- •
Provision of individually tailored optimal nutrition is an important goal of pediatric critical care.
- •
Malnutrition is prevalent in the pediatric intensive care unit (PICU) and is associated with increased physiologic instability and resource utilization.
- •
Failure to accurately estimate or measure energy expenditure during critical illness may result in unintended underfeeding or overfeeding. Indirect calorimetry is the gold standard for energy expenditure assessment and helps guide energy prescription.
- •
Protein catabolism and nitrogen loss are characteristic features of the metabolic stress response to critical illness, resulting in net negative protein balance and loss of lean body mass. Accurate measurement of muscle mass and efforts to preserve it during critical illness may improve functional outcomes.
- •
Failure to deliver optimal energy and protein has been associated with poor outcomes in critically ill adults and children. The precise amount and timing of nutrient delivery during acute illness must be individualized and is an important area of investigation.
- •
Enteral nutrition (EN) is the preferred mode of nutrient delivery in patients in the PICU with a functioning gastrointestinal tract. Early EN has been associated with positive outcomes in critically ill patients.
- •
The gastric route is preferred for enteral nutrition. Postpyloric (small bowel) feeding may be considered for patients at risk of aspiration or when gastric feeding is not feasible or has not been tolerated.
- •
Use of EN algorithms and the presence of a dedicated dietitian in the PICU may decrease the barriers to EN and improve nutrient delivery.
- •
Parenteral nutrition (PN) is associated with mechanical, infectious, and metabolic complications and should be used in carefully selected patients for whom EN is contraindicated, not tolerated, or has failed to provide adequate nutrition. Delaying the initiation of PN and attempting EN delivery during the early phase of critical illness is prudent. Timing of PN initiation during acute critical illness may need to be individualized.
Malnutrition is prevalent in critically ill children at the time of admission to the pediatric intensive care unit (PICU). , Further nutritional deficiencies during their illness course are often incurred due to the burden of illness or suboptimal nutrient intake and may result in poor outcomes. Safe provision of optimal nutrients during hospitalization is an important goal of pediatric critical care. The prediction, estimation, and measurement of true energy expenditure in PICU patients can be challenging, resulting in unintended underfeeding or overfeeding. Although underfeeding has long been recognized as a problem, a significant proportion of critically ill children are at risk of being overfed. A number of barriers impede the delivery of prescribed nutrients to the critically ill child and result in a delay or failure to achieve the prescribed energy and protein goal. Although the complexities of critical care or the nature of illness frequently conflicts with nutrient provision, many perceived barriers to bedside nutrient delivery may be avoidable. This chapter will review some of the evidence-based and consensus statements around procedures for assessment of a critically ill child in the PICU and the approach to optimal and safe nutrient delivery in this group.
Malnutrition in the pediatric critically ill patient
Critical illness increases metabolic demand on the host in the early stages of the stress response, when nutrient intake may be limited (see also Chapter 80 ). As a result, children admitted to the PICU are at risk of deteriorating nutritional status and anthropometric changes with increased morbidity. This effect is more pronounced in a subgroup of patients who are already malnourished or at risk of malnutrition on admission. The prevalence of malnutrition in children admitted to the ICU has remained largely unchanged since the 1990s. One in every four children admitted to the PICU shows signs of acute or chronic malnutrition on admission. , Malnutrition is associated with increased physiologic instability and the need for increased quantity of care in the ICU. Despite its high prevalence and consequences, awareness of malnutrition is lacking. The nutritional status of hospitalized patients is often not routinely assessed, and only a minority of patients are referred for expert nutritional consultation. Careful nutritional evaluation upon admission to the PICU will allow identification of children who are at risk for further nutritional deterioration and, hence, candidates for early intervention to optimize intake. Efforts by organizations such as the American Society for Parenteral and Enteral Nutrition (ASPEN) have renewed the focus on malnutrition and facilitated a uniform approach to detecting and managing malnutrition in hospitalized adults and children. ,
Assessment of nutritional status
The recently revised definition of malnutrition extends beyond anthropometric thresholds to include etiology and pathogenesis of malnutrition and its impact on patient outcomes. Uniform use of reference charts and statistics have been proposed to compare individual patient measurements to the reference population and thereby classify malnutrition. Furthermore, the new definition of malnutrition accounts for its association with disease states, in particular inflammation. Assessment of the nutritional status of the critically ill child is vital but remains challenging. Clinicians use a combination of anthropometric and laboratory data to diagnose undernourishment. Carefully elicited past history—with details of weight gain/loss, dietary history, recent illness, and medications—helps identify risk factors and early indicators of malnutrition. Weight on admission to the hospital is important and may be the only measure of the actual dry weight before capillary leak syndrome results in edema and weight gain. Unless regular and accurate weights are obtained, acute changes in nutritional status may be missed or detected late. Physical examination should be directed toward specific signs of nutritional and metabolic deficiencies. Hair, skin, eyes, mouth, and extremities may reveal stigmata of protein-energy malnutrition or vitamin and mineral deficiencies.
A variety of other measurements—including arm anthropometry (mid-upper arm circumference and triceps skin fold), body length, and body mass index (BMI)—have been used to monitor growth in children. Recommendations for anthropometric variables and thresholds to classify malnutrition in children are shown in eTable 99.1 . Although bedside anthropometric methods are inexpensive, they may be prone to significant interobserver variability. Furthermore, weight changes and other anthropometric measurements in critically ill children should be interpreted in the context of edema, fluid therapy, volume overload, and diuresis. In the presence of ascites or edema, ongoing loss of lean body mass may not be evident using weight monitoring alone.
| Primary Indicators ( z -Score) | Mild Malnutrition | Moderate Malnutrition | Severe Malnutrition |
|---|---|---|---|
| Weight for height | −1 to −1.9 | −2 to −2.9 | −3 or lower |
| Body mass index (BMI) for age | −1 to −1.9 | −2 to −2.9 | −3 or lower |
| Length/height | No data | No data | −3 |
| Mid-upper arm circumference | −1 to −1.9 | −2 to −2.9 | −3 |
Body composition
Body composition may be the primary determinant of health and predictor of morbidity and mortality in children. The characteristic protein catabolism seen in the metabolic stress response, described in more detail later in this chapter, may cause significant alterations in body composition. Preservation and accrual of lean body mass during illness are important predictors of clinical outcomes in patients with sepsis, cystic fibrosis, and malnutrition. , Body composition is assessed in clinical practice by a variety of techniques, including anthropometry, bioelectrical impedance assessment (BIA) and dual-energy x-ray absorptiometry (DEXA). DEXA is a radiographic technique that can determine the composition and density of different body compartments (fat, lean tissue, fat-free mass, and bone mineral content) and their distribution in the body. However, DEXA is not practical for application in the PICU. BIA, in contrast, is a bedside technique that can be applied to pediatric patients without exposure to radiation and with relative ease. Electrical current is conducted by body water and is impeded by other body components. BIA estimates the volumes of body compartments, including extracellular water and total body water (TBW). TBW measures can be used to estimate lean body mass by applying age-appropriate hydration factors. BIA has not been validated in critically ill populations; hence, its application in the PICU outside clinical studies is currently being investigated. In specific populations, such as children with sepsis, BIA has been studied, along with other anthropometric measurements to estimate outcomes. Body impedance spectroscopy (BIS) uses the principle of frequency-dependent conduction of electric current through different tissues and employs a larger spectrum of currents to determine the composition of body tissues.
Biochemical assessment
The nutritional status can also be assessed by measuring the visceral (or constitutive) protein pool, the acute-phase protein pool, nitrogen balance, and resting energy expenditure. Visceral proteins are rapid turnover proteins produced in the liver. Low circulating levels of visceral protein are seen in the setting of malnutrition, inflammatory states, and impaired hepatic synthetic function. The reliability of serum albumin as a marker of visceral protein status is questionable. Albumin has a large pool and a half-life of 14 to 20 days, and it is not an indicator of the concurrent nutritional status. Serum albumin may be affected by changes in fluid status, albumin infusion, sepsis, trauma, and liver disease; these changes are independent of nutritional status. Prealbumin (also known as transthyretin or thyroxine-binding prealbumin) is a stable circulating glycoprotein synthesized in the liver. It binds with retinol-binding protein and is involved in the transport of thyroxine and retinol. Prealbumin, so named because of its proximity to albumin on an electrophoretic strip, is readily measured in most hospitals and is a good marker for the visceral protein pool. , It has a half-life of 24 to 48 hours and reflects more acute nutritional changes. Prealbumin concentration is diminished in liver disease. Acute-phase reactant proteins are elevated proportional to the severity of injury in response to cytokines released during stress response and have been used to longitudinally monitor the inflammatory response. Serum levels of acute-phase protein are elevated in children within 12 to 24 hours after burn injury due to hepatic reprioritization of protein synthesis. When measured serially, serum prealbumin and C-reactive protein (CRP) are inversely related (i.e., serum prealbumin levels decrease and CRP levels increase with the magnitude proportional to injury severity and then return to normal as the acute injury response resolves). In infants after surgery, decreases in serum CRP values to less than 2 mg/dL have been associated with the return of anabolic metabolism and are followed by increases in serum prealbumin levels. The relationship between inflammation and malnutrition is intriguing; the use of acute-phase inflammatory markers as well as cytokines and nutrition requires further exploration. Chemistry profiles should be monitored on admission and repeated periodically. Serum electrolytes, blood urea nitrogen, glucose, coagulation profile, iron, magnesium, calcium, and phosphate levels are routinely monitored. Adequacy of cellular immunity can be estimated through the measurement of total lymphocyte count and by delayed-type hypersensitivity testing with a series of common antigens (e.g., Candida, Trichophyton, tuberculin).
Metabolic consequences of critical illness
The energy burden imposed by the metabolic response to injury, surgery, or inflammation cannot always be accurately estimated, as it varies in intensity and duration between individuals. Importantly, nutritional support itself cannot reverse or prevent the metabolic stress response but may help offset the catabolic losses, particularly protein losses, during this state. Failure to provide optimal calories and protein during the acute stage of illness may exaggerate existing nutritional deficiencies and further exacerbate underlying nutritional status. Respiratory compromise involving loss of respiratory muscle mass, cardiac dysfunction and arrhythmias involving loss of myocardial muscle tissue, and intestinal dysfunction involving loss of the gut barrier contribute to the morbidity and mortality of critical illness. On the other hand, overestimation of this energy cost of metabolic stress may result in provision of energy in excess of requirement and is associated with poor outcomes. Hence, large energy imbalances attributable to underfeeding and overfeeding in critically ill children must be avoided. This can be prevented by individualized nutritional regimens that are tailored for each patient and reviewed regularly during the course of illness. In a trial of a 12-week individualized nutritional intervention in home-ventilated children, significant improvements were observed in respiratory and body composition variables. A basic understanding of the metabolic events that accompany critical illness and surgery is essential for planning appropriate nutritional support in critically ill children.
The unique hormonal and cytokine profile manifested during critical illness is characterized by an elevation in serum levels of insulin, glucagon, cortisol, catecholamines, and proinflammatory cytokines. Increased serum counterregulatory hormone concentrations induce insulin and growth hormone resistance, resulting in the catabolism of endogenous stores of protein, carbohydrate, and fat. Fig. 99.1 illustrates the basic pathways involved in the metabolic stress response. In general, the net increase in muscle protein degradation, characteristic of the metabolic stress response, results in a high concentration of free amino acids in the circulation. Free amino acids are used as the building blocks for the rapid synthesis of proteins that act as inflammatory response mediators and are used for tissue repair. Protein breakdown may continue for an extended period in an attempt to channel the amino acids through the liver, wherein their carbon skeletons are used for gluconeogenesis to produce glucose as the preferred energy substrate for the brain, erythrocytes, and renal medulla. Reprioritization of protein during metabolic stress increases the synthesis of acute-phase reactant proteins such as CRP, α 1 -acid glycoprotein, haptoglobin, α 1 -antitrypsin, α 2 -macroglobulin, ceruloplasmin, and fibrinogen. Plasma concentrations of other proteins, including transferrin and albumin, decrease with injury or sepsis. Overall, the intense protein catabolism during critical illness outstrips anabolism with net negative protein balance. Prolonged stress response may result in significant loss of lean body mass. The intense catabolism seen in metabolic stress cannot be suppressed by supplying calories, and negative protein balance continues relentlessly. This is one of the principal differences between stress response and starvation.
Starvation, or protein-calorie malnutrition, may be caused by socioeconomic, psychosocial, disease-related, or iatrogenic factors. The metabolic response to starvation involves decreased secretion of insulin and thyroid hormones, normal secretion of glucocorticoids and catecholamines, and decreased oxygen consumption. In starvation states, the body tries to preserve itself by using less energy for basic metabolic functions; thus, overall metabolic rate decreases. Metabolism shifts to use fat as a primary energy source, and the corresponding ketones help provide fuel for the brain and spare glucose and protein utilization. However, body tissues still must be broken down to supply amino acids for other critical functions, eventually leading to a loss of lean body mass, vital organ wasting, and possibly death. Unlike the stress response, muscle catabolism from starvation is reversed by supply of macronutrients. Table 99.2 summarizes the basic differences between starvation and metabolic stress.
| Metabolic Stress | Starvation | |
|---|---|---|
| ↑↑ BMR | ↑ | ↔↓ |
| Oxygen consumption (V o 2 ) | ↑↑ | ↓ |
| Protein catabolism | ↑↑↑ | ↔ |
| UUN | ↑↑ | ↔ |
| Weight loss | Rapid | Slow |
| LBM loss | Early | Late |
| Response to caloric intake | Protein catabolism continues | Protein catabolism halted |
| Insulin, cortisol, and catecholamines | ↑↑ | ↓ |
| Ketones | ↑↑ | ↔ |
| Gluconeogenesis | ↑ | ↓ |
Carbohydrate turnover is simultaneously increased during the metabolic stress response, with a significant increase in glucose oxidation and gluconeogenesis. The administration of exogenous glucose does not blunt the elevated rates of gluconeogenesis, however, and net protein catabolism continues unabated. A combination of dietary glucose and protein may improve protein balance during critical illness, primarily by enhancing protein synthesis. The stress response also stimulates lipolysis and results in increased rates of fatty acid oxidation. Increased fat oxidation reflects the premier role of fatty acids as an energy source during critical illness. Triglycerides in adipose tissue are cleaved by hormone-sensitive lipase into fatty acids and glycerol. Fatty acids are oxidized by β-oxidation in the liver generating acetylcoenzyme A for energy production in the tricarboxylic acid cycle and mitochondrial electron transport chain. As seen with the other catabolic changes associated with the stress response, the provision of dietary glucose does not decrease fatty acid turnover in times of illness. The increased demand for lipid use in the setting of limited lipid stores puts the metabolically stressed neonate or previously malnourished child at high risk for the development of essential fatty acid deficiency. , Preterm infants are most at risk for developing essential fatty acid deficiency after a short period of a fat-free nutritional regimen. , Nutritional therapy should support the metabolic changes occurring during the acute catabolic stage. With resolution of a hypermetabolic stress response, an anabolic phase typically follows, with increased release of growth hormone and insulin-like growth factor-1. Supply of adequate nutrition is essential for this recovery phase. In summary, the metabolic response to critical illness results in glucose and lipid intolerance and increased protein breakdown.
Underfeeding and overfeeding in the pediatric intensive care unit
Individual assessment of energy requirements and provision of optimal energy should be the standard of care in the PICU. Both underfeeding and overfeeding are prevalent in the PICU, with resultant nutritional deficiencies or excesses that are associated with complications. , True energy expenditure during acute illness may not be easily predicted; several studies have documented discrepancies in measured versus equation-estimated energy expenditure. Unless increased energy requirements during the acute stage of illnesses are accurately measured and matched by adequate intake, cumulative energy deficits will ensue with a decrease in weight, loss of critical lean body mass, and worsening of existing malnutrition. A variety of barriers, both unavoidable as well as some avoidable, exist that impede optimal nutrient delivery at the bedside and contribute to the likelihood of underfeeding in the PICU. , Underfeeding during acute illness with cumulative negative energy balance has been associated with poor outcomes in critically ill adults. In a large multicenter cohort study, delivery of a higher percentage of the prescribed energy goal was associated with significantly lower 60-day mortality. On average, percentage daily nutritional intake (enteral route) compared with the prescribed goals in this study was 38% for energy and 43% for protein ( Fig. 99.2 ).
On the other hand, overfeeding in the PICU is prevalent but may be underrecognized and have a potential negative impact on patient outcomes. Children do not predictably mount the characteristic hypermetabolic stress response as is seen in adults. The metabolic response to stress from injury, surgery, or illness is variable. The degree of hypermetabolism is unpredictable and unlikely to be sustained during a prolonged course in the PICU. Critically ill children cannot be presumed to be hypermetabolic following acute illness—energy expenditure may actually be decreased in some groups of patients. Children on extracorporeal life support or after surgery have failed to show any significant hypermetabolism; measured energy expenditure is close to resting energy expenditure in these populations. Critically ill children who are sedated and mechanically ventilated may have a significant reduction in actual total energy expenditure due to multiple factors. Stress or activity correction factored into basal energy requirement estimates in an attempt to account for the perceived hypermetabolic effects of the illness may result in overfeeding in hypometabolic patients. , Overfeeding can have deleterious consequences for the critically ill child, , such as net lipogenesis, hepatic steatosis, liver dysfunction, and increased carbon dioxide (CO 2 ) production resulting in difficulty with ventilator weaning. Investigators have proposed hypocaloric diets in critically ill adults. , Hypocaloric diets may have a protein-sparing effect and have demonstrable benefits in critically ill obese patients. However, it is uncertain whether administration of energy intake lower than the measured expenditure is appropriate for the pediatric patient. There is not enough evidence to recommend its general use in critically ill children. In general, the energy goals in critically ill children should be individualized and based on accurate and serial assessment of energy requirement during the illness course.
Current recommendations for nutritional requirements of the critically ill child are derived from limited data based on studies in healthy children and on limited methodological approaches. The components of total energy expenditure in children include (1) basal metabolic rate (BMR) 70%, (2) diet-induced thermogenesis (DIT) 10%, (3) energy expended during physical activity (PA) 20%, and (4) energy expended for growth. The sum of these components determines the energy requirement for an individual. These traditional components of energy expenditure in healthy children may not apply during critical illness ( eTable 99.3 ). Thus, prescribing optimal energy for the critically ill child requires careful review of each component of total energy expenditure. Recommendations for energy requirements were based on estimates of BMR or resting energy expenditure (REE) derived by either indirect calorimetry (IC) or standard equations. , REE estimates are unreliable with large individual variability, particularly in underweight, overweight, or critically ill children. , , Newer equations have attempted to improve the prediction of REE in children by accounting for weight-based groups or including pubertal staging, with variable success. , These equations have not satisfactorily been validated in critically ill children. The variability of the metabolic state may be responsible for the failure of estimation equations in accurately predicting the measured REE in critically ill children. As a gold standard, the guidelines recommend IC, when available, for the most accurate assessment of energy expenditure in critically ill children. When IC is not available, the Schofield or Food Agriculture Organization/World Health Organization/United Nations University equations may be used without the addition of stress factors to estimate energy requirement.
| Component | Normal Health | Critical Illness |
|---|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| REE + DIT + PA + growth |
|
Indirect calorimetry
The volume of oxygen consumed (V o 2 ) and the volume of CO 2 produced (V co 2 ) are measured by IC over a period of time. The V o 2 and V co 2 values are then used to calculate REE using the modified Weir equation: REE = [V o 2 (3.941) + V co 2 (1.11)] × 1440. This technique has been validated in healthy children by using a whole-body chamber to allow 24-hour measurement. For obvious reasons, the whole-body chamber cannot be used in critically ill children. The application of IC in different PICU populations has shown the variability in energy expended during illness. In the past, studies have demonstrated a relatively higher resting metabolic rate in critically ill children (37% higher than the resting metabolic rate of age-matched healthy controls). However, contrary to beliefs held for years, more recent studies have shown that the total energy expenditure is not increased in head-injured children, postoperative general surgical patients, or children after major cardiac surgery. The muted metabolic responses to major surgeries and injuries in studies may reflect advances in surgical and intensive care over the years. In critically ill mechanically ventilated children, use of sedation and muscle paralysis decreases the component of energy requirement related to physical activity. Thus, caloric needs in the critically ill child may be lower than previously considered. IC remains sporadically applied in critically ill children despite mounting evidence of the inaccuracy of estimated BMR using standard equations. This could potentially subject a subgroup of children in the PICU to the risk of underfeeding or overfeeding. However, IC application is not feasible in all patients due to (1) specific subject requirements, (2) device limitations, and (3) the need for expertise and resources. Table 99.4 describes some of the common problems associated with IC testing in critically ill children. In the era of resource constraints, IC may be applied or targeted for certain high-risk groups in the PICU. Selective application of IC may allow many units to balance the need for accurate REE measurement and limited resources (see eBox 99.1 for suggested criteria for targeted IC). In centers where IC is not available, the use of a simplified equation to determine REE based on V co 2 values alone was recently shown to be more accurate than estimating equations. Although IC application has illuminated our understanding of energy expended during critical illness, this is yet to be translated into improving patient outcomes. Studies examining the role of simplified IC technique, its role in optimizing nutrient intake, its ability to prevent overfeeding or underfeeding in selected subjects, and the cost/benefit analyses of its application in the PICU are desirable. The effect of energy intake on outcomes needs to be examined in pediatric populations, especially in those on the extremes of BMI.
| Error in V co 2 and V o 2 Measurement | Limitations or Mechanical Issues With the Device | Failure to Reach Steady State |
|---|---|---|
| Air leak >10% around endotracheal tube | High inspired Fi o 2 (>60%) | Recent interventions (suctioning, painful procedure) |
| Air leak in the circuit | Calibration issues | Fever, seizures, dysautonomia |
| Chest tube for pneumothorax | Moisture or obstruction due to water in the circuit | Recent change in ventilator settings |
| Study period too short |
Children at high risk for metabolic alterations who are suggested candidates for targeted measurement of REE in the PICU
- •
Underweight (BMI <5th percentile for age), at risk of overweight (BMI >85th percentile for age) or overweight (BMI >95th percentile for age)
- •
Children with >10% weight gain or loss during ICU stay
- •
Failure to consistently meet prescribed caloric goals
- •
Failure to wean or need to escalate respiratory support
- •
Need for muscle relaxants for >7 d
- •
Neurologic trauma (traumatic, hypoxic, or ischemic) with evidence of dysautonomia
- •
Oncologic diagnoses (including children with stem cell or bone marrow transplant)
- •
Children with thermal injury
- •
Children requiring mechanical ventilator support for >7 d
- •
Children suspected to be severely hypermetabolic (status epilepticus, hyperthermia, systemic inflammatory response syndrome, dysautonomic storms, etc.) or hypometabolic (hypothermia, hypothyroidism, pentobarbital or midazolam coma, etc.)
- •
Any patient with ICU LOS >4 wk may benefit from indirect calorimetry to assess adequacy of nutrient intake.
In summary, energy expenditure must be carefully evaluated throughout the course of critical illness, using actual measurements when available. In patients meeting the requirements for this test, IC provides an accurate measurement of REE. IC may need to be targeted to specific patient groups due to the risk of metabolic instability, but it may help prevent unintended underfeeding and overfeeding in these patients. In the absence of measured REE, equation=estimated REE may be used. However, the uniform application of stress factors is not advisable and must be used only in individual cases after careful evaluation. Once energy needs are determined, the optimal substrate required for maintenance of energy needs is then administered as mixed fuel with glucose and fat; the proportion of each varies according to the clinical situation.
Protein requirements
Protein turnover and catabolism are increased severalfold in critically ill children. This is one of the most characteristic features of the metabolic stress response and probably represents an adaptive response that was critical to survival of prehistoric humans during periods of illness or injury prior to the agricultural age. The continuous flow of amino acids from protein breakdown allows for maximal physiologic adaptability at times of injury or illness. Specifically, this process involves a redistribution of amino acids from skeletal muscle to the liver, wound, and other tissues involved in the inflammatory response. Although this is an excellent short-term adaptation, it is ultimately associated with morbidity because of the limited protein reserves available in children and neonates. Although children with critical illness also have increases in whole-body protein synthesis, it is the whole-body protein degradation that predominates during the stress response and results in a net negative protein balance. , The role of protein intake in offsetting these losses by improving protein balance and eventually improving patient outcomes is being investigated. In a large international cohort study of more than 1200 mechanically ventilated children, 60-day mortality was lower in children who received a higher proportion of their daily prescribed protein goal. Fig. 99.3 depicts the linear relation between enteral protein intake adequacy (percentage of the protein goal delivered) and 60-day mortality in mechanically ventilated patients in this cohort. A similar association between lower protein intake and mortality was also shown in a large international study of adult critically ill patients. These reports suggest a strong association between protein intake and outcomes, which is both biologically plausible and independent of energy intake.
