Nutritional support is an important component of the care of the traumatically brain injured patient. Patients in a coma after traumatic brain injury (TBI) , even without other major injuries, are in a metabolic state similar to patients with major trauma and burns [1]. Caloric expenditure can be almost twice that of the expected resting energy expenditure (REE) and may be influenced by temperature, muscle tone, Glasgow Coma Scale (GCS) score, and time of measurement in relation to injury [2]. The REE in turn is used to calculate caloric needs.

Indirect calorimetry (IDC) is the most accurate form of nutritional assessment. It measures the oxygen consumption and carbon dioxide production from the patient, from which the REE can be calculated [3]. Several studies that include patients with TBI utilize IDC to determine REE in the intensive care unit (ICU) [4]. Most used single measurements of REE to determine total caloric needs but failed to identify optimal timing, duration, or frequency of REE measurement. The use of repeated measurements of REE in the ICU to monitor ongoing nutritional requirements has not been fully assessed.

Predictive equations such as the Harris–Benedict equation, with adjustments based on activity level [5], have been used to determine the REE [6]. Unfortunately, a direct and accurate relationship of these equations to a patient’s daily nutritional requirements has not been established. Currently the American Society for Parenteral and Enteral Nutrition guidelines recommend the range of 20–35 kcal/kg/day for adults, depending on the severity of stress or illness. In the critically ill obese patient, 11–14 kcal/kg actual body weight per day or 22–25 kcal/kg ideal body weight (IBW) per day is recommended [7]. In general, ICU patients should receive hypocaloric, high protein feeding. Fat calories from propofol infusions may be factored into the equation.

Protein metabolism after TBI is also similar to that after major systemic injury [5]. Protein energy metabolism is assessed in the ICU by measurement of the nitrogen balance, which is the only marker for this measurement widely reported in the neurological ICU population [8]. The nitrogen balance is the daily difference between nitrogen intake and output. The percentage of calories derived from protein has been shown to increase from the normal range of 10–15 % to upwards of 30 % after TBI [9]. The consumption of lean body mass during this catabolic state may be decreased by providing 100–150 % of expended calories and further decreased by providing a higher protein intake [10], however this incurs the risks of overfeeding such as excess carbon dioxide production, and currently a hypocaloric approach is recommended. Attempts to lessen nitrogen wasting after TBI by increasing protein intake beyond 17 g/day results in greater protein catabolism, so that only 50 % of administered nitrogen may be retained. The level of nitrogen intake that results in a nitrogen loss of 10 g/day is 15–17 g/day, or 0.3–0.5 g/kg/day. This represents about 20 % of the caloric composition of most feeding formulas designed for a hypermetabolic patient.

There is a significantly greater mortality rate as a consequence of undernutrition for a 2-week period after injury, as compared to receiving full nutrition by 7 days [11]. Fewer infectious and overall complications have been demonstrated by starting feedings that meet the estimated energy and nitrogen requirements on day one following injury [8]. Patients receiving early feeding are also more likely to have energy and nitrogen requirements met by 1 week. Early feeding is recommended once resuscitation and hemodynamic stability is achieved and therefore should begin within 72 h following injury to achieve full nutritional support by day seven [12].

There is no validated means of measuring the response to nutrition in the ICU setting. However some nutritional assessment parameters, such as the serum albumin level, are excellent prognostic indicators of morbidity, mortality, and ICU and hospital length of stay [13]. Studies have examined the predictive value of a single albumin level measured upon admission to the hospital, within the first few days of admission, or prior to surgery or other planned treatment. Yet the value of using sequential albumin levels for monitoring nutrition progress is low. Spontaneous changes in albumin concentration are expected in critically ill patients. They occur slowly and are affected by acute phase responses and compartmental fluid shifts which occur during an ICU stay [14]. In some patients albumin levels may not change significantly within the acute setting due to the long half-life of 19 days.

Prealbumin has been used frequently as a marker of nutritional response because of its short half-life of 2–3 days, so that significant changes may be detected in days to weeks. Its use has fallen out of favor, as prealbumin levels are also influenced by acute phase responses and do not correlate with outcomes [15].

Body weight is the most commonly used indicator of nutrition adequacy in nonhospitalized patients, but due to the confounding effects of fluid retention and gradual weight loss during acute illness, changes in weight more often indicate alterations in fluid balance [16]. Anthropometric indices such as the mid-arm circumference, triceps skin fold thickness, and calculated mid-arm muscle circumference are typical indicators of somatic protein and fat reserves and may be used to monitor response to nutrition therapy under normal circumstances, but are also confounded by fluid balance in the ICU setting.

The most commonly employed options for nutrition delivery currently are the enteral (gastric or jejunal) and parenteral routes. The enteral method is superior to parenteral nutrition in patients with a functional gastrointestinal tract [17]. Enteral formulations utilize more effective substrates to support cell and organ function, result in lower risks of hyperglycemia or hyperosmolarity, are administered at rates which may avoid overfeeding, and maintain the mass and barrier function of the gut. However, enteral feeding in patients with gastrointestinal intolerance is associated with underfeeding and subsequent malnutrition. Parenteral formulations deliver more dependable nutrient bioavailability, result in greater nutrition effects in a shorter time period, act independently of gastrointestinal function, and avoid gastrointestinal feeding complications such as intolerance, abdominal distention, and diarrhea. However overfeeding, administration of excess dextrose, triglycerides, or calories, and refeeding syndrome (from rapid feeding with preexisting malnutrition) may occur. This may cause certain metabolic complications such as hyperglycemia, hypertriglyceridemia, hypervolemia, and hypercapnia [18]. It has been estimated that at least 20 % of patients with TBI exhibit gastric feeding intolerance within the first week. In this situation, parenteral nutrition may be utilized for initial nutrition support, though improved outcomes from early parenteral nutrition have not been demonstrated. Jejunal feeding can also be used and provides the benefits of enteral feedings while avoiding the complications of the parenteral route. Although older evidence had suggested that pneumonia rates may be reduced by jejunal feeding due to the reduced aspiration presumed to occur during gastric feeding [19], more recent investigations have found no influence of the delivery route on this complication [20]. Currently, routine nasojejunal feeding is not routinely recommended.

Nutrition support and therapy are important factors in the management of the critically ill patient. Nutritional issues pertaining specifically to the brain injured trauma population have been underestimated and certainly less studied. Additional work is needed to realize outcome benefits that could result from improved nutrition delivery in the TBI population.

On any given day in intensive care units (ICU) across the world, about half of the patients carry a diagnosis of infection, and 71 % are receiving antibiotics [21]). This extensive antibiotic use, both appropriate and indiscriminate, has resulted in increased bacterial antibiotic resistance and emergence of multi-drug-resistant (MDR) pathogens that are increasingly difficult to treat [22]. Infections with antibiotic-resistant organisms result in the death of about 23,000 people a year in the United States according to the Centers for Disease Control and Prevention [22].

Indiscriminate use of antibiotics has other serious consequences. In addition to the risk of allergic reactions, ranging from a mild rash to anaphylaxis, organ damage and other adverse events may result from errors in dosing or incorrect choice of drug. Clostridium difficile infection is directly linked to antibiotic use [23]. While this risk increases with the duration of antibiotic administration, and with broad-spectrum as opposed to narrow coverage agents, even one dose of antibiotic may result in fulminant C. difficile infection. The price of MDR infections is high, both financially [22] and in terms of human health. Not only are MDR infections difficult to treat, but they are associated with higher mortality. [21, 24].

Patients with neurotrauma are highly susceptible to harm from both hospital-acquired infections and from the antimicrobial agents used to prevent or treat them. Coma or altered mental status, bedrest, increased risk of aspiration, extended ventilator dependence, and prolonged ICU stays increase the risk of ventilator-associated pneumonia, catheter-associated urinary tract infection, and central line-associated bloodstream infection. In this section, we discuss a common scenario for neurotrauma patients, that is, the use of prophylactic antibiotics for intracranial pressure monitors .

The use of prophylactic antibiotics (PAB) to prevent infection of fiberoptic intracranial pressure monitors (ICPM) and external ventricular drains (EVD) is a common and traditional practice [25], though the efficacy of this practice is still under investigation. The paucity of randomized or prospective data and the significant heterogeneity of the remaining studies account for the lack of definitive recommendations on the use of PAB for these devices. Many design characteristics must be accounted for when examining the literature on this topic, including: the type of device used (ICPM or EVD); method of infection diagnosis (insertion site infection, cultures from drains or lumbar puncture, or clinical signs); the duration of PAB administration (one pre-procedure dose, a few days, or for the duration of the device); the location of insertion (ICU, operating room); the length of time the monitor is in place; degree of sterile protocol and other technical factors pertaining to insertion; the expertise of the proceduralist (attending surgeon, trainee, midlevel practitioner); and not least, the condition for which the monitor is needed (trauma, non-traumatic subarachnoid hemorrhage, hydrocephalus, tumor).

In the setting of traumatic brain injury , the infection rate of fiberoptic intracranial pressure monitors ranges from 0 to 3.7 % [26–28]. Use of PAB has not been shown to affect central nervous system or monitor infection rates in multiple retrospective studies, when administered either as a perioperative dose or continuously for the duration of the monitor [26–29]. Intuitively this makes sense, when one considers the lack of evidence that prophylactic antibiotics affect infection rates for other percutaneous devices such as non-tunneled vascular catheters [30] and drains [31].

The risk of infection for EVDs is higher than for ICPMs, most likely due to multiple factors such as their more invasive nature, wider diameter, and attachment to drainage systems that allow one to break the circuit (e.g., to flush the catheter or change the fluid collection container). Whereas ICPM infection rarely involves more than the local surgical site, ventriculostomy-associated infections (VAI) involving ventriculitis or meningitis pose a greater risk to patients. Most retrospective studies fail to show benefit of PAB on VAI rates [27–29, 32, 33]. Only two randomized studies directly address the question of PAB for EVDs [34, 35].

Blomstedt [34] showed less early (but not late) infections with use of trimethoprim-sulfamethoxazole versus placebo in patients undergoing shunting procedures, and no difference in patients undergoing ventriculostomy procedures. A study from Hong Kong [35] randomized 228 patients receiving EVD to perioperative antibiotics only or antibiotics for the duration of the EVD, and reported a lower rate of CSF infection with prolonged antibiotics (3 %) than with perioperative antibiotics (11 %). However, no statistical methodology was reported in this paper, and with only 15 patients diagnosed with VAI its statistical power is limited. When VAI was diagnosed, more MDR pathogens were isolated in the prolonged antibiotic group, a finding replicated in other studies as well [26, 27]. It is important to keep in mind that sterile techniques in the intensive care unit have changed significantly in only the past decade (chlorhexidine skin preparation instead of betadine, head-to-toe draping of the patient instead of local draping, mandatory full sterile garb for all bedside procedures) such that the results of even fairly recent studies may have limited application to today’s healthcare environment.

The 2013 joint guidelines from the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, the Surgical Infection Society, and the Society for Healthcare Epidemiology of America state that data are insufficient to make a recommendation on use of PAB for EVD or ICPM [24]. Given the current state of evidence, the risk of MDR pathogens, and considering that use of extended prophylactic antibiotics in non-immunosuppressed patients has very few other indications, a restrained approach is preferable. If used for ICP monitors in trauma patients, PAB should be utilized in the manner as for other clean neurosurgical procedures: a single dose of pre-procedure cefazolin; clindamycin for documented beta-lactam allergy; or vancomycin with known MRSA colonization. One should also be mindful of other modifiable risk factors for VAI, including EVD presence for more than 5–7 days, suboptimal sterile placement technique, unnecessary or frequent flushing, and routine catheter exchange. [29, 32, 36, 37].

Approximately 2 % of patients with TBI who seek medical attention have a post-traumatic seizure (PTS) at some time. The risk of seizures increases with head injury severity, with rates reported as high as 12 % in patients with severe TBI, and approaching 50 % when seizure activity is diagnosed by electroencephalography [38]. A penetrating mechanism is associated with a 50 % rate of PTS [39].

Neurologic damage that occurs after a TBI occurs over hours to days. Mechanisms of injury may be divided into primary and secondary. The primary insult is the initial traumatic injury, can be focal or diffuse, and triggers a cascade of events that ultimately result in cell death. The secondary insult includes damage that occurs as a result of physiological responses to the initial injury. Since the primary insult currently cannot be therapeutically modified, therapeutic interventions target the secondary insult in an attempt to improve outcomes [40]. Secondary insults include impairment in cerebral blood flow, oxygenation, autoregulation, and metabolic function as well as PTS.

By convention, PTS that occur within seven days of injury are termed early, and those occurring after seven days are referred to as late. Risk factors for developing PTS include age, history of alcoholism, penetrating mechanism, loss of consciousness, focal neurologic deficits, GCS score <10, seizure within 24 h of injury, depressed skull fracture, hemorrhagic mass lesions, presence and location of cerebral contusion, and retained bone or foreign bodies [12, 41]. Risk factors for PTS lean heavily, though not exclusively, on findings from brain computed tomography (CT) scans. Assessment of these risk factors is important for determining the need for a prophylactic regimen against PTS.