental trauma, when use of the GI tract for feedings is not possible. In some instances, intravenous nutrition may be used to supplement inadequate oral intake. The safe and successful use of parenteral nutrition requires proper selection of patients with specific nutritional needs, experience with the technique, and an awareness of the associated complications. As with enteral nutrition, the fundamental goals are to provide sufficient calories and nitrogen substrate to promote tissue repair and to maintain the integrity or growth of lean tissue mass. Listed below are situations in which parenteral nutrition has been used in an effort to achieve these goals:

1. Newborn infants with catastrophic GI anomalies, such as tracheoesophageal fistula, gastroschisis, omphalocele, or massive intestinal atresia.
2. Infants who fail to thrive because of GI insufficiency associated with shortbowel syndrome, malabsorption, enzyme deficiency, meconium ileus, or idiopathic diarrhea.
3. Adult patients with short-bowel syndrome secondary to massive small bowel resection (<100 cm without colon or ileocecal valve, or <50 cm with intact ileocecal valve and colon).
4. Enteroenteric, enterocolic, enterovesical, or high-output enterocutaneous fistulas (>500 mL/d).
5. Surgical patients with prolonged paralytic ileus following major operations (>7–10 days), multiple injuries, blunt or open abdominal trauma, or patients with reflex ileus complicating various medical diseases.
6. Patients with normal bowel length but with malabsorption secondary to sprue, hypoproteinemia, enzyme or pancreatic insufficiency, regional enteritis, or ulcerative colitis.
7. Adult patients with functional GI disorders such as esophageal dyskinesia following cerebrovascular accident, idiopathic diarrhea, psychogenic vomiting, or anorexia nervosa.
8. Patients with granulomatous colitis, ulcerative colitis, and tuberculous enteritis, in which major portions of the absorptive mucosa are diseased.
9. Patients with malignancy, with or without cachexia, in whom malnutrition might jeopardize successful delivery of a therapeutic option.
10. Failed attempts to provide adequate calories by enteral tube feedings or high residuals.
11. Critically ill patients who are hypermetabolic for more than 5 days or when enteral nutrition is not feasible.

Conditions contraindicating hyperalimentation include the following:

1. Lack of a specific goal for patient management, or in cases in which instead of extending a meaningful life, inevitable dying is delayed.
2. Periods of hemodynamic instability or severe metabolic derangement (e.g., severe hyperglycemia, azotemia, encephalopathy, hyperosmolality, and fluid-electrolyte disturbances) requiring control or correction before attempting hypertonic intravenous feeding.
3. Feasible GI tract feeding; in the vast majority of instances, this is the best route by which to provide nutrition.
4. Patients with good nutritional status.
5. Infants with less than 8 cm of small bowel, because virtually all have been unable to adapt sufficiently despite prolonged periods of parenteral nutrition.

Total Parenteral Nutrition

Total parenteral nutrition (TPN), also referred to as central parenteral nutrition, requires access to a large-diameter vein to deliver the entire nutritional requirements of the individual. Dextrose content is high (15–25 percent) and all other macro-and micronutrients are deliverable by this route.

Peripheral Parenteral Nutrition

The lower osmolarity of the solution used for peripheral parenteral nutrition (PPN), secondary to reduced dextrose (5–10 percent) and protein (3 percent) levels, allows for its administration via peripheral veins. Some nutrients cannot be supplemented because of inability to concentrate them into small volumes. Therefore PPN is not appropriate for repleting patients with severe malnutrition. It can be considered if central routes are not available or if supplemental nutritional support is required. Typically, PPN is used for short periods (< 2 weeks).

Intravenous Access Methods

Temporary or short-term access can be achieved with a 16-gauge, percutaneous catheter inserted into a subclavian or internal jugular vein and threaded into the superior vena cava. More permanent access, with the intention of providing long-term or home parenteral nutrition, can be achieved by placement of a catheter with a subcutaneous port for access, by tunneling a catheter with a substantial subcutaneous length, or threading a long catheter through the basilic or cephalic vein into the superior vena cava.

COMPLICATIONS OF PARENTERAL NUTRITION

Technical Complications

One of the more common and serious complications associated with long-term parenteral feeding is sepsis secondary to contamination of the central venous catheter. Complications related to catheter insertion such as pneumothorax, hemothorax, subclavian artery injury, thoracic duct injury, cardiac arrhythmia, air embolism, catheter embolism, and cardiac perforation with tamponade have also been described. Contamination of solutions should be considered, but is rare when proper pharmacy protocols have been followed. This problem occurs more frequently in patients with systemic sepsis, and in many cases is because of hematogenous seeding of the catheter with bacteria. It is prudent to delay reinserting the catheter by 12–24 hours, especially if bacteremia is present.

Metabolic Complications

Hyperglycemia may develop with normal rates of infusion in patients with impaired glucose tolerance or in any patient if the hypertonic solutions are administered too rapidly. This is a particularly common complication in latent diabetics and in patients subjected to severe surgical stress or trauma. Treatment of the condition consists of volume replacement with correction of electrolyte abnormalities and the administration of insulin. This complication can be avoided with careful attention to daily fluid balance and frequent monitoring of blood sugar levels and serum electrolytes.

Increasing experience has emphasized the importance of not overfeeding the parenterally nourished patient. This is particularly true of the depleted patient in whom excess calorie infusion may result in carbon dioxide retention and respiratory insufficiency. Additionally, excess feeding also has been related to the development of hepatic steatosis or marked glycogen deposition in selected patients. Cholestasis and formation of gallstones are common in patients receiving long-term parenteral nutrition. Mild but transient abnormalities of serum transaminase, alkaline phosphatase, and bilirubin may occur in many parenterally nourished patients. Failure of the liver enzymes to plateau or return to normal over 7–14 days should suggest another etiology.

Intestinal Atrophy

Lack of intestinal stimulation is associated with intestinal mucosal atrophy, diminished villous height, bacterial overgrowth, reduced lymphoid tissue size, reduced IgA production, and impaired gut immunity. The full clinical implications of these changes are not well realized, although bacterial translocation has been demonstrated in animal models. The most efficacious method to prevent these changes is to provide nutrients enterally. In patients requiring total parenteral nutrition, it may be feasible to infuse small amounts of trophic feedings via the GI tract.

Special Formulations

Glutamine and Arginine

Glutamine is the most abundant amino acid in the human body, comprising nearly two-thirds of the free intracellular amino acid pool. Of this, 75 percent is found within the skeletal muscles. In healthy individuals, glutamine is considered a nonessential amino acid because it is synthesized within the skeletal muscles and the lungs. Glutamine is a necessary substrate for nucleotide synthesis in most dividing cells, and hence provides a major fuel source for enterocytes. It also serves as an important fuel source for immunocytes such as lymphocytes and macrophages, and a precursor for glutathione, a major intracellular antioxidant. During stress states such as sepsis, or in tumor-bearing hosts, peripheral glutamine stores are rapidly depleted and the amino acid is preferentially shunted as a fuel source toward the visceral organs and tumors, respectively. These situations create, at least experimentally, a glutamine-depleted environment, with consequences including enterocyte and immunocyte starvation.

Arginine, also a nonessential amino acid in healthy subjects, first attracted attention for its immunoenhancing properties, wound-healing benefits, and improved survival in animal models of sepsis and injury. As with glutamine, the benefits of experimental arginine supplementation during stress states are diverse. Clinical studies in which arginine was administered enterally have demonstrated net nitrogen retention and protein synthesis compared to isonitrogenous diets in critically ill and injured patients and following surgery for certain malignancies. Some of these studies also are associated with in vitro evidence of enhanced immunocyte function. The clinical utility of arginine in improving overall patient outcome remains an area of investigation.

Omega-3 Fatty Acids

The provision of omega-3 polyunsaturated fatty acids (canola oil or fish oil) displaces omega-6 fatty acids in cell membranes, which theoretically reduce the proinflammatory response from prostaglandin production.

Nucleotides

RNA supplementation in solutions is purported, at least experimentally, to increase cell proliferation, provide building blocks for DNA synthesis, and improve T-helper cell function.

Suggested Readings

Henneke P, Golenbock DT: Innate immune recognition of lipopolysaccharide by endothelial cells. Crit Care Med 30:S207, 2002. Lin E, Calvano SE, Lowry SF: Inflammatory cytokines and cell response in surgery.

Surgery 127:117, 2000. Lin E, Lowry SF: Human response to endotoxin. Sepsis 2:255, 1999. Marshall JC, Vincent J-L, Fink MP, et al: Measures, markers, and mediators: To

ward a staging system for clinical sepsis. A report of the Fifth Toronto Sepsis Roundtable, Toronto, Ontario, Canada, October 25–26, 2000. Crit Care Med 31:1560, 2003.

Vincent J-L, Sun Q, Dubois M-J: Clinical trials of immunomodulatory therapies in severe sepsis and septic shock. Clin Infect Dis 34:1084, 2002. Wilmore DW: From Cuthbertson to fast-track surgery: 70 years of progress in reducing stress in surgical patients. Ann Surg 236:643, 2002.

Cerra FB, Benitez MR, Blackburn GL, et al: Applied nutrition in ICU patients: A consensus statement of the American College of Chest Physicians. Chest 111:769, 1997.

Guirao X: Impact of the inflammatory reaction on intermediary metabolism and nutrition status. Nutrition 18:949, 2002. Heslin MJ, Brennan MF: Advances in perioperative nutrition: Cancer. World J Surg 24:1477, 2000.

Zhou Y-P, Jiang Z-M, Sun Y-H, et al: The effect of supplemental enteral glutamine on plasma levels, gut function, and outcome in severe burns: A randomized, double-blind, controlled clinical trial. J Parenter Enteral Nutr 27:241, 2003.