Role of Nutritional Support for Critically Ill Children

Contemporary nutritional support depends upon a balanced approach that provides macronutrients (e.g., protein or amino acids, carbohydrate, and fat) as well as a corresponding amount of micronutrients (e.g., minerals, trace elements, vitamins). While clinicians tend to take each of these components for granted, one must consider that the macronutrients must be processed via intermediary metabolic pathways to produce adenosine triphosphate (ATP) when and where needed and to provide structural molecules for tissue repair, maintenance of organ integrity, and immunoglobulin synthesis. The metabolic processes associated with biotransformation of macronutrients for both kidney and liver may produce further demands on organs that often are already working at limited capacity owing to perfusion deficits and humoral factors such as lipopolysaccharides (LPS), proinflammatory cytokines that tend to limit reserve capacity. Excess calories can be stored, whereas excess protein beyond what the body can utilize is rapidly degraded to urea as the final byproduct.

The protein pools of the body are conceptualized as existing in two compartments. The first compartment, referred to as the visceral protein pool, represents those proteins that can easily be accessed and degraded to provide amino acids when nutritional intake is inadequate. The visceral compartment includes plasma proteins, immunoglobulins, cytosolic enzymes, and so on that can be turned over readily. The second compartment, referred to as the somatic protein pool, represents primarily structural proteins in brain, heart, kidney, bone, and the like; this compartment is less accessible than the visceral pool. Protein is stored to a very limited extent when taken in beyond momentary need. Carbohydrate is generally stored to a limited extent as glycogen, beyond which it is converted into triglycerides under the influence of insulin. Fats must be taken up and cleared from the plasma through complex mechanisms involving lipoprotein lipase, which is impaired during stress states. Thus all three major macronutrients, when provided to patients by enteral or parenteral approach, depend upon a multitude of internal processes frequently altered by critical illness, leading to additional stress with no survival benefit. After decades of clinical research, the conclusion must be that excess nutritional support is to be avoided in favor of thoughtful moderation in providing nutrients.

Contemporary nutritional support has three primary goals: (1) preservation of lean body mass to minimize the catabolic consequence of critical illness, (2) provision of suitable substrates to permit restoration of immune function and repair of body tissues, and (3) prevention of nutrition-related complications including aspiration risks in patients receiving enteral nutrition and the avoidance of nutrient-induced organ overload, whether through excess carbohydrate (increased CO₂ production and hepatic steatosis) or excess protein/nitrogen load to the liver and kidney. Understanding the risks and realistic benefits of nutritional support are vital in current PICU care.

Impact of Physiologic Stress on Children

Alterations in protein and energy metabolism are hallmarks of critical illness and have been studied for many decades.⁴ This work has demonstrated a great difference between short-term starvation states in otherwise healthy individuals and the dramatic “autocannibalism” seen in critically ill patients who are not receiving appropriate nutritional support as summarized in Table 95-1.

TABLE 95-1 Comparison of Nutrient Metabolism in Starvation Versus Sepsis/Trauma

Adapted from Barton R, Cerra FB. The hypermetabolism-multiple organ failure syndrome. Chest 1989;96:1153-60.

The events that lead to ICU admission are extremely varied, yet the body’s response to acute physiologic stress tends to be similar whether the inciting event is sepsis, ischemia-reperfusion, trauma, burns, or other inflammatory conditions. Beyond low levels of stress, such as minor elective surgery, life-threatening illness, burns, organ transplantation, or major surgical procedures elicit dramatic systemic inflammatory responses due to activation of the immune system, clotting mechanisms, and the endothelium. The patient’s ability to withstand the metabolic responses to such stresses and ultimately to reverse the process is central to recovery. A complete discussion of the metabolic response to stress is beyond the scope of this chapter; the reader is referred to other sources.^5,6

The initial response to injury is to activate endothelial cells and to prime inflammatory cells such as neutrophils, macrophages, and lymphocytes through proinflammatory mediators including tumor necrosis factor, interleukin 2, histamine, eicosanoids, heat-shock proteins, free radicals, platelet-activating factor, and tryptases.⁷ These same signals that produce activation of the endothelium lead to permeability changes, activation of clotting mechanisms, and changes in hepatic and peripheral protein metabolism.⁸ If recovery is to occur, this process must be extinguished by a decrease in the inflammatory state and an increase in tissue repair.⁹ Although it may seem that simply shutting off the proinflammatory signals should lead to resolution, the process of resolving inflammation appears more complex.¹⁰ Studies show the importance of many of the proinflammatory stimuli in regeneration and repair, and the timing of interventions is important.¹¹ In response to injury, a wide range of neurohumoral reactions occur, forming the classic “stress response,” which includes elevation of growth hormone, endogenous catecholamines, glucagon, and cortisol. Recognition of the role of insulin-like growth factor-1 along with growth hormone in promoting protein synthesis and counter-regulating inflammatory states suggests important potential treatment options that have been best studied in burns. Despite these studies showing benefit from growth hormone supplementation, evidence of increased mortality rate after growth hormone supplementation also has been reported.¹² Clinicians must balance the relative benefit of hormonal manipulation with potential risks.

In the inflammatory state, unremitting gluconeogenesis occurs through the release of glycerol and gluconeogenic amino acids from the periphery with their conversion to glucose in the liver and kidney. Hyperglycemia frequently is associated with this state and may induce glycosuria and an osmotic diuresis. Insulin activity becomes impaired at the tissue level, leading to so-called insulin resistance in the face of the powerful gluconeogenesis driven by the stress hormones. It seems that the impairment of insulin results from decreased phosphorylation of the insulin receptor and second messengers.¹¹ In the last decade, evidence from adult ICU experience has suggested a benefit from the use of insulin infusions to maintain tight control over serum glucose level.¹³ Although much of the preceding information derives from adult studies, it has found its way into contemporary pediatric practice in many centers in children of various ages. This question is receiving intense scrutiny in critically ill children through multicenter trials which are currently underway. The use of insulin infusions to control hyperglycemia in premature infants continues to be standard practice; however, the potential to produce marked hepatic steatosis under the influence of insulin should be born in mind when choosing the amount of carbohydrate to provide.

The breakdown of protein is a central theme in the body’s response to stress, which has wide-ranging significance beyond simple protein losses. The conversion of certain amino acids to glucose and the oxidation of others in peripheral tissues lead to the liberation of large quantities of amino-nitrogen, which would become toxic if not for the efficient conversion to urea. A dramatic increase in the rate of urea production is seen in critically ill patients. Concomitantly, other non-urea nitrogen is liberated in the form of uric acid and creatine and accounts for the dramatic increase in nitrogen wasting seen during stress states. Total urinary nitrogen losses in critically ill children may be 0.3 g/kg/d, which represents the loss of approximately 1.8 g/kg/d of whole protein catabolized. In parallel with the increased turnover of proteins, the metabolic rate for oxidation of energy substrates may increase following acute critical illness during the recovery phase (see subsequent section on energy expenditure).

The body’s response to withholding feeding (i.e., starvation) in healthy individuals is qualitatively and quantitatively different than that seen when nutrient intake is absent during critical illness. These differences are fundamental to understanding nutritional support in the ICU and are summarized in Table 95-1. In simple starvation, the body’s regulatory mechanisms for sparing lean tissue and using triglycerides as the primary energy source are intact, whereas under the influence of the stress response, rapid depletion of lean tissues occurs with oxidation of amino acids, carbohydrate, and fat as energy substrates.

One of the major consequences of life-threatening physiologic stress is the net depletion of body protein representing the somatic protein pool (e.g., skeletal muscle mass) and functional (e.g., plasma proteins, enzyme systems, antibodies) tissues contained in the visceral protein pool. With protein catabolism rates increased up to twofold, synthesis does not keep pace, and a state of negative nitrogen balance ensues when patients are not given adequate calories and protein.¹⁴ These changes produce depressed function of T and B lymphocytes, monocytes, and neutrophils as cumulative protein loss increases. The synthesis of antibodies, chemotaxis, phagocytosis, and bacterial killing is impaired in the face of advanced protein-calorie malnutrition.¹⁵ A decrease in total lymphocyte count may be seen in many patients, but a total lymphocyte count less than 1200/mm³ should raise concern for the presence of possible immune dysfunction. These alterations lead to impairment of host defense mechanisms. As noted earlier, for resolution of the inflammatory response, the patient’s immune system plays a central role in recovery of wound healing and recovery of immune competence.¹⁶ It is likely that the syndrome of multiple organ dysfunction seen in critically ill patients is due in part to the inability of the immune system to down-regulate the inflammatory response to injury in specific organs, as well as acquired mitochondrial dysfunction leading to ineffective cellular energy production.¹⁷ Nutritional support of a critically ill patient is thought to be essential to achieving recovery and minimizing the subsequent period of convalescence.

Considerable attention currently is focused on the use of modified nutritional support regimens in critically ill adults to modify the inflammatory response and reduce secondary organ system dysfunction.¹⁸ A wide range of substances have been studied in an attempt to improve outcome or minimize nitrogen loss during critical illness in specific populations of patients. Glutamine supplementation appears to benefit critically ill adults, particularly those with burns.¹⁸ Omega-3 fatty acids appear to also be beneficial in patients with sepsis and systemic inflammatory response syndrome (SIRS). The results in adults suggest that formulas supplemented with these products improve oxygenation and reduce the alveolar inflammatory response during acute respiratory distress syndrome (ARDS). While trials of these agents are underway in critically ill children, there is still not a strong enough consensus among pediatric specialists to consider their use as standard therapy.³

Nutrition Assessment

The nutrition assessment of hospitalized children is a central and critical part of the initial examination and evaluation of all patients. The existence of chronic malnutrition as well as the development of acute malnutrition during critical illness has been recognized in pediatric critical care for many years^19–21 and appears to be an unmet need even today.²² Therefore, clinicians must assess newly admitted patients for the presence of malnutrition that may complicate the response to therapies or impair recovery (Box 95-1). The presence of previous severe malnutrition may complicate critical care management through the presence of marasmic cardiomyopathy, severe intracellular energy deficiency, and the development of refeeding disequilibrium when nutrients are provided in the ICU.²³

Box 95-1

Assessment of Nutrition Status on Admission

*Actual height (cm) × 100/expected height at 50th percentile for age.

^†BMI (kg/m²) = body mass index: actual weight (kg)/[actual weight × height (m)] ²

From Statistics NCfH. CDC growth charts, United States, 2000. Available at: http://www.cdc.gov/growthcharts; and from Waterlow J. Classification and definition of protein-calorie malunutrition. Br J Med 1972;3:566-9.

The initial nutrition evaluation consists of assessing the patient’s weight, height, historical evidence for recent weight loss, and anthropometric measurements including midarm circumference and skinfold determination (when edema is not present). Nutrition history must include the presence and duration of nausea, vomiting, diarrhea, fever, frequent infections, fatigue, food aversion, abdominal discomfort, or feeding intolerance. For growth standards, norms exist reflecting age and gender.²⁴ Ethnic background and considerations such as the presence of certain syndromes (e.g., Down syndrome) or the child’s birth status (e.g., premature, growth restricted, etc.) may affect the child’s growth status.

In particular, determination of body mass index (BMI, previously known as weight-for-height) for children older than 2 years of age provides important information regarding the previous nutritional status (see Box 95-1). In children younger than 2 years, the weight-for-age in light of the previous growth status is most useful. These straightforward measurements have withstood the test of time and were used by Pollack and coworkers to estimate the risk of malnutrition in critically ill children admitted to a multidisciplinary PICU.^19,20,25 Their findings demonstrated higher rates of preexisting malnutrition than had been previously thought. In addition, there was an unexpected deterioration in nutrition indices following admission, suggesting the powerful effects of life-threatening illness on nutritional stores and status even with excellent clinical care. Clinicians caring for children who will experience more than a few days of hospitalization must therefore be especially aware of the potential for acquired nutritional depletion. Potential sources of error exist in interpreting anthropometric measurements that are primarily related to changes in body water associated with many acute critical illnesses in children (i.e., conditions producing capillary leak syndrome or defects in renal water clearance). Such conditions may invalidate the measurement of skinfold or midarm circumference; however, their longitudinal use in patients can be very useful in estimating the accretion of fat and lean tissue stores. It is standard practice to measure these parameters in patients at risk for malnutrition, such as those with cystic fibrosis, short bowel syndrome, and other conditions in which malabsorption or chronically elevated metabolic demands exist (e.g., congenital heart failure, bronchopulmonary dysplasia, and similar chronic conditions).

The triceps and scapular skin folds measure the subcutaneous tissue compartment (consisting primarily of adipose tissue) but also tissue edema in patients with anasarca from any cause. Triceps skin fold is measured by standardized skin caliper and is subject to considerable error if not performed in a consistent manner midway between the acromion and olecranon. The midarm circumference should be measured at the same point with a nonstretchable tape measure. The two indices taken together permit a reliable estimate of muscle mass. In general, good correlation exists between skinfold and arm circumference and weight-for-height percentile.²⁶ During critical illness, anasarca may obscure the loss of lean tissue, which may only be apparent following resolution of edema when successful diuresis has occurred. A very reliable indicator of global loss of lean body mass can be seen in the wasting of the interosseous and thenar muscles of the hand, which becomes apparent 2 or 3 weeks after hospitalization with resolution of edema.

In addition to anthropometric measurements, longitudinal determination of specific plasma proteins including albumin, transferrin, and prealbumin have demonstrated value in assessing the response of patients to nutritional support. Frequently serum proteins will be decreased during acute critical illness without reflecting preceding malnutrition. This phenomenon occurs with capillary leak syndrome, seen in the first hours following PICU admission in patients with sepsis, cardiopulmonary bypass operations, ischemia-reperfusion injury, and similar stresses. Loss of endothelial barrier function causes large molecules such as albumin, which are normally three to four times more concentrated in the vascular compartment than in the interstitial fluid, to move into the extravascular space, lowering their concentration without a concomitant decrease in the total body pool of albumin. This effect may be very pronounced in patients who have received large volumes of crystalloid fluid during their resuscitation. Clinicians must guard against the tendency to replace albumin during acute critical illness solely based upon a low albumin level. Measures to correct the underlying pathophysiology should be considered before administering albumin. Serum albumin in healthy children is generally above 3.0 g/dL, and edema is rarely seen in otherwise healthy children until the albumin falls below 2.0 g/dL, such as in nephrotic syndrome.

Shorter half-life serum proteins such as prealbumin [T_1/2 = 2 days] and transferrin [T_1/2 = 7 days] also reflect nutrition status and respond more quickly to changes in anabolic state.²⁷ As noted earlier, the pool of proteins in the plasma, interstitial space, and some intracellular proteins represent a relatively labile pool of protein referred to as the visceral protein pool. Visceral proteins are rapidly turned over relative to structural proteins that comprise the somatic protein pool. In critical illness, the synthesis of specific proteins such as C-reactive protein, ceruloplasmin, and α₂-macroglobulin is increased, whereas the synthesis of other proteins such as albumin [T_1/2 = ~20 days] is decreased.²⁸ These changes may be seen within 6 hours of the onset of severe physiologic stress. This response to physiologic stress is under the regulation of complex neurohumoral control and is referred to as the acute phase response. It is largely responsible for the increase in erythrocyte sedimentation rate associated with acute inflammatory conditions.²⁹ When followed longitudinally, the return of previously depressed levels of certain visceral proteins such as albumin, transferrin, retinol-binding protein, or prealbumin represents the abatement of physiologic stress or improvement in nutrition when levels are low due to protein-calorie malnutrition. Such positive changes herald the impending return to a state of growth and tissue accretion, barring reentry into a new inflammatory state.

Energy Expenditure

All cellular processes require energy, generally in the form of ATP which is produced through oxidation of metabolic fuels, with heat and water as byproducts. The production of ATP is closely coupled to cellular metabolism and must be maintained to prevent cell death. As ATP levels fall, ionic gradients cannot be maintained, excitatory cells cannot depolarize, the synthesis of new cells and repair of damaged cell constituents cannot occur, and mechanical work such as cardiac pump function and respiratory activity cease. Thus, the body has numerous mechanisms for efficiently producing energy from a wide variety of substrates including protein, fat, and carbohydrates. Following the adaptation to decreased nutrient intake, an otherwise healthy individual will rely upon ketone bodies derived from the breakdown of fat stores to provide critical intracellular energy. Protein stores are relatively spared as the decrease in insulin output allows the metabolism to shift to a ketone-based state. As indicated in Table 95-1, critical illness prevents the body’s conservational mechanisms in response to decreased intake, leading to relatively rapid depletion of carbohydrate and available protein stores.

The close coupling between oxidative metabolism and substrate utilization is reflected in the amount of oxygen consumed (VO₂) and carbon dioxide produced (VCO₂) through the pathways of intermediary metabolism, which include the glycolytic pathway and the tricarboxylic acid cycle. Specific substrates such as fat, protein, and various carbohydrates have a characteristic relationship between VO₂ and VCO₂ based upon the stoichiometry of their unique oxidation. This relationship is referred to as the respiratory quotient (RQ = VCO₂/VO₂) and may be measured through the quantification of respiratory gas exchange through the patient’s lung. The overall metabolic rate is most easily determined in the clinical setting through the process of indirect calorimetry, a process that estimates the resting energy expenditure (REE) based upon VO₂ and VCO₂.³⁰ Indirect calorimetry is well established in clinical nutrition but has been elusively difficult to perform with consistent results and easily applied technology in children. The respiratory quotient for fats is around 0.707 and for proteins around 0.80 and, in conjunction with urinary nitrogen determination, forms the basis for determining the specific substrates being utilized.³⁰ This concept is demonstrated for the aerobic metabolism of glucose:

The availability of equipment to reliably perform indirect calorimetry in children has been a major obstacle to its widespread application. Several factors limit the reliability with which indirect calorimetry can be performed in young children, including non–steady state due to patient movement and nursing interventions, use of uncuffed endotracheal tubes producing loss of respiratory gases, high bias flows on infant ventilators, use of elevated inspired oxygen in nonintubated infants, as well as the small tidal volumes seen in the smallest patients. When indirect calorimetry is not feasible, VO₂ can be calculated in many patients via the Fick equation (A × VdO₂ × cardiac output) when a reliable measure of cardiac output is available. Based upon a conversion factor of approximately 5 kcal of energy per liter of oxygen consumed, one can closely estimate metabolic rate³⁰ if the oxygen consumption is known and the RQ is assumed to be in a normal range.

Through indirect calorimetry it has become clear that patients with similar clinical appearances may have widely differing metabolic rates when adjusted for age and weight.^3,31–34 The differences may be as great as 300%, suggesting the potential for severe over- or undernutrition depending upon the values assumed.³⁵ Thus, clinicians generally must rely upon information provided in controlled studies to guide the delivery of calories, since most will not have a means of easily determining the REE. A wide range of predictive equations have been devised which attempt to predict energy requirements of critically ill children, but it is clear that no single method of estimating caloric expenditures will be successful for all critically ill children.^3,4,36

In very young infants, the effects of environmental cold stress is recognized as a source of unnecessary morbidity.³⁷ The thermal neutral zone in infants up to 1 year of age tends to be several degrees higher than that for burned adults or older children. Heat lost to the environment produces rapid drops in core temperature in young children, with concomitant increase in metabolic demands. Maintaining the environment in a range of 30°C to 34°C with servo-controlled heaters or other means can significantly reduce energetic requirements in critically ill infants.

Nutritional Support for the Critically Ill Child

Nutritional support for critically ill children is fundamentally different than conventional nutrition of healthy children because of the alterations in metabolism outlined previously. During periods of critical illness, utilization of nutrients for growth is markedly inhibited by hormonal response to stress and circulating inflammatory mediators. Utilization of calories for activity is much lower than under normal conditions. In addition, diet-induced thermogenesis is also affected in hospitalized patients by the different routes and formulations of nutrients provided. Estimates of increased caloric and protein requirements during acute illness and recovery indicate that compared to critically ill adults, children have greater requirements for both on a body weight basis. Therefore, one of the most important points for clinicians prescribing nutritional support is to provide calories in a thoughtful manner based upon the guidelines that follow and to avoid excess caloric intake during the acute phase of illness. During acute critical illness in children, many investigators have found REE to be less elevated than previously expected, with significant risk for overfeeding.^33,35

Assessment of Response to Nutritional Support

It is important to monitor the response to nutritional support. Intolerance of enteral support frequently manifests through abdominal distension, vomiting, or other physical signs. With TPN, intolerance manifests in iatrogenic derangements of minerals, electrolytes, and acid-base status. Hyperglycemia was discussed previously but may represent a complication of TPN administration. Standard nutrition assessment should be considered for each patient after the initial stress phase of critical illness. End-organ response to nutritional support is monitored by assessing whether serum transferrin or prealbumin is rising or falling and whether genuine weight gain is occurring in convalescing patients.

In some circumstances, a patient may not respond adequately to nutritional support and may benefit from a more detailed examination including the measurement of albumin, total protein, and transferrin, a 24-hour urine collection for nitrogen balance determination, and if possible, measurement of energy expenditure via indirect calorimetry. However, in general such a detailed and cumbersome approach has not consistently improved the status of the critically ill child. In circumstances in which clinicians believe the response to nutritional support could be improved, a consultation with a pediatric dietitian or gastroenterologist may be required. The use of total urinary nitrogen determination to assess nitrogen balance in critically ill children remains more useful in research studies than in practical patient care. Finally, indirect calorimetry can provide practical information regarding overall energy expenditure and substrate utilization when cardiopulmonary function is stable and lactic acidosis is not present; however, its utility in improving patient outcomes has not been confirmed.

Summary

Nutritional support of critically ill children is a central part of modern intensive care medicine. The complexity of pediatric disease and the wide range of nutrient options available necessitate close collaboration with dietary specialists who are familiar with children’s nutrition requirements during critical illness. It has become almost axiomatic in critical care nutrition that giving ever more nutrition will only produce undesired complications while not improving outcomes. Enteral nutrition will continue to be the preferred route of nutrition when it is tolerated and provides the more efficient and cost-effective means of transitioning patients to conventional dietary intake when critical illness has resolved.

Key Points