Parenteral and Enteral Nutrition in the Intensive Care Unit

David F. Driscoll

Bruce R. Bistrian

Nutritional and metabolic support during acute illness is an integral part of the clinical care of critically ill patients. The significance of such interventions is predicated on three main factors: (a) degree of metabolic stress; (b) dysfunction of major organ systems; and/or (c) presence of protein-calorie malnutrition (PCM). In the first case, metabolic stress can arise from a variety of sources including, for example, severe injuries sustained by major trauma such as closed head injury, multiple long-bone fractures, third-degree burns of greater than 25% body surface area, and severe sepsis and stress of lesser intensity such as thoracoabdominal surgery, pulmonary infection, systemic infection, or any source of active systemic inflammation. Often, more than one form of metabolic stress may be present that can accentuate and/or dysregulate the injury response. Concerning the second factor, metabolically stressed patients may develop acute failure of vital organs during the critical care period or have underlying chronic end-organ dysfunction. Acute or chronic disease, particularly of the cardiopulmonary, renal, or hepatic system, often further complicates the clinical course and requires modification of nutritional support during critical illness, especially in the elderly [1]. Finally, the presence of preexisting or the likely early development of PCM is key to identifying those patients who will derive the greatest clinical benefits from nutritional and metabolic support therapy.

Approximately 35 years ago, the prevalence of PCM in hospitalized general medical and surgical patients was reported to be as high as 50% of all adult admissions to a large teaching hospital [2,3]. More recent reports continue to document high rates of malnutrition in hospitalized patients [4,5,6,7,8,9]. When moderate to severe PCM accompanies severe metabolic stress, an increase in nutrition-related complications can be expected to occur, including wound dehiscence, nosocomial infections, and severe fluid, electrolyte, and acid–base disturbances. During stress, substantial catabolism of both endogenous and exogenous protein and energy occurs coincident with the injury response. In support of the metabolic response to injury, the breakdown of body protein, principally from muscle and connective tissue stores, supports amino acid and energy needs to mount various beneficial components of the systemic inflammatory response by the release of amino acids for accelerated synthesis of such proteins as leukocytes, hepatic acute phase and cellular proteins, and wound tissue, and gluconeogenesis for the optimization of energy requirements for such tissues as cardiac, leukocytes, and fibroblasts. An assessment of the degree of this response can be estimated by application of the catabolic index [10]. However, if protein calorie malnutrition complicates injury or infection, the systemic inflammatory response is less intense than that found in normally nourished individuals with a similar degree of injury. Consequently, the degree and duration of the metabolic response, with respect to nitrogen breakdown, is greatly diminished. In terms of the degree of catabolism, for example, a malnourished elderly patient with significant catabolic injury could manifest nitrogen losses that may be as a much as 50% less than normally nourished younger counterparts with the same injury [1]. Although this might imply a less severe catabolic response sparing lean tissue, the pathologic consequences are more severe as a result of the muting of the beneficial aspects of the systemic inflammatory response, and these adverse effects tend to occur sooner. Moreover, the time course to intervene with nutritional and metabolic support to limit the likelihood of nutrition-related complications is also shortened by as a much as 50% (i.e., 5 to 7 days) in the moderate to severely malnourished versus normally nourished individuals (i.e., 7 to 10 days) with the same metabolic stress. Ultimately, the consequences of ongoing depletion of the metabolically active body cell mass in the malnourished reduce the ability to recover from acute illness, can be associated with severe deficiencies in minerals that are typically found in muscle (potassium, magnesium, and phosphorus), and often lead to severe impairments in immunocompetence, wound healing, and organ repair.

Once the decision to provide nutrition support is made, parenteral or enteral nutritional therapies are available options. In every case, if the gastrointestinal tract is functional and the patient is hemodynamically stable, enteral nutrition (EN) should be instituted. However, if significant malnutrition also exists and a prolonged recovery is anticipated, it should be recognized that the time frame to achieve eucaloric intakes for EN often takes much longer due to associated gastrointestinal intolerance, compared with parenteral nutrition (PN). As central venous access is generally necessary during critical illness, EN support can often be supplemented with PN [11] so as to avoid the prolongation of caloric deficits during acute illness, which are particularly of concern in initially malnourished patients or the most critically ill with closed head injury, multiple trauma, major burns, and severe sepsis. In such patients it appears that early feeding within the first 72 hours, whether by enteral, parenteral, or the combination, has the greatest impact on outcome in terms of mortality. Although mild decrements in energy balance in the critical care setting may well be tolerated and in certain circumstances appropriate, at least 1 g of protein per kg and 15 kcal per kg advancing to 1.5 g protein per kg and 20 to 25 kcal per kg as soon as possible should be the goal to avoid adverse, nutrition-related outcomes. Moreover, intensive metabolic support (i.e., the provision of electrolytes and acid–base therapy) can also be accomplished efficiently through the PN admixture. The amount of parenteral nutrients can be gradually reduced as the patient is transitioned to EN coincident with remission of the stress response and return of full gastrointestinal tolerance to tube feeding. Thus, in the intensive care unit (ICU), nutrition support is often provided to patients using both enteral and parenteral means, especially during the acute care period. The purpose of supplying both EN and PN where appropriate should not be motivated by attempts to meet protein and energy needs as soon as possible, but rather as a means of providing trophic stimulation to enterocytes and hopefully a quicker transition to full enteral feedings, while PN is used to treat severe metabolic disorders such as hypokalemia, hypophosphatemia, and metabolic alkalosis, that can only be safely and effectively addressed by the
intravenous route of administration. The greatest challenge facing the critical care clinician is to appropriately identify those patients who are in greatest need of nutrition support therapy and to provide it in a manner that is both effective and does not produce iatrogenic complications.

Clinical Consequences of Delaying Nutrition Support

Although at times it is difficult to pinpoint the cause and effect of nutrition-related complications during critical illness, it should be intuitively obvious that withholding nutrition will ultimately lead to death from starvation. This message was poignantly illustrated in the deaths of Maze prisoners in Belfast, Ireland, as detailed in a report from Leiter and Marliss [12] in 1982. Ten Irish Republican Army prisoners went on a hunger strike that led to their deaths over a period of 45 to 73 days of fasting. All were young lean males and the critical weight loss that resulted in death was approximately 35% calculated from the first day of the fast. It is also generally acknowledged that patients who approach 35% to 40% losses from their ideal or usual body weight through inadequate nutritional intake are at greatest risk of malnutrition-related death. Presumably, at these extreme levels of body mass depletion, both the size and function of vital organs of the viscera are considerably diminished. At some critical point, presumed to be when fat stores become limited, protein catabolism now coming from both skeletal and visceral organs accelerates. If one discontinues providing life-sustaining needs for energy, the loss of a critical mass of body protein is ultimately reached and death from organ failure is imminent.

The effects on the vital organs can be catastrophic, since oxygen consumption of the visceral organs is much higher than that of resting skeletal muscle. The imbalance between loss of skeletal muscle and visceral organ mass initially favoring visceral organs has also been suggested to explain the higher energy expenditures per body weight seen in severely depleted hospitalized patients (average of approximately 70% of ideal body weight) as a result of an approximate 10-fold difference in resting oxygen consumption between skeletal muscle compared to visceral tissues such as the liver [13]. During starvation (with adequate water intake), and in the absence of metabolic stress, a normally nourished, thin individual can survive for periods of approximately 6 to 10 weeks. In terms of total body nitrogen, it is estimated that the loss of 350 to 500 g of nitrogen is potentially lethal. In terms of body mass index (BMI), which is weight in kg per height in meters squared, it is generally considered that a BMI less than 13 kg per m² in males and less than 11 kg per m² in females is incompatible with life [14]. However, the rapidity of weight loss is also a factor, since lesser degrees of semistarvation (i.e., smaller energy deficits) are better tolerated. Table 191.1 depicts the relationship of BMI with nutritional status.

By way of comparison, the metabolically stressed patient experiences greater catabolism coincident with acute illness and can lose as much as 30 g of nitrogen per day, representing about 1 kg of lean tissue from the breakdown of lean body mass. Generally, the majority of these losses can be measured in a 24-hour urine collection as urea nitrogen and used for nitrogen balance estimation. Nitrogen balance studies assess the difference between dietary protein (nitrogen) intake and nitrogen excretion. Healthy individuals consuming an adequate diet in terms of essential nutrients including protein (0.8 g protein per kg per day) and sufficient energy to provide energy balance will be in zero nitrogen balance. That is the nitrogen in is equaled by the nitrogen out in urine (mostly) and feces, reflecting no net change in lean body mass. Net nitrogen losses in patients receiving parenteral or enteral feeding can vary from 0 to 30 g per day, depending on the extent of the injury response and the level of feeding. With the systemic inflammatory response, the utilization of protein to maintain lean body mass is impaired, making the daily requirement increase to about 1.5 g protein per kg per day. Similarly, energy requirements increase, which are offset to some degree by the reduction in physical activity characteristic of the hospitalized patient. With the development of renal dysfunction, the proportionate amounts of nitrogen found in the urine become substantially less, with a concomitant rise in blood urea nitrogen (BUN). In general, in a 70-kg male every 5 mg% change in BUN represents 2 g of nitrogen catabolized and not excreted, and 1.5 g of nitrogen for a 60-kg female, based on average total body water of 60% and 50% for males and females, respectively. Protein intakes must be adjusted to limit the rise in BUN, but nutrition efficacy should not be sacrificed to renal function beyond a reduction to the 1 g protein per kg for other than very brief periods. Renal replacement therapy such as dialysis or hemofiltration should be considered in those circumstances. Once the BUN becomes stable, even if elevated by impaired renal function, a 24-hour urine urea nitrogen excretion represents the amount catabolized over that period. The catabolic index (CI) (CI = 24-hour urine urea nitrogen – [0.5 × dietary nitrogen + 3]), adjusts for the effects of dietary intake and obligatory nitrogen loss on urinary urea nitrogen excretion. The catabolic index is the difference between measured and predicted urine urea nitrogen excretion. For example, the major catabolic stresses that produce the highest nitrogen losses and catabolic indices include burns, head injury, severe sepsis, and multiple trauma. The clinical application of nitrogen balance and CI assessments are illustrated in Table 191.2.

Table 191.1 Body Mass Index and Nutritional Status

Body mass index	= weight in kg ÷ (height in m)²
Assumptions:	weight: 75 kg; height: 1.84 m
BMI	= 75/(1.84)²
	= 22.2
Body mass index	Nutritional status
≥30	Obese
≥25 – < 25	Normal
< 18.5	Moderate malnutrition
< 16	Severe malnutrition
< 13	Lethal in males
< 11	Lethal in females

There are potential clinical scenarios that may affect the accuracy of nitrogen balance studies. This is especially true in patients with renal dysfunction that may reduce nitrogen output and could erroneously suggest an improvement in nitrogen balance. A correction of the nitrogen balance study can be applied to account for the nitrogen losses that do not appear in the urine, but result in an increase in the BUN concentration. Assuming nitrogen intake remains constant, two important pieces of data are required to correct for the nitrogen losses not appearing in the urine and include the patient’s BUN and body weight at the beginning and end of the 24-hour collection period. These are important because most of the urea is distributed in total body water. A clinical example that applies to this method of correction appears in Table 191.3.

In terms of lean body mass, each gram of nitrogen lost represents approximately 30 g of (hydrated) lean tissue (hydration ratio: approximately 4 or 5 to 1). For patients with daily
nitrogen losses of 30 g, which represents the highest catabolic nitrogen loss in the absence of dietary protein intake, approximately 1 kg of lean tissue would be lost each day. Such losses cannot be sustained for protracted periods, and under these circumstances, nutrition support is clearly indicated within the first 24 to 36 hours even in the previously well-nourished patient to address this extraordinary rate of loss. Using cumulative nitrogen deficits of 350 to 500 g, a sustained loss of this magnitude could theoretically result in death in approximately 2 to 3 weeks, although catabolic rates usually diminish in the later weeks of injury. For the severely malnourished patient of 75% ideal body weight, one can estimate the critical survival period to be in the range of 1.5 to 2 weeks under the same circumstances. Finally, a cumulative caloric deficit of 10,000 kcal or more during acute illness has been associated with significant morbidity and mortality in the surgical ICU [15]. However, it is likely that the associated protein deficit played the larger role, since normal individuals have more than 150,000 stored calories as fat, which always makes up the greater proportion of the caloric deficit. A study in the medical ICU has shown that intakes of less than 25% of requirements were associated with a higher rate of bloodstream infections [16].

Table 191.2 Clinical Application of the Nitrogen Balance and Catabolic Index Assessments^a

Table 191.3 Nitrogen Balance Correction in Renal Dysfunction^a

Of course, projections of survival or complications are estimates and may be highly variable depending on other factors (i.e., nutritional status, metabolic stress[es], end-organ function, and so forth). Moreover, in the clinical setting, such high outputs of nitrogen over long periods will not likely be sustained, as medical and surgical therapies will usually be successful in reducing the stress response. Furthermore, both the rate of reduction in lean body mass and the intensity of the systemic inflammatory response diminish as PCM develops. Such patients will invariably receive calories (dextrose) and electrolytes from various parenteral infusions, so that some form of supplementation is given, which also slows the loss of lean tissue. Consequently, the outcome of death from the total lack of nutrition support is rare. However, nutrition-related complications, such as impaired wound healing and immunocompetence leading to nosocomial infection, are the common proximate causes of increased morbidity and mortality under such circumstances.

Identifying Patients in Need of Nutrition Support

In the ICU setting, it is often difficult to identify those patients who are at greatest risk of developing nutrition-related
complications due to preexisting malnutrition. Such patients are often volume overloaded due to massive administration of parenteral fluids from multiple drug therapies and often acute volume resuscitation, as well as maintenance intravenous therapy to support intravascular volume. This fluid retention and weight gain is often compounded by the hormonal consequences of the systemic inflammatory response such as enhanced insulin, aldosterone, and antidiuretic hormone secretion, which favor salt and water retention. Consequently, the weight of the patient is artificially high, and major efforts of the ICU team are often directed at reducing volume intake in order to mobilize third-space fluids. A weight history may be difficult to obtain or overlooked entirely because of more acute clinical issues. Moreover, an accurate patient weight is also important to optimize drug therapy. Under these circumstances, a moderately to severely malnourished patient may escape detection by the primary care team, and only be recognized as malnourished after fluid homeostasis is achieved, or worse, a potentially preventable nutrition-related complication, such as wound breakdown, occurs. Clearly, at this point, the opportunity to minimize such complications from expert nutrition support has passed, and the course toward rehabilitation may be long and costly.

To avoid this scenario, a more substantial effort must be undertaken to identify the patients at greatest risk. Nutrition screening programs on admission, especially by dieticians, can greatly assist in identifying these patients, but many patients, especially acute admissions for emergent care, may escape this surveillance process. In these cases, the premorbid weight is very important and should be obtained if at all possible. It will at least provide a baseline prior to the numerous medical and surgical maneuvers that may take place over the ensuing 24 to 48 hours that could dramatically change the patient’s weight in the critical care setting.

If the admission weight is not obtained, then the clinician may need to estimate the patient’s body weight from available hospital data. Estimations may be made based on the most recent weight recorded, and then backtracked through the medical chart using the intake and output records to reconstruct the original weight history. For critically ill patients, such records are usually reliable, and a reasonable estimate may be made. This estimate may be confirmed by subsequent discussions with the patient or family. When confirmed, the body weight can then be compared to standard measures for population-based body weight for height tables such as the ideal body weight or the BMI. A patient weight less than 85% of the ideal body weight (IBW) or BMI less than 18.5 indicates moderate malnutrition. Severe malnutrition would be considered likely if weight is less than 75% of IBW or BMI is less than 16 kg per m². Thus, a greater sense of urgency to intervene with nutrition support is present under these conditions and should be undertaken within several days of the acute injury. If the patient is deemed well nourished, then intervention may be delayed unless the systemic inflammatory response is severe (i.e., major third-degree burns, closed head injury with a Glasgow Coma Score less than 8, multiple trauma with very high acute physiology and chronic health evaluation [APACHE] or injury severity scores, severe pancreatitis with a positive CT scan and more than three Ranson criteria, and so forth). Then, because the systemic inflammatory response is likely to endure beyond 1 week, very early nutritional support is indicated. The serum albumin level, which reflects the presence of a recent systemic inflammatory response, is not often helpful in this setting because the invariable systemic inflammatory response and common disturbances in volume status make hypoalbuminemia universal. However, severe hypoalbuminemia (less than 2.4 g per dL) usually reflecting a greater degree and/or longer duration of systemic inflammation identifies a population at much greater nutritional risk. Finally, if the weight-based data are not reliable, a formal nutrition support consult or indirect calorimetry may be indicated.

Nutritional Requirements

Protein

The amount and type of protein administered to the critically ill depend on the clinical circumstances of each patient. Nevertheless, there is an upper limit to the quantity of protein that can be given based on net protein utilization during metabolic stress. In general, providing protein in amounts above 1.75 g per kg per day exceeds the capacity of the body to use the administered protein to increase synthesis [17,18]. Amounts above this level of intake are essentially completely converted to urea and serve no nutritional purpose. At intakes ranging between 0.6 to 1.75 g per kg per day, each increment of intake increases net protein synthesis at a cost of increasing the proportion going to ureagenesis. In patients with nitrogen accumulation disorders (of either renal or hepatic origin), a compromise must often be made between greatest rates of net protein synthesis and lowest rates of urea or ammoniagenesis. For example, as the BUN increases, especially above 100 mg%, the risk of uremic complications increases, including bleeding, or, increasing the production of ammonia in encephalopathic patients. Generally, the optimal protein intake in critically ill patients is given at twice the recommended daily amount (approximately 0.8 g per kg per day) of normal adults, at approximately 1.5 g per kg per day. With renal impairment, at least 1 g per kg should be provided and greater amounts given if tolerated or dialysis is initiated. In patients with liver failure at least 1 g per kg of standard protein should be provided and up to 1.5 g per kg if tolerated. This is done recognizing the overall impairments in protein utilization that accompanies metabolic stress, as well as the heightened needs during catabolism.

The type of protein administered varies with the patient’s condition and the route of administration. For PN support, standard protein mixtures are given in their monomeric form as individual crystalline amino acids and levorotatory isomers, which comprise the essential amino acids (histidine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan, and valine) and the nonessential amino acids (alanine, aminoacetic acid, arginine, cysteine, proline, serine, and tyrosine). In standard amino acid formulations, the branched-chain amino acids (leucine, isoleucine, and valine) comprise approximately 18% to 25% of the amino acid profile. Collectively, they are available in commercial intravenous solutions in concentrations ranging from 3% to 15%. On average, for every 6.25 g of the amino acids in the mixture, 1 g of nitrogen is available, although this number is lower with a number of the specialized amino acid formulas. The caloric value of protein is 4.1 kcal per g, and such calories should be counted in critically ill patients when tracking energy intakes.

Specialized amino acid mixtures have evolved that include selected profiles. For example, renal formulations have been devised that principally provide the essential amino acids (histidine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan, and valine), while hepatic formulations have eliminated or reduced aromatic amino acids (phenylalanine, tryptophan, and tyrosine) and the sulphur-containing amino acid methionine and increased the proportion of branched-chain amino acids (BCAAs) (isoleucine, leucine, and valine). However, the routine use of these expensive formulations in these conditions over conventional or standard amino acid mixtures has not been convincingly demonstrated, and in certain cases when used to meet full protein needs, may be harmful [19]. For patients with nitrogen accumulation
disorders, the use of branched-chain enriched amino acid formulas in the range of 45% to 50% of the total amino acid profile have been shown to improve protein utilization when total amino acid intakes are given in the 40- to 70-g range and may reduce the risk of encephalopathy when compared to a standard formula. Finally, other attempts at modifying the profiles of amino acid mixtures, such as the extemporaneous preparation by the hospital pharmacy of sterile glutamine in total parenteral nutrition (TPN), have shown some benefits in selected settings but they require a considerable level of parenteral compounding expertise. In addition, in order to safely provide this compounded sterile preparation, ongoing quality assurance measures as outlined by the United States Pharmacopeia must be performed and therefore such practices are subject to Federal Drug Administration oversight [20]. A glutamine-containing dipeptide formulation, which is commercially available in Europe, has been the subject of some positive trials, but its ultimate place in the care of the critically ill is not yet established.

For EN support, protein is typically provided in either an oligomeric form as protein hydrolysates containing various peptides ranging from di- and tripeptides to polymers of eight or more, or as whole protein usually provided as casein or in its polymeric form as, for example, casein hydrolysates. Less commonly, they can even be provided as the individual amino acids. Most formulations contain a fixed amount of protein in the range of 30 to 40 g per L and thus for fluid-restricted patients in the ICU cannot meet the protein needs of most patients. Alternatively, more concentrated enteral formulae exist that may be used, or the clinician may opt to add protein modules to conventional products to increase protein density. However, in either case, both approaches result in higher osmolarities that may affect gastrointestinal tolerance.

Carbohydrate

The amount and type of energy provided to improve the utilization of the prescribed protein intake also varies with the individual patient. As well, there are physiologic limits to the amounts given, beyond which significant complications are more likely. For most patients, providing 25 kcal per kg per day is sufficient to support the protein synthetic response to metabolic stress. This is the total energy expenditure of most critically ill, postoperative patients. Amounts above 30 kcal per kg per day exceed the energy expenditure of most hospitalized patients except those with severe burns, closed head injury, and multiple trauma where measured caloric expenditures are usually 30 to 40 kcal per kg. However, providing nutritional support in amounts greater than 30 kcal per kg leads to higher rates of hyperglycemia in both types of patients; in the postoperative setting, due to overfeeding, and in the trauma unit, due to the severity of systemic inflammatory response. Although better glycemic control through the use of insulin would be one way to reduce the infectious risk in the latter instance, it is interesting to note that in several trials of immune-enhancing diets that improved outcomes and reduced infection rates have been seen at energy intakes at 30 kcal per kg or less, in diets that are likely to have been hypocaloric [21]. For carbohydrates, the physiologic limits are linked to the normal endogenous hepatic production rates for glucose, which approximate 2 mg per kg per minute or about 200 g per day for a 70-kg healthy adult [22]. This is the amount of glucose needed by the body to meet the obligate needs of tissues dependent on glucose (i.e., brain, renal medulla, red blood cells, and so forth), and it is derived from body stores of glycogen (glycogenolysis) or made from noncarbohydrate sources such as from protein breakdown to gluconeogenic amino acid precursors (gluconeogenesis). Glycogen stores are limited and therefore can be rapidly depleted during acute metabolic stress (i.e., within 24 hours) [23]. Thus, the major source of glucose in the hypocaloric state following stress comes from gluconeogenesis, and higher amounts than usual are produced to support the metabolic response to injury, accelerated by the hormonal milieu produced by the increased secretion of catecholamines, glucagon, cortisol, and growth hormone [24]. The judicious provision of nutrition support is designed to attenuate the extent of protein breakdown without exacerbating significant changes in nutritional and metabolic homeostasis. Similar to the case with protein, as carbohydrate intake increases net oxidation occurs, but with an increasing proportion going to nonoxidative pathways (glycogen synthesis and particularly de novo lipogenesis). However, glycogen synthesis is limited by available storage capacity of about 500 g in normal adults and perhaps 1,000 g in a critically ill patient receiving TPN, with its resultant very high insulin levels. There is effectively no limit for fat storage. The optimal balance is at intakes at about 400 g per day, with maximal glucose oxidized of 700 g per day. Thus, in a 70-kg adult, glucose to amino acid of 2:1 TPN formula providing glucose at 400 g per day and 1.5 g of protein per kg per day represents about 25 kcal per kg per day.

For PN, glucose is the only reasonable carbohydrate fuel or energy source that is widely available for intravenous administration. Generally, it is provided as a monohydrate and its caloric equivalent is therefore 3.4 kcal per g rather than 4 kcal per g for its anhydrous form. It is commercially available in a variety of concentrations ranging from 2.5% to 70% in sterile water for injection. Glucose is the primary energy source of any PN admixture prescribed for central venous alimentation and typically is given in final admixture concentrations from 10% to 25%. Higher concentrations can be given, but are associated with an increase in the number of dextrose-associated complications if the amounts given are too large.

For EN, carbohydrates may be given in a number of chemical forms. For example, they can be given as the monosaccharide, glucose, frequently found in monomeric or elemental formulas. Alternatively, in less refined formulas, carbohydrates may be provided as oligosaccharides, such as hydrolyzed cornstarch, or more complex polysaccharides, such as corn syrup, are frequently used. The selection of a particular enteral formula is largely based on a number of clinical factors such as gastrointestinal function, fluid status, and end-organ function.

Fat

Lipids serve as an alternative energy source that is used to substitute for a portion of the carbohydrate calories. PN support prescribed in this fashion, it is referred to as a total nutrient admixture, all-in-one or 3-in-1 mixed-fuel system [25]. As with protein and carbohydrates, the amount and type of lipids used will vary depending on the clinical condition of the patient. For the most part, long-chain triglycerides (LCTs) derived from vegetable oils have been the principal source of lipid calories used in the clinical setting. Specifically, soybean oil, which is rich in polyunsaturated omega-6 fatty acids, has been extensively used, especially for intravenous nutrition. It is a major source of the essential fatty acids, linoleic, and alpha linolenic acids. However, ill-considered prescribing habits, where either excessive quantities or infusion rates have been used, have led to clinically significant adverse effects such as immune dysfunction and pulmonary gas diffusion abnormalities in critically ill patients. The excessive administration of intravenous lipid emulsions (IVLE) can accumulate in the liver and impair Kupffer cell function, thus interfering with a major component of the reticuloendothelial system [26,27]. In addition, lipid injectable emulsions are composed of various oils that serve as prostaglandin precursors that are immunosuppressive,
especially those of the n6 series such as PGE₂, which suppresses lymphocyte proliferation and natural killer cell activity [28], and can reverse hypoxic vasoconstriction in patients with adult respiratory distress syndrome [29]. In contrast, the oxidation and subsequent plasma clearance of lipids is significantly improved when IVLEs are given over 24 hours versus briefer intervals [30]. Impaired plasma clearance of lipids can result in fat overload syndrome and is a particularly significant clinical issue in children [31,32,33,34,35,36,37,38,39,40,41,42]. Fat overload syndrome can result from the administration of a stable fat emulsion over brief intervals [29,30,43,44,45,46,47] or from more modest doses of lipid that might be physicochemically unstable [48]. In fact, a review of the literature regarding stable fat emulsions has concluded that virtually all of the adverse effects associated with LCTs have occurred when the infusion rate exceeds 0.11 g per kg per hour [49]. For a 70-kg adult this limit would be approximately 13 hours for 500 mL of 20%, which makes 3-in-1 admixture infusions safer and easier to administer as a continuous infusion over 24 hours rather than as a separate “IV piggyback” over a brief period, which would require an infusion rate almost twice as fast. In addition, piggyback infusion of lipids is not recommended beyond 12 hours [50].

Recent reports regarding the clinical significance of unstable fat emulsions have emerged. On December 1, 2007, the United States Pharmacopeia (USP), which is recognized by the Food and Drug Administration (FDA) as the official compendium for drug standards, was the first pharmacopeia worldwide to establish globule size limits for intravenous lipid emulsions [51]. This is notable because intravenous lipid emulsions had been used clinically in the United States for more than 30 years (and Europe for more than 45 years), when most drugs have official USP specifications within 5 years of FDA approval [52]. The USP limits specified two size limits: (i) mean droplet size (MDS < 0.5 μm) and (ii) large diameter tail, expressed as the percent of fat globules > 5 μm (PFAT₅ < 0.05%). The primary motivation for these limits was to avoid the development of microvascular pulmonary embolism from an excessive population of large-diameter fat globules indicating instability of the emulsion.

Around the time the USP announced its intentions to adopt these limits in 2004 [53], a major lipid emulsion product also changed its conventional packaging from glass to plastic containers. With this change in packaging, the lipid emulsion product now failed the large-diameter globule limits of the USP [54]. Lipid emulsions failing USP limits were also shown to produce less stable emulsions when packaged in syringes for neonates [55], when mixed in TPN admixtures [56] and when used in a multichamber bag premixed for TPN therapy [57]. Moreover, lipid emulsions not meeting pharmacopeial limits were also shown to be associated with significant hypertriglyceridemia in premature neonates when compared to lipid emulsions meeting USP limits [58], although this has not been confirmed in a randomized clinical trial. Finally, in animal studies lipid emulsions failing USP limits were shown to be hepatotoxic [59]. A recent study intended to explore the extent of physiologic damage from the infusion of unstable lipid emulsions produced evidence of hepatic accumulation of fat associated with oxidative stress, liver injury and a low-level systemic inflammatory response [60].

Triglyceride clearance is maximal at serum triglyceride levels of up to about 400 mg per dL, and patients who initially have serum triglycerides at this level will tolerate even lesser amounts of fat without adverse consequences. In patients who have normal serum triglyceride levels at initiation of TPN, serum triglyceride levels are usually not monitored. For those with levels greater than 200 mg per dL it is reasonable to check the triglyceride again after a stable regimen has been attained with lipids below 0.11 g per kg per hour. Stable levels below 400 mg per dL are acceptable while receiving lipid emulsions.

For PN therapy, soybean oil emulsions continue to dominate the United States market. However, there are a number of different lipid compositions presently available in Europe and under investigation in the United States [61]. They include various mixtures of soybean oil with medium-chain triglycerides (MCTs), olive oil, and fish oil. In nearly every case, soybean oil is included in sufficient proportions to provide adequate amounts of the essential fatty acids [62].

For EN therapy, a number of the lipid types available for parenteral use in Europe are widely available in the United States for enteral administration in complete nutritional diets. Typically, they contain 30% to 40% of the total calories as fat and often contain blends of corn and soy oil. However, in the more specialized enteral formulas, MCTs, fish oil, and even structured lipids are available. Moreover, in some of these products the fat content is either severely restricted (i.e., 3% to 10% of total calories for the fat-intolerant patient) or may be as high as 55% for the patient with pulmonary compromise.

Volume

The maintenance of fluid homeostasis is an important goal in critical care. At times, many patients in the ICU become severely volume-overloaded as a consequence of parenteral fluid administration and the fluid-retentive state characteristic of critical illness [63,64,65]. For this reason, when assessing fluid status, it is important to bear in mind the usual contribution of water to body weight or total body water (TBW) of the patient under normal, unstressed conditions. In normal adults, TBW comprises approximately 50% to 60% of body weight. As lean body mass is hydrated in a ratio of approximately 4 parts water to 1 part protein, lean tissue is a significant component of TBW. In the clinical setting, acute changes in weight over short intervals primarily reflect net changes in TBW which almost never reflect lean tissue gains in the hospital setting. For example, a 10% increase in weight over 24 to 48 hours represents a proportional increase in TBW and may be associated with adverse clinical consequences, such as greater ventilator dependence, impaired cardiovascular function, and disturbances in electrolyte homeostasis. Even when the patient is considered euvolemic, the contributions to volume from nutrition support are generally limited to approximately 25 mL per kg per day, as other reasons for fluid administration are usually indicated.

Depending on the volume assessments by the primary care team, the amount of nutrition support that may be provided by either PN or EN may be affected. The most significant effect occurs when volume restrictions are imposed. When this happens, hypocaloric nutrition is usually provided due to the limitations associated with caloric density. Caloric or macronutrient density is the sum total of calories from protein, carbohydrates, and fat, expressed in kilocalories per milliliter (kcal per mL). Generally, the caloric density of typical formulations routinely prescribed for either PN or EN support is approximately 1 kcal per mL, but special forms of each therapy are available that reasonably allow up to 1.5 kcal per mL to be formulated. However, most enteral formulations are commercially available in fixed concentrations and therefore are less easily manipulated to the specific needs of the critically ill patient than with the PN admixture. For example, with a 1,000 mL fluid restriction allotted for PN support, the increased macronutrient density could be achieved to attain eucaloric nutrition for adult patients weighing up to 60 kg (25 kcal per kg). Of course, these special dosage forms are generally more expensive than conventional products, and the cost-to-benefit ratio has not been fully demonstrated. The usual parenteral formula provided when fluid restriction is necessary is a more standard PN admixture [66], providing 70 g of amino acids and 210 g of glucose (A7D21) approximating 1,000 kcal in a 1 L final volume when
compounded from the standard 10% amino acid (700 mL) and 70% dextrose (300 mL) stock solutions, and is usually given for short periods of up to 10 days. Such a formula offers a compromise of the usual desired protein and caloric goals and may provide for a clinical outcome not distinguishable from higher protein, eucaloric regimens [67]. Tables 191.4 and 191.5 provide examples of PN formulations that may be used in the acute critical care setting in adult patients who are fluid restricted (i.e., 1,000 mL for TPN), whose regimens are often hypocaloric for clinical and practical reasons (see Table 191.4), as well as for goal amounts of nutrients in TPN in the absence of fluid restrictions [68]. A recent analysis of highly concentrated TPN admixtures, using a 16% crystalline amino acid solution containing lipid injectable emulsions in eucaloric amounts, showed them to be stable for up to 30 hours with a net fluid savings of approximately 20% compared with conventional 10% amino acids [69]. Patient-specific PN therapy for pediatric patients (premature, neonate, infant, and adolescent) may be devised using specific practice guidelines [70].

Table 191.4 HYPOCALORIC 1,000 mL Total Parenteral Nutrition Regimens as A single- Versus Mixed-Fuel System in Intensive Care Unit Patients

		Single-fuel		Mixed-fuel
Weight (kg)	Total kcal/d^a	Amino acids^b (%)	Glucose^c (%)	Amino acids (%)	Glucose (%)	Lipids^d (%)
40	600	40 g or 266 mL (4)^e	128 g or 183 mL (12.8)^e	40 g or 266 mL (4)^e	75 g or 107 mL (7.5)^e	20 g or 100 mL (2)^e
50	750	50 g or 333 mL (5)^e	160 g or 228 mL (16)^e	50 g or 333 mL (5)^e	96 g or 137 mL (9.6)^e	24 g or 120 mL (2.4)^e
60	900	60 g or 400 mL (6)^e	192 g or 275 mL (19.2)^e	60 g or 400 mL (6)^e	115 g or 164 mL (11.5)^e	29 g or 145 mL (2.9)^e
70	1,050	70 g or 466 mL (7)^e	224 g or 320 mL (22.4)^e	70 g or 466 mL (7)^e	135 g or 192 mL (13.5)^e	34 g or 170 mL (3.4)^e
80	1,200	80 g or 533 mL (8)^e	256 g or 366 mL (25.6)^e	80 g or 533 mL (8)^e	154 g or 220 mL (15.4)^e	39 g or 195 mL (3.9)^e
^aCalories from the hypocaloric regimen consists of 1 g/kg/day of protein and 15 kcal/kg/day total or approximately 50% to 60% of needs. Hypocaloric regimens that are intended as permissive underfeeding are often intended for patients whose present weight is within 10% of ideal body weight. ^bAssumes a stock bottle of 15% amino acids at 4.1 kcal/g. ^cAssumes a stock bottle of 70% hydrated dextrose at 3.4 kcal/g. ^dAssumes a stock bottle of 20% lipid emulsion at 9 kcal/g and providing approximately 20% of total calories. ^eFinal concentration of nutrient in 1,000 mL of total parenteral nutrition fluid. From Driscoll DF: Formulation of enteral and parenteral mixtures, in Pichard C, Kudsk KA (eds): Update in Intensive Care Medicine. Brussels, Springer-Verlag, 2000, pp 138–150, with permission.