Nutrition 23 (2007) 308 –317

www.elsevier.com/locate/nut
Applied nutritional investigation

Nutritional treatment of patients with severe traumatic brain injury

during the first six months after injury
Karolina Krakau, B.Sc.a,b,*, Annette Hansson, B.Sc.c, Torbjörn Karlsson, M.D., Ph.D.d,
Catharina Nygren de Boussard, M.D., Ph.D.b, Christer Tengvar, M.D., Ph.D.e,
and Jörgen Borg, M.D., Ph.D.a
a
Department of Neuroscience, Rehabilitation Medicine, Uppsala University Hospital, Uppsala, Sweden
b
Department of Rehabilitation Medicine Stockholm, Danderyd University Hospital, Stockholm, Sweden
c
Department of Neurosurgical Intensive Care, Karolinska University Hospital, Stockholm, Sweden
d
Department of Surgical Science, Anesthesiology and Intensive Care, Uppsala University Hospital, Uppsala, Sweden
e
Department of Rehabilitation Medicine, Central Hospital, Västerås, Sweden

Manuscript received September 28, 2006; accepted January 23, 2007.

Abstract Objective: This study explored current nutritional treatment policies and nutritional outcome in
patients with severe traumatic brain injury.
Methods: We performed a retrospective, structured survey of the medical records of 64 patients up
to 6 months after injury or until the patients were independent in nutritional administration.
Results: Enteral nutrition was administered to 86% of patients. Fourteen patients (22%) had a
gastrostomy; after 6 months four were still in use. At 6 months, 92% of patients received all food
orally and 84% had gained nutritional independence. Energy intake was equal to the calculated basal
metabolic rate throughout the first month after injury and increased by 21% during the second
month. Sixty-eight percent exhibited signs of malnourishment with weight losses of 10 –29%.
Conclusion: This study suggests that most patients with severe traumatic brain injury regain their
nutritional independence within the first 6 months after injury, but also that most develop signs of
malnutrition. © 2007 Elsevier Inc. All rights reserved.

Keywords: Traumatic brain injury; Parenteral nutrition; Enteral nutrition; Oral feeding; Level of independence; Malnu-
trition

Introduction postinjury period [5–22]. The increased energy demand

leads to a catabolic state and a net muscle protein break-
The pathophysiology of a severe traumatic brain injury down, with possible unfavorable effects on immune func-
(TBI) is dependent not only on the primary injury but also tion and morbidity [23–26]. To prevent excessive muscle
on unfavorable secondary events, e.g., epileptic seizures, protein breakdown, adequate nutrition balancing the energy
high intracranial pressure, and fever. Early detection and expenditure is crucial. There are data indicating that mor-
treatment of such events improve outcome [1– 4]. Less is bidity and mortality are decreased in patients receiving early
known about the effect of systemic metabolic disturbances feeding compared with those given standard nutritional
caused by TBI, i.e., hypermetabolism and catabolism. Con- treatment [25–29]. However, evidence is still not strong
clusive data show that patients with moderate and severe enough to establish a standard [30 –32], and evidence is
TBI have increased energy expenditure during the early lacking for the post–acute rehabilitation period.
Moreover, the need of assisted nutrition may continue for
a long period. Thus, the care will involve several units with
This study was supported by the Centre for Clinical Research, Dalarna, nutritional guidelines that may differ considerably. This
the Department of Rehabilitation Medicine, Falu Lasarett, and Fresenius
Kabi. urges the contemporary nutritional treatment of all units
* Corresponding author. Tel.: ⫹468-655-5484; fax: ⫹468-655-7754. caring for these patients during the intensive care period and
E-mail address: karolina.krakau@ds.se (K. Krakau). during the early rehabilitation period to be scrutinized, as a

0899-9007/07/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.nut.2007.01.010
 K. Krakau et al. / Nutrition 23 (2007) 308 –317 309

basis of future prospective studies. We therefore conducted Data collection

a three-center, retrospective, descriptive study in Sweden to
delineate the nutritional treatment of adult patients with Data were extracted from patient records covering the
severe TBI and to analyze their nutritional outcome during period from the injury until the patients became nutrition-
the first 6 mo after injury or until the patients were inde- ally independent or, if not, until 6 mo after injury. Data
pendent regarding administration of nutrition. extraction and analysis aimed at answering the following
questions.
1. How long was the need for parenteral nutrition (PN)
Materials and methods and enteral nutrition (EN)?
2. How many patients received a gastrostomy? When
Patients was it placed and for how long was it needed?
3. How long was the need for assisted feeding?
The study took place in the Stockholm, Uppsala, and 4. What was the relation between energy intake (EI) and
Gothenburg regions, with a total population of approxi- calculated basal energy metabolism?
mately 5.3 million inhabitants. All patients with TBI admit- 5. How did the patients’ body mass change over time?
ted to one of the three regional neurosurgical intensive care 6. How many patients fulfilled criteria for malnourish-
units during a period of 1 y (July 1, 2003 to June 30, 2004) ment?
were screened for inclusion into the study. We included 7. What was the frequency and possible effect of “im-
patients 16 to 64 y of age with severe TBI, i.e., patients portant factors” in malnourished compared with non-
scoring 3– 8 on the Glasgow Coma Scale (GCS). To control malnourished patients?
for confounding factors such as hypothermia or intoxica- Documentation systems varied considerably between
tion, rating scores during the first 24 h were examined. In units having different computerized and manual systems.
deteriorating patients, the worst score before medical seda- Data sources were mainly medical and nursing records but,
tion was recorded, and in patients improving without any when available, also records kept by occupational thera-
interventions, the best score was recorded. Exclusion crite- pists, speech-language pathologists, physiotherapists, and
ria were other major injuries or diseases with an influence dietitians. Data were extracted to a preformed protocol by
on metabolism or nutritional treatment, as presented in the two of the present investigators. To secure adequate data
flow chart in Figure 1. Patients were contacted directly or by collection, each care unit provided an experienced em-
an authorized representative. The study was approved by the ployee who introduced the documentation system to the
regional ethical review board of Uppsala University (Upp- investigators when they were unfamiliar with it. Further
sala, Sweden). contact was made whenever data to the preformed protocol

Fig. 1. Flow chart of inclusion and exclusion procedures of patients. Values are expressed as numbers of patients. G, Gothenburg region; S, Stockholm
region; TBI, traumatic brain injury; U, Uppsala region.
310 K. Krakau et al. / Nutrition 23 (2007) 308 –317

were not found by the investigators. Data collection took Energy intake
place locally in all but nine units, where instead journal
copies were sent for. The daily delivered intake of energy was recorded on
The daily recordings made at the units were carefully separate nutritional protocols mainly during PN and EN
examined for the selected data. The nutritional data sought periods, but occasionally even when patients were fed
for and, when available, extracted to the preformed protocol orally. Nutritional deliveries noted in other parts of the
concerned the ways of administering nutritional support, the medical record were also checked and EI recalculated when
nutritional prescriptions, the predicted energy need, the en- necessary, i.e., when the intake of PN and EN were listed
ergy expenditure measured by indirect calorimetry, and the separately and not added together, or when intake was
EI. Moreover, patients’ body height and body weight data registered in text and volume instead of translated into the
were extracted. When body height was not found in the caloric amount. Standardized tables on the energy content of
medical record, the patient, a family member, or nursing different types of food, drinks, EN, and PN were used to
staff was contacted, the choice of contact depending on the recalculate these protocols. When patients were on oral
recovery of the patient. Family members were asked to seek feeding and, e.g., half a portion was registered as eaten, we
for this information in a valid passport of the patient and the regarded this as half of the energy content of this patient’s
nursing staff was asked to measure it. Another part of the specific diet even though it was unknown what had been left
protocol concerned factors influencing energy demands, i.e., on the plate, the potatoes or the meat. Information on the
other injuries, body temperature, pharmacologic sedation, specific diets energy content per portion was supplied by the
physical anxiety, wasted residual volumes, vomiting, and kitchen of the local hospital.
appetite, where any data available concerning this were also
extracted. The protocol also concerned the recovery of pa- EI and BMR
tients, which is the reason all scoring results from the rating
scales GCS, Functional Independence Measure (FIM), and Data on the total energy expenditure according to indi-
Glasgow Outcome Scale were extracted. rect calorimetry [34] were too sporadic to allow any eval-
uation of the energy balance. When analyzing the EI of the
patients, we related it to the basal metabolic rate (BMR),
Length of stay
i.e., what the energy expenditure of these patients would
have been before the trauma, while at rest and fasting. We
The length of stay was defined as the period from the
used the Harris-Benedict equation [35], based on admission
injury until discharge to home without any further assis-
weight, height, age, and gender to calculate this.
tance needed, adjusted living with assistance given at home,
nursing home, or death.
Factors with a possible effect on energy demands

Assisted nutrition and level of nutritional independence The length of sedation, i.e., the use of benzodiaz-
epines, barbiturates, and propofol, was extracted to pro-
Assisted nutrition was defined as the state when help was tocol. Due to difficulties in evaluating the sedative effect
required for administering nutrition parenterally, enterally retrospectively, the dosage was not specified. Body tem-
or orally. Any verbal or physical guidance needed for ini- perature data were extracted as days of fever (ⱖ38°C) or
tiation or completion of meals was also considered as as- hypothermia (⬍36°C).
sisted feeding, i.e., scorings ⱕ5 according to the FIM [33]. No scales to measure physical anxiety of the patients
The FIM was the most common assessment tool used in were identified in the records. Therefore data in this respect
scoring the actual level of independence in eating. However, were extracted from free text in the records, e.g., from notes
FIM was used only in rehabilitation settings; some patients on the patient being easily stressed, sweating profusely,
were not scored at all and some began eating while cared for restless, significantly spastic, or agitated.
in other care settings. In these patients we based the re- Vomiting and wasted residual volumes were recorded
corded level of independence on the daily nutritional notes. and patients with such problems for ⱖ5 d were defined as
If it was stated in the nutritional notes that “the patient now having possible upper gastrointestinal intolerance.
eats unattended and with good appetite” and the weight
development from this day did not diverge in any direction, Malnutrition
the date for this statement was considered the time point for
when assisted feeding ended. If no FIM had been used and The term malnutrition is used to describe an imbalance
only scarce nutritional notes were found, we used change in between intake and depletion that leads to over- or under-
body weight as a proxy measurement. Thus, patients with feeding. In this study the term is used only in the sense of
considerable weight loss or weight gain were not defined as energy deficiency, i.e., undernourishment. The anthropo-
independent in eating, although they might have been al- metric data available to evaluate malnourishment were body
lowed to eat unassisted. mass index (BMI) and weight change over time. According
 K. Krakau et al. / Nutrition 23 (2007) 308 –317 311

to the Malnutritional Universal Screening Tool [36], there is

a risk for malnutrition when BMI is 18.5–20 kg/m2 or if
there is an involuntary weight loss of 5–10% over 3– 6 mo
and the risk for malnutrition is considered high under the
following conditions.
1. Both of the previously mentioned risk factors occur at
the same time.
2. The BMI is ⬍18.5 kg/m2.
3. Involuntary weight loss exceeds 10%.
Patients in this study were considered malnourished
when any of these three high-risk criteria was met.
Fig. 2. Number of patients receiving PN or EN or capable of OF during the
first 6 mo after injury. EN, enteral nutrition; OF, oral feeding; PN, paren-
Statistical analysis teral nutrition.

In text, diagrams, and tables, the frequencies and quote

(45%) and falls (22%); less common were assaults and
data are presented as mean ⫾ SD or, with non-normal
sports or other injuries.
distributions, as median and interquartile range (IQR). Clin-
Total length of hospital stay was 9 to 356 d and number
ical factors that might influence body mass were analyzed
of care units during this period was three to eight. During
by comparing malnourished with non-malnourished pa-
the first 2 mo, there were 4.7 ⫾ 0.96 transferals between
tients. These variables were tested for association by use of
different care units per patient. For most patients (73%),
Pearson chi-square test, defining statistically significant dif-
time in the hospital included three periods: an intensive care
ferences at P ⬍ 0.05.
period followed by a period in different ward settings,
mainly surgical or medical wards but also orthopedic care,
psychiatric care, and nursing homes, and then in a special-
Results ized rehabilitation setting. Six patients went straight from
intensive care units to rehabilitation, whereas 11 patients
Demographics, injury data, and levels of care were never transferred to a rehabilitation setting. At dis-
charge from hospital care, 64% of patients (n ⫽ 41) were
Patient characteristics are summarized in Table 1. Al- able to go home without any assistance, whereas 22% (n ⫽
most 50% (29 of 64) of the patients exhibited GCS scores 14) needed further assistance at home. Eight patients
within the lowest part of the severity range, i.e., 3 to 5. The (12.5%) were discharged to nursing homes and one patient
most common injury mechanisms were traffic accidents died during hospital stay 1 mo after injury due to a major
stroke.

Table 1 Duration of PN and EN

Patient and injury characteristics*
Age (y) 35 ⫾ 15.7 All patients received PN while in the intensive care unit.
Women 11 The total number of days on PN was 2 to 64 d (19 ⫾ 12.2 d).
Men 53
Body mass (kg) 75 ⫾ 12.6
Most patients (86%) also received EN, whereas 14% never
Height (cm) 178 ⫾ 8.4 had EN and could start oral feeding immediately after PN.
BMI (kg/m2) 24 ⫾ 4.3 EN started 4 ⫾ 2.9 d after injury and the total days of EN
Isolated EDH or SDH 12 varied between 1 and 178 (median 14 d, IQR 33) during the
Cerebral contusion and other brain injuries 52 first 6 mo after injury.
GCS score 3–5 29
GCS score 6–8 35
As demonstrated in Figure 2, change in the way of
No fractures 25 administering nutrition was most frequent during the first
ⱖ1 fracture 39 month. After approximately 3 mo, a final change had been
Total LOS (d) 76.5 (95) made in most patients (⬎90%), but a few patients continued
LOS in ICU (d) 17 (14) to make progress in this respect up to 5 mo after injury.
LOS in different ward settings (d) 17 (15)
LOS in rehabilitation unit (d) 55 (68)
Use of gastrostomy
BMI, body mass index; EDH, epidural hematoma; GCS, Glasgow Coma
Scale; ICU, intensive care unit; LOS, length of stay; SDH, subdural
hematoma Fifty-five patients received EN and 14 of these patients
* Values are presented as numbers, mean ⫾ SD, or median (interquartile received a gastrostomy approximately 1 mo after injury
range). (31 ⫾ 8.6 d). Ten of these patients could return to total oral
312 K. Krakau et al. / Nutrition 23 (2007) 308 –317

feeding after 77 ⫾ 31.3 d. The remaining four patients protocols of EI were mainly registered while the patients
remained dependent on gastrostomy 6 mo after injury. were on PN and/or EN, but when they started to eat, the
recordings were more sparse. From the onset of oral feeding
Course of assisted feeding the daily total EI was registered in only 25% of the orally
fed patients (n ⫽ 15/n ⫽ 59) for 8 ⫾ 8 d. Thus, data for
From being totally dependent on assisted feeding, pa- delivered energy decreased over time and after 2 mo fewer
tients gradually gained more control of their nutritional than 10 patients could be evaluated. The analysis covers
intake. The time to the first recorded swallowing test varied only the first 2 mo (Fig. 3). BMR could be calculated in 57
but was mostly within 1 mo after injury (median 16 d, IQR patients and 2 patients had no EI registrations. This left 56
22). According to measurements of independence (FIM, patients to be evaluated. The EI throughout the first month
nutritional notes, or unfavorable changes in body mass), the was just below the calculated BMR (EI/BMR ⫽ 0.99 ⫾
period requiring assisted feeding varied a lot, from 9 to 0.16) and stayed at 21% above BMR (EI/BMR ⫽ 1.21 ⫾
180 d (median 37.5 d, IQR 61). At 6 mo after injury 54 0.09) during the second month.
patients (84%) had gained total nutritional independence.
This group of patients started to receive all their food orally Change in body mass
4 to 119 d after injury (median 21 d, IQR 18). Five (8%) of
the remaining dependent patients received all their food Weight development was evaluated in 56 patients, but
orally but needed assistance; two (3%) were fed with a the frequency of measurements varied between individuals
combination of oral nutrition and EN and two could not and between care units, which hampered our evaluation.
swallow any food orally and were given all food enterally. Body weight was measured throughout the study period in
The patient who died 1 mo after injury had received PN and 32 patients. Eleven patients were weighed initially and then
EN throughout this period. only sporadically and 13 patients were weighed only during
the initial intensive care period. Eight patients were with-
EI in relation to BMR drawn from analysis because they were weighed no more
than three times. Adequate data for evaluation of mean
Energy intake recordings were done in all but two pa- values were restricted to the first month because the fre-
tients at an average of 31 ⫾ 38 d (range 1–176 d). Daily quency of weight recordings then decreased (Fig. 4).

Fig. 3. Daily mean ratio of EI to BMR during the first 2 mo after injury in 56 patients. Percentages of patients evaluated per day: 100 –50% days 1–18,
50 –25% days 19 –28, 25–14% days 29 – 60. BMR, basal metabolic rate; EI, energy intake.
 K. Krakau et al. / Nutrition 23 (2007) 308 –317 313

Fig. 4. Daily alteration in body mass during the first month after injury in 56 patients. Each mark represents the mean value of ratios between current and
initial body weights in the group. n, number of patients included per day.

Compared with the first recorded measurement after 3 ⫾ criterion 3. Fourteen of the patients fulfilled both these
2.9 d, all but four patients had initial weight loss. In eight high-risk criteria.
patients (14%), this was ⬍5%; in another eight, it was In total, the number of malnourished patients according
5–10%; and in 37 (66%), it was ⬎10%. The weight de- to the criteria was 38 (68%). Most patients had their lowest
crease lasted 1–2 mo after injury in most patients and was body weight recordings during the second month after in-
followed by a slow return toward the initial weight. Indi- jury (Table 2).
vidual measurements are delineated in Figure 5 and dem-
onstrate weight development during the entire study period.
Malnourished versus non-malnourished patients
Malnourished patients
The malnourished and non-malnourished patients
Anthropometric data included body mass, height, and were compared regarding factors that might have an
BMI. No data on nutritional status in the patients before
injury was found.
The change in weight during the study period could be
calculated in 56 patients and the development of BMI was
possible to calculate in 53 patients. Nine patients had a BMI
⬍20 kg/m2 (mean 18.5, range 16.9 –19.9) at the time of
injury. Malnutrition was defined according to the three
high-risk criteria of the Malnutritional Universal Screening
Tool (Table 2).
Sixteen patients fulfilled one of the low-risk criteria. No
patient had the two low-risk criteria combined, but 7 of
these 16 patients were weighed only during the period early
after injury. Thus, it is uncertain how they developed during
the proceeding hospital period.
Fifteen patients were malnourished according to high- Fig. 5. Each mark represents the ratio between the current weight and
risk criterion 2, and 37 patients were according to high-risk initial weight in patients (n ⫽ 56) during the first 6 mo after injury.
314 K. Krakau et al. / Nutrition 23 (2007) 308 –317

Table 2
Malnourishment in patients with severe traumatic brain injury according to risk criteria of the Malnutritional Universal Screening Tool
Risk criteria Percentage Malnourished/total Patients, Patients, Days after injury,
of patients patients mean ⫾ SD minimum–maximum mean ⫾ SD

Low-risk criteria
BMI 20–18.5 kg/m2 15% 8/53 19.3 ⫾ 0.62 18.5–19.9 29 ⫾ 20.4
Weight loss 5–10% 14% 8/56 6.8 ⫾ 1.4 5–8.6 24 ⫾ 11.8
High-risk criteria
Both low-risk criteria combined 0% 0 0 0 0
BMI ⬍18.5 kg/m2 28% 15/53 16.7 ⫾ 1.4 13.9–18.3 48 ⫾ 26.1
Weight loss ⬎10% 66% 37/56 16.2 ⫾ 4.9 10.1–29.2 49 ⫾ 26.5

BMI, body mass index

effect. There was no difference in age, gender, severity, Discussion

days of hypothermia, fever, or days of physical anxiety
between malnourished and non-malnourished patients Findings
(Table 3). However, associated fractures, days on seda-
tion, upper gastrointestinal intolerance, length of stay, The two main findings of this detailed, retrospective
and length of stay in an intensive care unit differed review of medical records were that 68% of patients with
between groups. severe TBI exhibited signs of malnutrition extending to the
Energy intake data according to the nutritional protocols second month after injury and that 84% became nutrition-
could be compared only 18 d after injury because the nu- ally independent within the first 6 mo after injury.
tritional protocols in non-malnourished patients ended by The increased risk associated with hypermetabolism in
then. During this time, the two groups were given the same patients with severe TBI during the first weeks after injury
amount of energy in relation to BMR. has been extensively described in the literature [5–22] and

Table 3
Proportions and differences between malnourished and non-malnourished patients
Variables Dichotomized variable Malnourished Non-malnourished Pearson chi-square asymptotic
patients (n ⫽ 38) patients (n ⫽ 18) significance (2-sided)

Age 16–30 19 10
31–64 19 8 0.698
Gender Women 6 4
Men 32 14 0.557
Severity GCS score 3–5 18 9
GCS score 6–8 20 9 0.854
Accompanied fractures 0 fracture 10 11
ⱖ1 fracture 28 7 0.012*
Days on sedation Median ⱕ12 d 13 15
Median ⬎12 d 24 3 0.001*
Missing 1
Days of hypothermia 0d 19 14
ⱖ1 d 18 4 0.061
Missing 1
Days of fever Median ⱕ9 d 13 11
Median ⬎9 d 24 7 0.068
Missing 1
Days of physical anxiety 0d 18 10
ⱖ1 d 20 8 0.567
Upper gastrointestinal intolerance ⱕ5 d 13 14
⬎5 d 25 4 0.002*
Length of stay Median ⱕ76.5 d 12 13
Median ⬎76.5 d 26 5 0.004*
Length of stay in intensive care unit Median ⱕ17 d 13 12
Median ⬎17 d 25 16 0.023*

GCS, Glasgow Coma Scale

* P ⬍ 0.05.
 K. Krakau et al. / Nutrition 23 (2007) 308 –317 315

thus malnutrition was not unexpected [37– 40]. However, are generally are avoided. Further, any sudden weight loss
the weight development in these patients showed that body approximately 3 mo after injury might be due to removal of
mass continued to decrease for weeks after the early hyper- plasters or external fixations, but there was no information
metabolic period. Presumably, the duration of hypermetab- identified in that respect. Thus the observed weight loss of
olism was not restricted to the early postinjury period but 10 –29% in the malnourished patients is most unlikely ex-
was extended over a longer period or other factors not plained by the regain in fluid balance or by the removal of
captured in this study mattered, as discussed below. external fixations or plasters.
First, the quality of nutritional care might have been It may also be pointed out that, to prevent deterioration
threatened by practical conditions. Ways of administering of nutritional status, the decision to place a gastrostomy
nutrition changed frequently during the first and second catheter should be considered at an early stage, i.e., when
months after injury and patients were often transferred be- nutritional intake is expected to be inadequate for 2–3 wk,
tween different care units. It is a demanding task to maintain taking the clinical situation and ethical issues of the patient
optimal nutritional care and to secure adequate information in account [43]. In this study, 14 of the 55 enterally fed
when handing over the patients. If routines for nutritional patients received gastrostomy at 1 mo after injury, when
assessment are missing or inadequate, this most likely en- weight loss had already advanced. Although 10 of these
dangers the nutritional status of the patients. For instance, patients needed gastrostomy for only a few months, it is
body height was seldom registered which is why BMI was possible that this prevented additional weight loss because
often missed as a risk indicator for development of malnu- repeated removals of nasogastric tubes are common during
trition. Nine patients had a BMI ⬍20 kg/m2 and one of these stages of anxiety and confusion [44 – 46].
had a BMI ⬍18.5 kg/m2 at the time of injury. Early detec- Special attention should probably be given to patients in
tion of low BMI is crucial because these patients may be vegetative state because they frequently develop complica-
more vulnerable to catabolism. tions that affect energy need e.g., epilepsy, dysautonomy
Second, assessing energy need is not a standardized (i.e., fever, tachycardia, pronounced sweating, high blood
procedure because no equation has yet been proved valid for pressure), muscular hypertonus, infections, venous throm-
patients with severe TBI [41,42]. If indirect calorimetry boembolism, and respiratory and gastrointestinal dysfunc-
measurements are not performed and instead generalized tions. Energy demands of 40 – 60 kcal · kg⫺1 · d⫺1 and a
equations regarding energy need are used for calculations, 20 – 48% frequency of malnourishment have been reported
there is a risk for over- and underfeeding. In this investiga- in this subgroup of patients [39].
tion indirect calorimetry was used in only a few patients, Moreover, pituitary hormone deficiencies have been
sporadically, and only during the intensive care period, identified in a substantial proportion of patients with TBI
which is the reason energy balances could not be adequately [47– 49], causing metabolic disruptions that may affect
evaluated. However, overfeeding did not seem to be a prob- weight development.
lem because EI during the first month covered merely the In our study the time needed for assisted feeding varied
BMR and thus obviously not the energy need of the patients. greatly. This should be interpreted with caution because the
During the second month it was increased by only 21%. timing for rating the independence at the units may not have
Energy balance was probably negative in most patients, but, been congruent with the actual time for the patient’s inde-
because not all developed malnutrition, other factors were pendence, which may have occurred earlier. One conspic-
important. It is also questionable whether the “indepen- uous finding, however, was that most patients with severe
dence in eating” really tell us that the patient is able to TBI regained their eating function within the first 6 mo after
control EI in an adequate way. We can not answer this injury. Only a minority depended on continued assisted
question in the present investigation, but the question needs feeding. Further, the first recorded swallowing test was done
to be asked. within 1 mo after injury. The time for the first safe oral
Third, body mass measurements should be interpreted feeding has been reported to be a reliable predictor of final
with some caution because various care units used different outcome in patients with severe TBI [50]. However, be-
scales and it was not always clear whether patients were cause the time for the first safe oral feeding could not be
weighed under standardized procedures. The recorded weight stated from our data, its predictive value for the final out-
values were, despite their impreciseness, the ones clinicians come could not be tested.
had based their interpretations on.
The first body mass measurement available was per- Limitations of the study
formed on day 3. Thus patients might then already have
gained weight due to positive fluid balance and fluid reten- The retrospective design of this study enabled the delin-
tion. Because the first measurement serves as a basic level eation of current nutritional treatment without interfering
in this study, the proceeding weight loss would then be with the treatment process and the results are useful for
explained by the regaining of fluid balance. However, fluids planning prospective studies. The findings in this study
are restricted to keep intracranial pressure down, explaining must however consider all the weaknesses inherent in a
why extensive positive fluid balances in patients with TBI retrospective review of medical records. Although the find-
316 K. Krakau et al. / Nutrition 23 (2007) 308 –317

ings are based on a large set of detailed, clinical data, [7] Chiolero R, Schutz Y, Lemarchand T, Felber JP, de Tribolet N,
extracted according to a preformed protocol, factors not Freeman J, Jequier E. Hormonal and metabolic changes following
severe head injury or noncranial injury. JPEN 1989;13:5–12.
measured in clinical practice and thus not present in this
[8] Haider W, Lackner F, Schlick W, Benzer H, Gerstenbrand F, Irsigler
delineation may have been important for the development of K, et al. Metabolic changes in the course of severe acute brain
malnutrition. Moreover, data do not allow any evaluation of damage. Eur J Intensive Care Med 1975;1:19 –26.
the possible effect of malnutrition on complication rate or [9] Clifton GL, Ziegler MG, Grossman RG. Circulating catecholamines
final outcome. and sympathetic activity after head injury. Neurosurgery 1981;8:
10 – 4.
[10] Clifton GL, Robertson CS, Grossman RG, Hodge S, Foltz R, Garza
Future studies C. The metabolic response to severe head injury. J Neurosurg 1984;
60:687–96.
Because evidence-based standards for nutritional assess- [11] Fell D, Benner B, Billings A, Siemens R, Harbison B, Newmark SR.
ment and treatment of patients with severe TBI are lacking, Metabolic profiles in patients with acute neurosurgical injuries. Crit
further studies are urgently needed. These studies should Care Med 1984;12:649 –52.
[12] Young B, Ott L, Norton J, Tibbs P, Rapp R, McClain C, Dempsey R.
define the level of EI in relation to energy expenditure Metabolic and nutritional sequelae in the non-steroid treated head
during a relevant period that has the best effect on outcome. injury patient. Neurosurgery 1985;17:784 –91.
They should also delineate the duration of catabolism. Eas- [13] Robertson CS, Clifton GL, Goodman JC. Steroid administration and
ily manageable assessment tools for clinical practice that nitrogen excretion in the head-injured patient. J Neurosurg 1985;63:
captures the diverse nutritional demands of these patients is 714 – 8.
[14] Moore R, Najarian MP, Konvolinka CW. Measured energy expendi-
also a topical area for research to prevent the development
ture in severe head trauma. J Trauma 1989;29:1633– 6.
of malnutrition. [15] Dickerson RN, Guenter PA, Gennarelli TA, Dempsey DT, Mullen JL.
Increased contribution of protein oxidation to energy expenditure in
head-injured patients. J Am Coll Nutr 1990;9:86 – 8.
Conclusion [16] Bruder N, Lassegue D, Pelissier D, Graziani N, Francois G. Energy
expenditure and withdrawal of sedation in severe head-injured pa-
tients. Crit Care Med 1994;22:1114 –9.
This investigation shows that most patients with severe [17] Bruder N, Raynal M, Pellissier D, Courtinat C, Francois G. Influence
TBI develop malnutrition, but also that most regain their of body temperature, with or without sedation, on energy expenditure
independence in eating within the first 6 months. in severe head-injured patients. Crit Care Med 1998;26:568 –72.
[18] Feldman Z, Contant CF, Pahwa R, Goodman JC, Robertson CS,
Narayan RK, Grossman RG. The relationship between hormonal
mediators and systemic hypermetabolism after severe head injury.
Acknowledgments J Trauma 1993;34:806 –16.
[19] Charlin V, Carrasco F, Ferrer L, Brito A, Poblete R. [Protein and
The authors express their gratitude to the patients and energy requirements in patients with severe head injury]. Rev Med
care units that participated in this study. They also thank Chil 1993;121:626 –32.
Marianne Omne-Pontén for most valuable comments and [20] Fugazza G, Aquilani R, Iadarola P, Dossena M, Catapano M, Boschi
F, et al. The persistence of hypercatabolic state in rehabilitation
Håkan Källmen, Ph.D., for statistical assistance.
patients with complicated head injury. Eur Medicophys 1998;34:
125–9.
[21] Weekes E, Elia M. Observations on the patterns of 24-hour energy
References expenditure changes in body composition and gastric emptying in
head-injured patients receiving nasogastric tube feeding. JPEN 1996;
[1] Elf K, Nilsson P, Enblad P. Outcome after traumatic brain injury 20:31–7.
improved by an organized secondary insult program and standardized [22] Raurich JM, Ibanez J. Metabolic rate in severe head trauma. JPEN
neurointensive care. Crit Care Med 2002;30:2129 –34. 1994;18:521– 4.
[2] Signorini DF, Andrews PJ, Jones PA, Wardlaw JM, Miller JD. Add- [23] Quattrocchi KB, Issel BW, Miller CH, Frank EH, Wagner FC Jr.
ing insult to injury: the prognostic value of early secondary insults for Impairment of helper T-cell function following severe head injury.
survival after traumatic brain injury. J Neurol Neurosurg Psychiatry J Neurotrauma 1992;9:1–9.
1999;66:26 –31. [24] Wolach B, Sazbon L, Gavrieli R, Broda A, Schlesinger M. Early
[3] Becker DP, Miller JD, Ward JD, Greenberg RP, Young HF, Sakalas immunological defects in comatose patients after acute brain injury.
R. The outcome from severe head injury with early diagnosis and J Neurosurg 2001;94:706 –11.
intensive management. J Neurosurg 1977;47:491–502. [25] Minard G, Kudsk KA, Melton S, Patton JH, Tolley EA. Early versus
[4] Elf K, Nilsson P, Ronne-Engstrom E, Howells T, Enblad P. Cerebral delayed feeding with an immune-enhancing diet in patients with
perfusion pressure between 50 and 60 mm Hg may be beneficial in severe head injuries. JPEN 2000;24:145–9.
head-injured patients: a computerized secondary insult monitoring [26] Sacks GS, Brown RO, Teague D, Dickerson RN, Tolley EA, Kudsk
study. Neurosurgery 2005;56:962–71. KA. Early nutrition support modifies immune function in patients
[5] McCall M, Jeejeebhoy K, Pencharz P, Moulton R. Effect of neuro- sustaining severe head injury. JPEN 1995;19:387–92.
muscular blockade on energy expenditure in patients with severe head [27] Young B, Ott L, Haack D, Twyman D, Combs D, Oexmann JB, et al.
injury. JPEN 2003;27:27–35. Effect of total parenteral nutrition upon intracranial pressure in severe
[6] Bruder N, Dumont JC, Francois G. Evolution of energy expenditure head injury. J Neurosurg 1987;67:76 – 80.
and nitrogen excretion in severe head-injured patients. Crit Care Med [28] Taylor SJ, Fettes SB, Jewkes C, Nelson RJ. Prospective, randomized,
1991;19:43– 8. controlled trial to determine the effect of early enhanced enteral
 K. Krakau et al. / Nutrition 23 (2007) 308 –317 317

nutrition on clinical outcome in mechanically ventilated patients [40] Thomson MA, Carver AD, Sloan RL. Nutritional status of traumatic
suffering head injury. Crit Care Med 1999;27:2525–31. and anoxic brain-injured patients on admission to rehabilitation
[29] Rapp RP, Young B, Twyman D, Bivins BA, Haack D, Tibbs PA, (abstract). Proc Nutr Soc Lond 2000;60:83A.
Bean JR. The favorable effect of early parenteral feeding on survival [41] Sunderland PM, Heilbrun MP. Estimating energy expenditure in
in head-injured patients. J Neurosurg 1983;58:906 –12. traumatic brain injury: comparison of indirect calorimetry with pre-
[30] Krakau K, Omne-Ponten M, Karlsson T, Borg J. Metabolism and dictive formulas. Neurosurgery 1992;31:246 –53.
nutrition in patients with moderate and severe traumatic brain injury: [42] Clifton GL, Robertson CS, Choi SC. Assessment of nutritional re-
a systematic review. Brain Inj 2006;20:345– 67. quirements of head-injured patients. J Neurosurg 1986;64:895–901.
[31] Brain Trauma Foundation, American Association of Neurological [43] Loser C, Aschl G, Hebuterne X, Mathus-Vliegen EM, Muscaritoli M,
Surgeons, Joint Section on Neurotrauma and Critical Care. Nutrition. Niv Y, et al. ESPEN guidelines on artificial enteral nutrition—
J Neurotrauma 2000;17:539 – 47. percutaneous endoscopic gastrostomy (PEG). Clin Nutr 2005;24:
[32] Yanagawa T, Bunn F, Roberts I, Wentz R, Pierro A. Nutritional 848 – 61.
support for head-injured patients. Cochrane Database Syst Rev 2002: [44] Stechmiller J, Treloar DM, Derrico D, Yarandi H, Guin P. Interrup-
tion of enteral feedings in head injured patients. J Neurosci Nurs
CD001530.
1994;26:224 –9.
[33] Kidd D, Stewart G, Baldry J, Johnson J, Rossiter D, Petruckevitch A,
[45] Ott L, Annis K, Hatton J, McClain M, Young B. Postpyloric enteral
Thompson AJ. The Functional Independence Measure: a comparative
feeding costs for patients with severe head injury: blind placement,
validity and reliability study. Disabil Rehabil 1995;17:10 – 4.
endoscopy, and PEG/J versus TPN. J Neurotrauma 1999;16:233– 42.
[34] Branson RD. The measurement of energy expenditure: Instrumenta-
[46] Klodell CT, Carroll M, Carrillo EH, Spain DA. Routine intragastric
tion, practical considerations and clinical application. Respir Care
feeding following traumatic brain injury is safe and well tolerated.
1990;35:640 –59. Am J Surg 2000;179:168 –71.
[35] Harris J, Benedict F. A biometric study of basal metabolism in man. [47] Aimaretti G, Ambrosio MR, Di Somma C, Gasperi M, Cannavo S,
Washington D.C.: Carnegie Institute of Washington, 1919;223–50. Scaroni C, et al. Hypopituitarism induced by traumatic brain injury in
[36] Kondrup J, Allison SP, Elia M, Vellas B, Plauth M. ESPEN guide- the transition phase. J Endocrinol Invest 2005;28:984 –9.
lines for nutrition screening 2002. Clin Nutr 2003;22:415–21. [48] Benvenga S, Campenni A, Ruggeri RM, Trimarchi F. Clinical review
[37] Denes Z. The influence of severe malnutrition on rehabilitation in 113: hypopituitarism secondary to head trauma. J Clin Endocrinol
patients with severe head injury. Disabil Rehabil 2004;26:1163–5. Metab 2000;85:1353– 61.
[38] Aquilani R, Viglio S, Iadarola P, Guarnaschelli C, Arrigoni N, [49] Lieberman SA, Oberoi AL, Gilkison CR, Masel BE, Urban RJ.
Fugazza G, et al. Peripheral plasma amino acid abnormalities in Prevalence of neuroendocrine dysfunction in patients recovering from
rehabilitation patients with severe brain injury. Arch Phys Med Re- traumatic brain injury. J Clin Endocrinol Metab 2001;86:2752– 6.
habil 2000;81:176 – 81. [50] Formisano R, Voogt RD, Buzzi MG, Vinicola V, Penta F, Peppe A,
[39] Dolce G, Sazbon L. The post-traumatic vegetative state. Stuttgard: Stanzione P. Time interval of oral feeding recovery as a prognostic
Thieme; 2002, p. 22. factor in severe traumatic brain injury. Brain Inj 2004;18:103–9.