John K. Stene
Thomas C. Vary
Immediate Concerns
The care of critically ill patients usually focuses on the immediate concerns of cardiorespiratory resuscitation, ventilatory support, maintenance of adequate hemodynamic parameters, and antibiotics to control infectious processes. Despite the increase in metabolic substrate utilization, as critically ill patients become catabolic, nutritional support is frequently overlooked. The extreme catabolism and negative nitrogen balance from critical illness has been well recognized for decades, as described in a 1976 review by Cuthbertson (1), but there still exists a conventional wisdom that starvation is not harmful to critically ill patients and no attempts to reverse it with nutrition support should be undertaken (1). However, there is no medical evidence that starvation is therapeutic. In fact, Kinney (2) pointed out more than 25 years ago that the catabolic response to multiple injuries resulted in loss of up to 20% of normal body weight within 3 weeks despite some oral intake started during the first week. Kinney also noted the catabolic melting (rapid wasting) of large weight-bearing muscle groups of critically injured patients in a hospital setting (2).
Critical illness leads to a change in the hormonal milieu in response to cytokines released from injured tissues; this change increases the demands for energy substrates. In addition, cytokines promote break down of muscle proteins to provide amino acids that are deaminated and carbon chains utilized for energy (3). The tidal metaphor of Cuthbertson has long been used to describe the metabolic response to acute injury. In this metaphor, the initial response to shock is a marked reduction in metabolic rate, termed the ebb phase, conjuring up an image of the patient's life ebbing away like the receding tide. The flow phase suggests the image of the returning tide, which involves the patient mounting a compensatory response to shock and tissue injury.
|
Table 64.1 Ebb Flow Phase of Response to Injury |
||||||||||||||||||||||||||||||||
|
Further refinement of this concept has described an early catabolic period of the flow phase when muscle protein is mobilized to provide energy and substrate for tissue repair (4). This is followed in the late flow phase, a period of anabolic activity and a tissue healing component of the flow phase. Both subintervals of the flow phase increase the patient's need for nutritional supplementation to optimize healing and reduce the length of the critical illness. The flow phase is associated with an increased metabolic rate as measured by oxygen consumption/heat production, which gradually returns to normal (Table 64.1). Nutritional support prescriptions need to change throughout the patient's period of critical illness to provide adequate protein-energy substrates during the flow phase and to prevent overfeeding when the patient's convalescence is complete. Although critical illness most frequently compromises a patient's ability to normally intake adequate nutrition, critical care technology offers several alternative routes of providing nutritional support. These routes of nutrition support include intravenous routes—total parenteral nutrition (TPN) and peripheral parenteral nutrition—and enteral routes with various formulae tailored to the patient's unique requirements (Table 64.2).
|
Table 64.2 Commonly Used Routes of Nutrition |
|||||||||||||||||||||
|
|||||||||||||||||||||
Nutritional Support Service
A nutritional support service aids those caring for critically ill patients in providing appropriate nutritional prescriptions. The nutritional support service is also responsible for evaluating the patients' responses to nutrition prescriptions. These monitored evaluations range from the patients' actual calorie expenditure to protein requirements so as to prevent catabolism with profoundly negative nitrogen balance. Finally, the nutritional support service should be a focus for research concerning the improvement of nutritional care of critically ill patients.
Classically, the nutritional support service consisted of a registered dietitian (RD) with advanced training in the biochemical aspects of critical illness, a physician consultant, and assistants who provide nutritional consults to metabolically unstable patients (Table 64.3). One of the team's duties is screening patients for risks of malnutrition and diagnosing various nutritional deficiency states. The team consults on patients either by specific request or on a systematic protocol-driven basis—such as for patients in the critical care unit—according to local hospital policy.
The nutritional consult includes estimates of the patient's metabolic rate and calorie expenditure, estimated either with the Harris-Benedict equation or actually measured via indirect calorimetry. The nutritional support team may collaborate with respiratory care professionals to perform indirect calorimetry measurements.
Protein requirements are estimated from the patient's critical illness diagnosis and monitored via visceral protein analysis or 24-hour nitrogen-balance measurements. The nutritional support service incorporates such measurements into their daily nutritional prescriptions to provide the patients with optimum nutritional support.
|
Table 64.3 Organization of a Nutrition Support Service |
|
|
Alternatives to Nutrition Support Services
The author (JKS) has participated as a physician member of a nutrition support service that rounded on all patients receiving artificial nutrition support, both parenteral and enteral, and participated in the redesigning of the service when physician shortages precluded using a dedicated physician. Currently, the author's institution uses a RD with an advanced degree and extensive critical care experience to oversee a team of RDs who see patients by request and estimate nutritional needs. The team works with the physicians in the various critical care units to provide recommended nutritional prescriptions.
The nutritional support service also provides an educational service to critical care physicians in training, advises attending critical care physicians on nutritional issues, and spearheads nutritional research projects in the various intensive care units (ICUs). This service appears to work as well as the service organized with an obligatory physician consultant and is more cost effective in terms of physician time.
|
Table 64.4 Indices of Malnutrition: Protein-Energy Malnutrition |
||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||
Nutritional Assessment
Critically ill patients are at risk for nutritional deficiencies. The diagnostic criteria for various protein and calorie-deficient states are noted in Table 64.4. It is wise for critical care practitioners to remember that kwashiorkor can be seen in the critically ill patient following prolonged fasting, not just in famine-stricken developing countries.
In general, protein-energy malnutrition, commonly known as protein-caloric malnutrition (PCM), has been classified as marasmus, kwashiorkor, and marasmic kwashiorkor (5). The differentiation has been the presence of edema and severe serum protein depletion in kwashiorkor as compared with body wasting without edema and preservation of plasma proteins in marasmus. Marasmic kwashiorkor exhibits features of both severe wasting (marasmus) and protein depletion with edema (kwashiorkor). Typically, marasmic kwashiorkor occurs when a chronically malnourished patient is subjected to added catabolic stress such as trauma or sepsis.
Because of the risk that a critically ill patient with underlying nutritional deficiency will develop marasmic kwashiorkor, they are screened for chronic undernutrition on admission to the hospital/intensive care unit. One such screening tool, the Malnutrition Universal Screening Tool (MUST), provides a numerical index: 0, minimal risk; 1, medium risk; or 2, high risk of malnutrition (6). The urgency of instituting nutritional support increases for patients with a score of 1 and especially 2. The MUST evaluates patients' body mass index (weight in kilograms [kg] divided by height in meters squared [m2]), prehospital involuntary weight loss as a percent of body weight, and the potential for critical illness-induced starvation for the next 5 days.
Anthropometric indices such as MUST are useful for identifying patients who need aggressive nutritional support. However, a more precise diagnosis of a critically ill patient's current nutritional state as well as a tool to monitor the patient's response to nutritional support requires a combination of anthropometric, serum chemical analysis, and physiologic measurements (7). Charting the patient's daily weight is one of the simplest anthropometric measurements that can identify both long-term loss of lean body mass and the development of edema. The body mass index, mid–upper arm circumference, and estimate of body fat from triceps skin fold are other measurements to track the critically ill patient's nutritional state (7).
Many biochemical markers are available to track a patient's nutritional state. These include albumin concentration, estimates of whole body potassium, water balance, and visceral proteins such as prealbumin, transferrin, and retinol-binding protein. Albumin concentration is a good prognostic marker of a patient's chronic nutritional state. Perioperative mortality has been shown to increase when albumin concentration is less than 30 gm/L (8). However, serum albumin has a long serum half-life and, hence, does not reflect acute responses to nutritional therapy.
Total body potassium represents lean cellular mass and can be a useful measurement of changes in lean tissue mass that occur rather slowly. In a similar fashion, 24-hour creatinine production is a marker of total skeletal muscle mass. However, both of these markers change relatively slowly and may not be useful to monitor the patient's response to nutritional therapy for critical illness.
Visceral proteins—prealbumin, transferrin, and retinol-binding protein—with their short half-lives, can be useful to monitor protein synthetic response to nutritional support therapy (7). Measuring nitrogen (N) balance via a 24-hour urine urea nitrogen excretion is a very useful technique to monitor response to nutritional support. Nitrogen balance, defined as protein N intake minus protein N excretion, is approximately zero (balanced) in the healthy, free-living state. The catabolism triggered by critical illness will lead to increased loss of protein N and a negative nitrogen balance. The goal of the intensivist is to provide nutritional support to lead to a positive N balance until the patient's depleted protein state is replenished.
The most direct method to monitor N balance is to measure 24-hour protein intake, divide the number by 6.25, which is the average ratio of molecular weight of amino acid to nitrogen content, then subtract protein loss, represented by the excretion of urine urea N (UUN) over the same 24-hour period. Usually, an empiric constant of 4 gm N per day is added to the UUN measurement to account for nonurinary losses of proteins (Table 64.5). If the patient's blood urea nitrogen (BUN) is not stable, corrections for retained N must be performed (Table 64.5).
Energy expenditure may be calculated by the Harris-Benedict equation or measured directly with an indirect calorimeter. The Harris-Benedict equation is fairly accurate if the patient's weight is near his or her ideal body weight. Because lipid has a metabolic rate different from lean body, the Harris-Benedict equation may not accurately estimate the patient's actual caloric expenditure, especially in the obese.
Indirect calorimetry takes advantage of the fact that oxygen consumption (VO2) and carbon dioxide production (VCO2) are stoichiometric products of aerobic metabolism. Thus, the patient's pulmonary gas exchange is measured over a several-hour period and converted to the patient's actual energy expenditure by empiric equation (9) (Table 64.6). The longer the time period over which the measurements occur, the closer to the patient's actual average daily metabolic rate. Because patients have a diurnal variation, as well as an activity variation, in metabolic rate, short time periods of indirect calorimetry may overestimate or underestimate the average 24-hour energy expenditure. Indirect calorimetry also provides information that is useful for managing a patient's respiratory status, because a high oxygen consumption and carbon dioxide production is associated with obligatory mechanical ventilatory support. Furthermore, mean expired carbon dioxide, a value required for accurate measurement of pulmonary dead space, is directly measured by indirect calorimetry.
|
Table 64.5 Assessment of Daily Nitrogen Balance |
||
|
||
|
Table 64.6 Indirect Calorimetry—Daily Energy Expenditure |
||
|
Overall, monitoring of a patient's response to nutrition support is performed by a battery of the tests described above. Daily weights are tracked to monitor a patient's response to nutritional support. Initially, weight gain will represent an increase in total body water, which will start decreasing (diuresing) as the patient becomes anabolic. During the anabolic phase, the patient will initially lose weight via diuresis of water, then gain weight as lean body mass. Obese patients may continue to lose weight from fat stores if they are fed a high-protein, restricted calorie diet. Repeated indirect calorimetry can follow the respiratory quotient (VCO2/VO2) that provides a measure of whether a patient is being underfed or overfed. Utilization of protein intake can be tracked by repeated 24-hour N balances via urine urea nitrogen measurements and nutrition intake.
Beyond direct physiologic measurements of adequacy of nutritional intake, functional measurements that indirectly depend on the nutrition response, such as immunologic function, can be followed. The severely malnourished patient exhibits immunocompromise by skin test anergy, low absolute lymphocyte count, low T-cell lymphocyte count, difficulty mounting a fever, and increased susceptibility to infections; these functions should improve with aggressive nutritional support.
Timing of Nutritional Support
Nutritional evaluation of the critically ill frequently reveals patients admitted to the ICU who exhibit signs of marasmic kwashiorkor and who have an urgent need for nutritional support. Once the decision is made to provide nutritional support for the critically ill, one has to decide on the best route of administration—parenteral or enteral. Data from one to two decades ago suggested that parenteral nutrition (total parenteral nutrition or TPN) was associated with a higher mortality rate than enteral nutrition (total enteral nutrition or TEN) in patients who had abdominal trauma (10,11). However, more recent studies in other patient populations have failed to confirm these results (12,13). In several studies comparing the administration routes of nutritional support, the major difference between TEN and TPN was the fact that enteral-fed patients had more frequent interruptions of nutrition (14,15). Therefore, TEN patients received somewhat less nutritional substrate than TPN patients.
Historically, there has been grave concern that TPN puts a patient at risk for serious infection complications. However, recent studies reveal that bloodstream infections are relatively rare during TPN, and TEN patients have a significantly higher incidence of feeding tube complications than TPN patients have from central venous catheter complications.
In one review of infectious complications in the ICU, it was noted that inadequate nutritional intake—less than 7 kcal/kg per day—was associated with sepsis. In this group of underfed patients, those who received TEN were equally likely to have infectious complications as those who received TPN (16). This study supports the practice of starting early and adequate nutrition support to prevent infectious morbidity.
One of the alleged benefits of early enteral feeding is to maintain the integrity of the gut mucosal lining and thus limit translocation of microorganisms from intestinal lumen to bloodstream. This hypothesis was tested using a macromolecular marker of gut permeability in two groups of patients: those with early enteral feedings and those kept NPO (nothing by mouth) (17). There was no demonstrated difference in permeability between the two groups, which suggests that bacterial translocation from the gut to the bloodstream is not the source of bloodstream infections in patients who are fed parenterally.
TEN does have some demonstrated advantages over TPN, including lower cost, using gut absorption to regulate total body water balance, and the ability to utilize larger molecules as a food source. TPN has advantages over TEN that include fewer interruptions for NPO status prior to procedures and no interference with drug absorption.
We recommend that patients receive early nutritional support to help reduce infectious complications. Because patients need adequate protein and calorie intake to prevent complications from infections, those patients who cannot tolerate TEN should be immediately switched to TPN. It is clearly an error to withhold TPN for the notion that the risks of TPN outweigh the risks of starvation for several days.
The marked improvement in managing central venous access that developed in recent years has markedly reduced the risk of infectious complications of TPN. Furthermore, accurate and widespread assessment of actual metabolic needs with indirect calorimetry and N balance has led to a significant reduction in TPN calorie load that has decreased metabolic complications of TPN (18,19).
Nutrition Support for Specific Organ Dysfunction
Nutritional support is often modified for specific organ dysfunction because certain pathophysiologic conditions lead to changes in the metabolic handling of nutrient substrates. These disease states include liver failure, kidney failure, diabetes/glycemic control, and brain injury.
Hepatic Failure
Liver failure is one of the most vexing morbidities requiring prescriptions for nutritional support. In severe hepatic failure, nitrogen from amino acid metabolism remains as ammonia because it is not metabolized to urea. Hyperammonemia leads to secondary neurologic dysfunction including hepatic coma. Thus, protein needs to be administered in amounts to replace catabolic losses, but not high enough to lead to high levels of ammonia. Sorbitol may be administered enterally to try to lower ammonia by increasing gut motility and decreasing intestinal transit time to decrease the absorption of proteins and ammonia from bacterial metabolism of intraluminal amino acids. However, sorbitol has no effect on hepatic amino acid metabolism.
Hepatic glycogen stores are also depleted in end-stage hepatic failure, making the maintenance of normoglycemia extremely difficult. High glucose feedings may be required to compensate for lack of hepatic glucose production. The bottom line in end-stage hepatic failure is that nutritional support needs to be modified to minimize plasma ammonia levels and maintain normal glucose levels; nutritional support will not, of course, reverse significant hepatic injury.
Renal Failure
Rising blood levels of the nitrogenous waste product urea and creatinine are the hallmarks of renal failure. Acute renal failure is, unfortunately, rather common in critically ill patients in response to intrarenal insults from circulating inflammatory mediators as well as changes in renal perfusion. Although past nutritional efforts for acute renal failure have focused on limiting protein intake to reduce urea production or using only essential amino acids to recycle amines to reduce urea production, modern critical care uses renal replacement therapy (20,21). Either hemodialysis or continuous venovenous hemoperfusion is used to maintain acceptable levels of urea, creatinine, and electrolytes until the kidney regains its function. One of the few evidence-based therapies that enhances the repair of acute renal failure is nutritional support with adequate protein for renal healing. Thus, efforts to restrict protein intake in patients with acute renal failure are misguided and should be discouraged. Renal replacement therapy should be instituted to maintain acceptable levels of urea while the healing kidney is supported by adequate protein replacement. Prior to the commencement of renal replacement therapy, serum electrolytes need to be carefully monitored because renal failure often leads to excess water (H2O), low sodium (Na), and high potassium (K). Nutrition support may have to be modified to correct serum electrolyte disorders.
Glycemic Control
Hyperglycemia is a common occurrence in critically ill patients. Not all hyperglycemic critically ill patients have diabetes, but the hormonal milieu of the stress response tends to cause hyperglycemia. Among the hormones that increase plasma glucose are growth hormone, cortisol, glucagon, and epinephrine (22). Furthermore, one of the diagnostic hallmarks of sepsis is glucose intolerance/insulin resistance. Some authors have postulated that hyperglycemia has survival advantages in providing energy substrate for collagen synthesis for wound healing. A large randomized trial compared glucose levels of 180 to 200 mg/dL in the control group, consistent with the survival advantage hypothesis, to normoglycemia of 80 to 110 mg/dL in the experimental group (23). The normoglycemic group had a significant improvement in outcome effects that persisted for several months post recovery. These effects included decreased ICU length of stay, reduced infectious complications, and survival.
Despite fears that tight glucose control might expose patients to risks of dangerous hypoglycemic episodes, it is recommended that critically ill patients be managed with normal glucose levels (23,24). Furthermore, in the controlled environment of the critical care unit, where it is feasible to monitor glucose levels every hour, an insulin infusion is the best method to manage glucose concentrations. Infusions of regular insulin can be rapidly changed to respond to hourly, or more frequent, bedside glucometer measurements that may vary widely in unstable critically ill patients. To protect patients from hypoglycemia from the insulin infusion, patients receive either nutritional support (TPN or TEN at goals) or, if needed, a 10% dextrose infusion (23,24,25).
Brain Injury
The injured brain is characterized by an impairment of the blood–brain barrier, which causes increased susceptibility to changes in plasma concentrations of glucose and electrolytes. The prevention of edema in the injured brain, an essential requirement to allow the brain to heal and reorganize neural pathways, places demands on the composition of nutritional support. Nutritional goals for the head-injured patient include avoidance of hyperglycemia, hyperosmolar state, and maintenance of normal sodium and potassium. Hyperosmolar coma, a known complication of TPN, is usually caused by extreme hyperglycemia and can be treated with insulin. This is certainly another good reason to accurately assess a patient's metabolic rate and calorie requirement from nutrition support. Hyperglycemia, even with normal osmolality, is detrimental to the injured brain by causing cerebral edema in ischemic areas of the brain. This occurs when one glucose molecule diffuses into the brain and is anaerobically metabolized to two lactic acid molecules that are highly polar and do not diffuse out into the bloodstream. Thus, the interior of the cell increases in osmolality (more nondiffusable particles) relative to plasma and attracts increased cell water. The above effect is magnified by hyperglycemia in patients who have brain lesions associated with low perfusion/O2 delivery (26,27).
Besides adjusting nutritional support to maintain euglycemia, with appropriate calorie administration and insulin use, sodium replacement is important in brain-injured patients. Many patients with brain injuries develop hyponatremia, which tends to increase the risk of cerebral edema from excess plasma water and decreased serum osmolality. Cerebral salt-wasting syndrome or inability to conserve sodium in the kidney needs to be differentiated from syndrome of inappropriate antidiuretic hormone (ADH) secretion (SIADH). Cerebral salt-wasting patients tend to be hypovolemic and need extra sodium replacement in their nutrition support, whereas patients with SIADH tend to be hypervolemic and generally need fluid restriction to correct their serum sodium levels.
Brain-injured patients will require adequate protein intake to compensate for their increased demands to provide brain healing. The metabolic rate should be measured directly and caloric needs met by appropriate amounts of nutritional support. There is some concern about excitatory amino acids having deleterious effects on the injured brain, but it is unclear that nutritional support affects the CNS concentration of these amino acids.
Summary
Nutritional support for critically ill patients is often overlooked by clinicians concerned with the immediate concerns of hemodynamic instability and respiratory failure. However, starvation is not therapeutic, and prolonged inadequate energy intake will lead to malnutrition states in critically ill patients. Increased metabolic activity stimulated by stress hormonal response and a generalized inflammatory state may accelerate the appearance of malnutrition. Adequate protein-energy intake reduces critical care unit–acquired bloodstream infections, and aggressive insulin management to maintain euglycemia will reduce mortality as well as infectious complications of critical illness. Supported in part by NIH GM 39277 (TCV).
References
1. Cuthbertson DP. Surgical metabolism: historical and evolutionary aspects. In: Wilkinson AW, Cuthbertson D, eds. Metabolism and the Response to Injury. Chicago, IL: Year Book Medical; 1976:1–34.
2. Kinney JM. The application of indirect calorimetry to clinical studies. In: Kinney JM, Buskirk ER, Munrot N, eds. Assessment of Energy Metabolism in Health and Disease. Columbus, OH: Ross Laboratories; 1980:42–48.
3. Vary TC, Siegel JH. Sepsis, abnormal metabolic control and multiple organ failure syndrome. In: Siegel JS, ed. Trauma: Emergency Surgery and Critical Care. New York, NY: Churchill Livingstone; 1987:411.
4. Kinney JM, Gump FE. The metabolic response to injury in American College of Surgeons Committee on Pre and Post Operative Care: Manual of Preoperative and Postoperative Care, ed 3. Philadelphia, PA: WB Saunders; 1983;15–37.
5. Odigwe C, Ejibe DK. Malnutrition. Student BMJ. 2005;13:404–405.
6. Malnutrition Advisory Group: Malnutrition Universal Screening Tool (‘MUST’). www.bapen.org.uk. Accessed August 30, 2007.
7. Halder M, Halder SQ. Assessment of protein-calorie malnutrition. Clin Che. 1984;30:1286–1299.
8. Mullen JL, Buzby GP, Waldman MT, et al. Prediction of operative morbidity and mortality by pre-operative nutritional assessment. Surg Forum. 1979;30:80–82.
9. Bursztein S, Elwyn DH, Askanazi J, et al. Energy metabolism, indirect calorimetry, and nutrition. Baltimor, MD: Williams & Wilkins; 1989.
10. Kudsk KA, Croce MA, Fabian TG, et al. Enteral versus parenteral feeding. Effects on septic morbidity after blunt and penetrating abdominal trauma. Ann Surg. 1992;215:503–513.
11. Heyland DK, MacDonald S, Keefe L, et al. Total parenteral nutrition in the critically ill patient: a meta-analysis. JAMA. 1998;280:2013–2019.
12. Lipman TO. Grains or veins: is enteral nutrition really better than parenteral nutrition? A look at the evidence. JPEN J Parenter Enteral Nutr. 1998;22:167–182.
13. Woodcock W, Zeigler D, Palmer M, et al. Enteral versus parenteral nutrition: a pragmatic study. Nutrition. 2001;17:1–12.
14. Jeejebhory KN. Total parenteral nutrition: potion or poison? Am J Clin Nutr. 2001;74:160–163.
15. Pacelli F, Bossola M, Papa V, et al. Enteral vs. parenteral nutrition after major abdominal surgery: an even match. Arch Surg. 2001;136:933–936.
16. Rubinson L, Diette GB, Song X, et al. Low caloric intake is associated with nosocomial blood stream infections in patients in the medical intensive care unit. Crit Care Med. 2004;32:350–357.
17. Reynolds JV, Kanwar S, Welsh FKS, et al. Does the route of feeding modify gut barrier function and clinical outcome in patients after major upper gastrointestinal surgery? JPEN J Parenter Enteral Nutr. 1997;21:196–201.
18. McCowen KC, Friel C, Sternberg J, et al. Hypocaloric total parenteral nutrition: effectiveness in prevention of hyperglycemia and infections complications—a randomized clinical trial. Crit Care Med. 2000;28:3606–3611.
19. Lowry SF, Brennan MF. Abnormal liver function during parenteral nutrition: relation to infusion excess. J Surg Res. 1979;26:300.
20. Giordano C. Use of exogenous and endogenous urea for protein synthesis in normal and uremic subjects. J Lab Clin Med. 1963;62:231.
21. Grovanetti S, Neaggiore Q. A low nitrogen diet with protein of high biologic value for severe chronic uremia. Lancet. 1964;1:1000.
22. Bone RC. Toward an epidemiology and natural history of SIRS (systemic inflammatory response syndrome). JAMA. 1992;268:3452–3455.
23. Vanden Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med. 2001;345:1359–1367.
24. Vanden Berghe G, Wilmer A, Hermans G, et al. Intensive insulin therapy in medical ICU. N Engl J Med. 2006;359:449–461.
25. Inzucchi SE. Management of hyperglycemia in the hospital setting. N Engl J Med. 2006;355:1903–1911.
26. Young B, Ott L, Haack D, et al. Effect of total parenteral nutrition upon intracranial pressure in severe head injury. J Neurosurg. 1987;67:76–80.
27. Sieber FE, Smith DS, Traystman RJ, et al. Glucose: a re-evaluation of its intraoperative use. Anesthesiology. 1987;67:72.
Haldane correction