representing 500 g of glucose per day for a 70 kg individual), the majority of hospitalized patients will have blood glucose levels greater than 200 mg/dL (Rosmarin et al., 1996). Because most patients suffering TBI, particularly military service members, are well nourished at the outset, serious protein-calorie malnutrition does not occur in the first seven days. However, because early plus adequate feeding in the critically ill can still improve outcome, the concept of permissive underfeeding (see above) has been developed (Burke et al., 2010; McCowen et al., 2000). A retrospective analysis of energy intakes and morbidity and mortality in the critically ill suggests that the middle tertile of intakes, 9–18 kcal/kg/day, provides the optimal outcome, with greater or lesser intakes associated with poorer outcomes (Krishnan et al., 2003). This strongly suggests that the mechanism for outcome improvement resulting from feeding early in the first week may be related to something other than optimal retention of lean body mass, such as providing sufficient glucose energy to meet the needs of key tissues including the brain, kidney, heart, and immune system, while maintaining protein synthetic rates for new protein synthesis, including for the immune system and tissue repair. Furthermore, permissive underfeeding may be beneficial by reducing the intensity of the systemic inflammatory response from a given level of injury (Burke et al., 2010). There is some support for this possibility in TBI, where patients receiving enhanced enteral nutrition showed lower levels of C-reactive protein than those receiving standard enteral nutrition during the first week after injury (Taylor et al., 1999). The greater likelihood of glucose homeostasis achieved with lower energy intakes may additionally reduce the known adverse impacts of hyperglycemia on morbidity and mortality outcomes in the critically ill (Fahy et al., 2009; McCowen et al., 2001; Pasquel et al., 2010). Most critically ill patients receive their invasive nutritional support by enteral nutrition rather than through TPN for a variety of reasons, including concerns about the relative safety and ease of administration of the two modes of feeding. However, enteral intakes are often limited by intestinal tolerance and temporary discontinuation for other procedures. Thus, although intakes of 25–30 kcal/kg/day are generally recommended in the critically ill (Cerra et al., 1997; McClave et al., 2009) to meet total energy expenditure and maintain lean tissue, intakes are usually substantially less than this, and generally less than 50 percent of goal (Krishnan et al., 2003), including in TBI (Hartl et al., 2008) during its initial phase. A small randomized trial of enhanced enteral feeding versus standard enteral feeding in TBI showed a significant reduction in infections and complications, with a suggestion of improved neurologic outcome (Taylor et al., 1999). Providing energy intake of at least 50 percent of energy needs up to 25–30 kcal/kg/day is likely appropriate for both the brain injury and for any associated critical injury. A corollary of this concept of permissive underfeeding is the interrelationship of energy and protein intake. When energy intakes are limited, supplying greater amounts of protein, up to 1.5 g/kg/day, will improve the preservation of lean body mass and improve protein synthetic rates (Hoffer et al., 1984; Ishibashi et al., 1998). Thus, it is beneficial to increase protein intakes when energy intake is less than energy expenditure. Because those enteral formulas generally employed have fixed compositions, either protein supplementation of present formulas or the development of new formula compositions incorporating these principles will need to be developed to test this hypothesis. Alternatively, enhanced enteral feeding techniques (Taylor et al., 1999) may need to be widely adopted. Beyond the initial two-week period, when the intensity of the systemic inflammatory response to TBI has generally remitted to some degree, attempting to provide sufficient energy to meet total energy expenditure in order to optimize protein metabolism is a reasonable goal. Estimating total energy expenditure by various formulas—including adding stress factors to the Harris-Benedict estimates—is inadequately sensitive, perhaps in part because nonseptic patients in medical ICU have less pronounced hypercatabolism than burn or trauma patients (Dickerson,

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement