http://surgerybook.net/

common in surgical ICU patients. Nearly 85% of those who remain in the SICU after a week and over

40% overall will be transfused, many of whom are not actively bleeding.83 It was reflexive practice a

decade or so ago to prescribe a transfusion for ICU patients whose hemoglobin levels fell below 10 g/dL

with at least two units being administered. However, packed cell transfusion may not dramatically

improve oxygen delivery, particularly in situations where the oxygen dissociation curve has been shifted

to the left, as with decreased 2,3-diphosphoglycerate concentrations (Table 10-8). In addition, in many

patient populations, transfusion of packed cells has been associated with increased infections and organ

failure. The Transfusion requirements in Critical Care (TRICC) trial published in 199984 provided data

to question the universal practice of red cell transfusion. This multi-institutional trial of patients in

Canadian ICUs was designed as a noninferiority trial comparing a restrictive transfusion strategy, aimed

at maintaining a hemoglobin level greater than 7 to 9 g/dL, to a liberal one, targeting a level of 10 to

12. Hebert and colleagues demonstrated no difference in mortality between the groups, the primary

outcome measured. Although patients with significant cardiac history were excluded a priori, nearly 300

were studied, and there was no outcome difference in this group either.85

Nonetheless, despite widespread dissemination of this landmark article, transfusion practice did not

widely change in many surgical ICU patients since and there is still a sense that older, cardiac, actively

bleeding, neurosurgical, and septic patients should be excluded from restrictive transfusion practices.

However, subsequent studies have further questioned this dogma. A large propensity matched study

analyzed the effect of transfusion on ICU patients who were not actively bleeding. It revealed that

nontransfused patients (including those with a significant cardiac history) had a lower hospital mortality

and incidence of infection and acute kidney injury.86 The RELIEVE trial was a multicenter randomized

controlled trial of ventilated ICU patients over the age of 55, including one-third with ischemic cardiac

disease. Those in the restrictive group received almost a quarter fewer transfusions and had over a 30%

lower mortality rate.87 A meta-analysis of the effect of blood transfusion during myocardial infarction

noted that the therapy was associated with a relative risk of 2.91 for mortality and 2.04 for a

subsequent MI with a number to treat of 8.88 Further, a British propensity-matched study of cardiac

surgery patients revealed odds ratios of 3.8 and 3.5, respectively, for infection and ischemia after

transfusion and ratios of 6.69 and 1.32, respectively, for 30-day and 1-year mortality in transfused

patients.89 This risk of transfusion in cardiac surgery patients is apparent even with administration of

small-volume (i.e., one or two units of packed cells) transfusions.90 A large, randomized controlled trial

of cardiac surgery patients over the age of 18 undergoing a bypass or valve replacement revealed no

change in 30-day all-cause mortality or morbidity utilizing a transfusion trigger of 8 versus 10 g/dL of

hemoglobin.91 Further, a large single center study randomized those with severe acute upper

gastrointestinal bleeding – almost 75% variceal – to a transfusion target of 7 versus 9 g/dL of

hemoglobin. Nearly half of those in the restrictive group did not receive a transfusion and the mean

time to endoscopy was 6 hours. The restrictive group had a 45% relative reduction in mortality and all

adverse events, mostly confined to those in Child’s groups A and B (there was no difference in those in

Child’s C classification group). This study reinforced the findings of a previous observation study that a

liberal transfusion target in active upper gastrointestinal bleeding doubled the risk of rebleeding and

increased mortality by 28%,92 presumably by promoting “popping the clot.” In a randomized clinical

trial of patients with traumatic brain injury, a hemoglobin target of 10 compared to 7 g/dL was

associated with no improvement in neurologic outcome at 6 months but a higher incidence of adverse

events including thromboembolism.93 Finally, in a multicenter, RCT of septic patients, a hemoglobin

target of 7 g/dL with transfusion of leukoreduced cells was not associated with difference in 90-day

hospital mortality of adverse events compared to 1 of 9 g/dL.94

Table 10-8 Factors Altering the Oxygen Dissociation Curve

Although administration of both erythropoietin95 and intravenous iron96 would seem sound practice

316

http://surgerybook.net/

to mitigate against transfusion and the harmful effects thereof, in clinical practice in the SICU, neither

has been effective in dramatically impacting transfusions or outcomes. It may be most prudent to “ride

out” the anemia seen in ICU patients, minimize unnecessary blood draws, and improve oxygen delivery

by other means (although they, too, have not afforded improved outcomes).

METABOLISM AND NUTRITION

Classic texts on surgical nutrition emphasize the metabolic and hormonal differences between the states

of starvation, one of decreased metabolic rate, and that of stress, one of increased metabolism driven by

so-called stress hormones. In reality, many surgical and trauma patients experience stress after a fast,

although the use of early nutrition in almost all has largely prevented this.

In starvation, fat oxidation is the principle energy source after glycogen stores are exhausted within

24 hours. Patients turn to gluconeogenesis (particularly in cells that are obligate glucose users such as

neurons and red blood cells) for fuel, resulting in increased free fatty acids, ketones, and glycerol. This

amplified gluconeogenesis results in exhaustion of amino acid levels and protein wasting. Eventually,

glucose levels drop as does insulin. Stress hormones, including cortisol and catecholamines, promote

protein catabolism. As glucose is used up, a switch is made to fat metabolism with the inability to keep

up via gluconeogenesis.

Critically ill patients are in a catabolic stress state with a systemic inflammatory response.

Malnutrition in critically ill patients leads to worse outcomes; inability to wean from the ventilator,

pressure ulcers, and increased risk of infectious complications. For these reasons, it is critical that ICU

patients undergo assessment and adequate support of their nutritional status. Goals of nutrition support

are to preserve lean body mass, maintain immune function, and avert metabolic complications.1,97

Metabolic Requirements and Assessment

Metabolic requirements for ICU patients for both calories and protein needs are based on ideal body

weight. Equations, such as the Harris–Benedict calculation, can assist in determination of caloric

administration and are based on the patient’s baseline energy expenditure, referred to as the basal

metabolic rate (BMR) and/or basal energy expenditure (BEE). BMR is the estimated rate of energy used

by the body at rest and is only sufficient for the functioning of vital organs. The BEE varies with age,

lean body mass, and sex. In most circumstances, ICU patients should receive 25 to 30 kcal/kg/day and

1.2 to 2.5 g/kg/day of protein (highest in stressed burn patients). In fact, hypocaloric, high-protein

diets may be best for many, if not most, ICU patients. Starting critically ill patients at 8 to 10

kcal/kg/day and attempting to achieve goal nutrition of 25 to 30 kcal/kg/day within a week is

optimal.98

The most important nutritional assessment of our ICU patients is the least high tech – that is, a careful

history and physical examination documenting diet status, weight loss, and the nature of the operation

or trauma. We often supplement this with quantitative assessment of protein and caloric adequacy.

These include simple laboratory tests and more complicated evaluations. A variety of visceral proteins

with relatively short half-lives have been assayed for their relation to protein status. These include

prealbumin, transferrin, and retinol-binding protein (with half-lives of 2, 8, and 10 days, respectively)

as opposed to albumin whose half-life is 21 days. However, even prealbumin levels may respond to

factors other than nutritional status, including ongoing inflammation. The most accurate assessment is

via indirect calorimetry. This allows for the direct measurement of oxygen consumption (VO2

). It is best

determined in those on stable ventilator settings with an inspired oxygen fraction of less than 65% and

no airway leaks. However, use of indirect calorimetry to guide nutritional support has not been

demonstrated to result in improved outcomes to justify its expense.

Energy Sources

Carbohydrates, fat, and protein each contain a unique amount of calories per gram and respiratory

quotients (RQ), a ratio of CO2 consumption (VCO2

) over oxygen consumption (VO2

). Ratios that exceed

one will result in the body’s need to eliminate additional CO2

, which can be of consequence in those on

ventilator support with lung disease. The goal of ICU nutrition is to provide energy from carbohydrate

and fat sources, so that exogenous proteins can be used for anabolism and not catabolism. This principle

is the protein sparing effect of carbohydrate and fat administration. Less than 500 kcal/day are protein

sparing, although in most instances, we provide many more calories with traditional ICU nutrition.

317

http://surgerybook.net/

Carbohydrates provide 3.4 kcal/g and have the highest RQ of 1.0. Thus, excess carbohydrate

calories can result in prolonged ventilator needs. An RQ of greater than one is seen with glycogen

storage. The ICU patient should be provided with 5 to 6 g/kg/day of carbohydrates.

Fat provides 9.0 kcal/g with an RQ of 0.7. ICU patients should receive 1 g/kg/day or less of fat. Fat

deficiency used to be fairly common in ICU patients before early and adequate nutrition. Deficiency can

result in anemia, thrombocytopenia, respiratory distress, and rash. Excess fat can promote cholestasis

and fatty liver resulting in coagulopathy. In fact, it is possible to administer nearly all of one’s

nonprotein calories as carbohydrates, providing essential fatty acids are administered. However, this

would result in an increased RQ (with ventilatory consequences) and issues of glycemic control. Fatty

acids exist in two varieties, omega-three and omega-six. Omega-three fatty acids are derived from

linolenic acid, found in fatty fish such as salmon, and are less pathogenic than omega-six fatty acids

derived from linoleic acid. In stress, after the body exhausts its supply of carbohydrates, it relies on

lipolysis which lowers the RQ (until the supply of fatty acids is also consumed and proteolysis occurs).

Protein delivers 4.0 kcal/g with an RQ of 0.8. Essential amino acids cannot be synthesized

endogenously and must be provided in the diet and include: methionine, threonine, tryptophan, valine,

phenylalanine, isoleucine, leucine, lysine, and histidine. Diets specialized for those with renal failure

deliver their protein in the form of essential amino acids. Semiessential amino acids such as arginine and

glutamine are difficult for the body to manufacture in times of stress and are obligately utilized by

certain cell types. Protein is the important component of nutrition due to the body’s obligate synthetic

needs. We are able to measure the efficacy of protein nutrition by assaying nitrogen balance – a balance

sheet between protein in and nitrogen out measured in the urine. This, of course, requires that we

remember that a gram of protein contains 6.25 g of nitrogen. Severely burned patients can lose 25 g N

(125 g pro) daily. In certain conditions, such as sepsis, we may not be able to provide enough protein to

avoid negative nitrogen balance.

Vitamins and minerals. Contemporary supplemental nutrition contains sufficient trace elements and

vitamins to make deficiency in ICU patients a thing of the past in most instances. A resurgence in this

has occurred with the care of bariatric patients and the deficiencies they experience post bypass, most

notably of thiamine. Interesting to note, virtually all enteral formulae contain vitamin K – some in

rather large amounts – so anticoagulation may be difficult (Table 10-9).

The nutritional support of the malnourished and nourished patient employs different strategies to

optimize outcome. Malnourished patients may benefit from preoperative nutritional support for at least

7 days that can be delivered enterally, if possible. Well-nourished patients should have nutrition

resumed within 48 hours; however, it is safe, if not wise to delay parenteral nutrition for 5 to 7 days if

the enteral route is not possible. There is much controversy and regional variability regarding this issue

with the Europeans preferring earlier initiation of parenteral nutrition, as will be discussed below.

Enteral Nutrition

9 In general, enteral nutrition should be used early and often in ICU patients. Standard enteral formulas

include fatty acids, carbohydrates, protein, vitamins, minerals, and micronutrients. A standard formula

is isotonic to serum and has a calorie density of 1 kcal/mL. A concentrated formula is used in patients

who may benefit from low volume nutrition. Concentrated formula is hyperosmolar with a calorie

density of 1.2 to 2.0 kcal/mL. There are specific formulations that address types of organ dysfunction.

So-called pulmonary formulations are low in carbohydrates; renal formulas are concentrated with

essential amino acids and hepatic formulas contain branched chain amino acids. Further, enteral feeds

are either polymeric or elemental – that is, broken down to the basic building blocks of the energy

substrates – proteins as free amino acids, fats as medium chain triglycerides, and carbohydrates as

oligosaccharides at a calorie concentrations of 1 to 1.5 kcal/mL. Elemental feeds may be useful in the

stressed patient as they require minimal intraluminal digestion. Elemental formulas should be

considered in patients with nutrition absorption disorders and those that fail to tolerate standard enteral

nutrition.

Table 10-9 Vitamin Deficiencies

318

http://surgerybook.net/

There are several complications related to enteral nutrition in the ICU, the most significant of which

is aspiration. To decrease the aspiration risk, the ICU staff should maintain head-of-bed elevation at 45

degrees or greater and consider promotility agents. Regular assay of gastric residuals may not decrease

aspiration risk.99

Parenteral nutrition requires the insertion of a central catheter or a centrally directed peripheral

one. As such, the complications of parenteral nutrition include the risks of catheter insertion and

infection in addition to the metabolic consequences of feeding (such as hyperglycemia) that can also be

seen with enteral nutrition. Parenteral nutrition can be prepared in standard or concentrated

formulations depending upon the glucose with added amino acids and fat. There may be evidence that

hypocaloric formulas with limitation of fat early on are beneficial.97

Immunonutrition

In addition to use of omega-three fatty acids, there are other ingredients that may be helpful to use in

the stressed patient. In small, early nonrandomized trials, semiessential amino acids arginine and

glutamine have been shown to be beneficial and the use of high-protein/low-calorie strategies may be

most optimal, particularly using branched chain amino acids. There was thought to be a notable

exception to this strategy – arginine, in particular, was not advised in the patient who is already

infected.100 Arginine is metabolized to nitric oxide which may increase permeability, hemodynamic

instability, and translocation in septic patients. However, this may be mitigated by increased production

of arginase in septic patients. Increasing literature actually supports increased perfusion by giving

arginine in sepsis.101,102 Thus, arginine administration is likely safe in sepsis, but the preferred dose is

still of debate. Glutamine is rapidly depleted from muscle stores and thought to be important in

minimizing translocation as it assists in maintaining gut barrier function as an enterocyte fuel. A

hypocaloric, high-protein strategy appears to be most beneficial with a calorie to nitrogen ratio of 100–

150:1 compare to the more traditional 300:1. However, large randomized trials discussed below

question the clinical efficacy of immunonutrition.

Controversial Issues in ICU Nutrition

Delivery of Early Enteral Nutrition

Most critically ill patients are unable to ingest enteral nutrition and have a need for access to the enteric

track. There is no compelling evidence that enteral feeds need be delivered in a specific location – that

is, there is little data to suggest that postpyloric feeding results in lower aspiration rates or in better

food tolerance with higher caloric provision.103 Gastric tube feeding is the most common method to

deliver enteric nutrition. Via an orogastric or nasogastric route, sump tubes or smaller-caliber feeding

tubes are placed into the stomach with confirmation by abdominal x-ray. Surgically placed gastric

feeding tubes, whether placed percutaneous or via an open procedure, can also be used for a more longterm enteral nutrition access. Postpyloric tube feeding should be considered in patients who cannot

tolerate gastric enteral nutrition or patients who have undergone surgical procedures that have altered

anatomy. Placement of a postpyloric tube can be done with blind placement via the nasal or oral route.

These tubes may also be placed via open surgical procedures with a single tube inserted directly in the

intestine distal to the pylorus or with a tube with a port to drain the stomach and a distal port to deliver

enteral nutrition into the duodenum. Postpyloric placement of enteral nutrition access has no benefit

over gastric feeding access in terms of energy delivered or rate of aspiration pneumonia, and so it is

recommended to proceed with enteral nutrition via the gastric route.103

319