The type of formula and the timing of enteral nutrition support in patients with thermal injuries has long been debated. What is known is that burn injury can result in profound metabolic abnormalities, and that malnourishment is associated with increased infection risks, decreased healing rates, and altered cell function. The rapid removal of devitalized burn tissue, combined with early nutritional support, appears to significantly attenuate this auto-destructive process (Demling 2000).
After the thermal injury has occurred a cascade of events are set in motion. An initial "Ebb" phase occurs immediately after the thermal injury, and corresponds to a period of hypovolaemia and tissue hypoxia, poor cardiac output and increased oxygen consumption. The "Flow" phase occurs within 12 to 24 hours post initial insult as the patient is fluid resuscitated and a return of an adequate oxygen transport occurs. At this stage an increase in glucose production and free fatty acid release takes place (Mahan 1996). The metabolic response to burns is complex and is illustrated by an accelerated catabolism leading to a negative nitrogen balance and loss of lean body tissue (primarily skeletal muscle).
Patients with large burns have significant energy requirements. Basal metabolic rates can double when burns are greater than 50% total body surface area (TBSA). These patients, as well as those with inhalation injury who require mechanical ventilation, are not able to meet their requirements for macro- and micronutrients and fluids via the oral route. Under these circumstances, enteral nutrition support is indicated.
Enteral nutrition is provided by placing a flexible feeding tube via the nose, mouth or a formed tract (i.e. gastrostomy) into the stomach or proximal small intestine. The feeding tube delivers liquid nutritional formula (enteral nutrition), which contains macro- and micronutrients. Enteral nutrition will be provided until sufficient oral intake is established to meet the patient's requirements.
The Eastern Association for the Surgery of Trauma (EAST) group (EAST 2001) recommends that enteral nutrition should ideally begin during resuscitation in a patient with severe burns, but because delays (>18 hours) result in a high risk of gastroparesis and need for intravenous nutrition, a fear of complications normally delays enteral support until after the resuscitation period has transpired (McDonald 1991).
Although providing nutrition is clearly essential in the successful management of the burn injured patient, there are several conflicting findings amongst research groups regarding the optimal method and timing of enteral nutritional support. Although providing early nutritional support has a number of advantages including increased caloric intake (Gottschlich 2002) and improving bowl mucosal integrity (Peng 2001), it has remained unclear whether early enteral nutritional support has any beneficial impact on a diverse field of nutritional, metabolic and biochemical outcomes and clinical indicators such as length of stay, infection rates and mortality (Peck 2004). In addition, it has been suggested that more complications occur with early enteral support than late enteral support, including the risk of intestinal necrosis (Gottschlich 2002).
As suggested by Peck 2004 the result of these and other studies reveal that the clinical efficacy of early enteral nutrition and its effect on metabolic, and hormonal and biochemical markers, remains unresolved. While the provision of enteral nutritional support in patients with major burn injuries continues to be an accepted therapeutic modality, there is an urgent need to review the evidence regarding its safety and efficacy, optimal timing and formulation.
The objective of this review was to establish the relative safety and effectiveness of early versus delayed enteral nutrition support in patients with burn injuries.
Criteria for considering studies for this review
Types of studies
We included all randomised controlled trials (RCTs) that evaluated the safety and effectiveness of early versus delayed nutritional support in patients with burn injury.
Types of participants
We focused on adults (>16 years) with any burn injury to the epidermis, subcutaneous tissues, vessels, nerve, tendons, or bone.
Types of interventions
Trials comparing early enteral nutritional support (within 24 hours of injury) versus late enteral nutritional support (greater than 25 hours). In this instance, enteral nutrition (EN) was defined as the delivery of a liquid nutritional formula, which contains macro- and micronutrients such as carbohydrates, fat, protein, vitamins, mineral and trace elements that passes any part of the digestive tract, regardless of the method of deliver (e.g. nasogastric, naso-jejunal, naso-enteric, oro-gastric, PEG or jejunal feeding tubes). Studies examining combined enteral and oral supplementation feeding regimes were also included.
Types of outcome measures
Clinically important measures of effectiveness were selected as follows:
1. All cause mortality at end of follow-up
2. Length of hospital stay
3. Frequency of infection
4. Number of adverse events such as bowel necrosis, acute respiratory distress syndrome, renal failure, multisystem organ failure
2. Biochemical markers such as albumin, white cell count and C-reactive protein
(Secondary parameters such as those listed above were considered if reported in the studies.)
Search methods for identification of studies
The searches were not restricted by date, language or publication status.
Cochrane Injuries Group's Specialised Register (Dec 2007),
CENTRAL (The Cochrane Library, issue 4, 2007),
MEDLINE (1966 to December, 2007),
EMBASE (1980 to December, 2007),
CINAHL (1982 to May, 2006),
Zetoc (searched 4th December 2007).
Details of the search strategies can be found in Appendix 1.
Searching other resources
In addition, we contacted experts in the field of burns and nutrition and contacted authors of relevant trials to identify any unpublished or ongoing studies.
Data collection and analysis
Selection of studies
Records retrieved by the initial search were scanned by JW, RJ and HC to exclude obviously irrelevant studies, and then two authors (JW and RJ) identified trials that may have met the inclusion criteria. Full-text articles were retrieved and reviewed by two authors (JW and RJ) for the purpose of applying inclusion criteria independently. In all instances, differences of opinion were resolved by discussion among the authors.
Data extraction and management
Data from the studies were extracted independently by two authors (JW and RJ) using standardised forms. The author of one primary study was contacted to provide information when missing or incomplete data was encountered. All differences were resolved by discussion among the review authors.
Assessment of risk of bias in included studies
Study quality was assessed using an adaptation of the method outlined in Schulz 1995. Results from the study quality are presented in a descriptive manner. The following characteristics were assessed:
- Method of randomisation;
- Allocation concealment;
- A) Adequate - measures to conceal allocations such as central randomisation; serially numbered, opaque sealed envelopes; or other description that contained convincing elements of concealment;
- B) Unclear - in which the author either did not report an allocation concealment approach at all, or reported an approach that did not fall into one of the categories in A);
- C) Inadequate - such as alternation methods or use of case record numbers;
- Patient baseline characteristics;
- Blinding (of treatment provider, patient, outcome assessor);
- Intention-to-treat analysis;
- Loss to follow-up.
Assessment of heterogeneity
We used a fixed-effect model where there was no evidence of significant heterogeneity between studies and planned to use a random-effects model when such heterogeneity was likely (DerSimonian 1986). Consideration was to be given to the appropriateness of meta-analysis in the presence of significant clinical or statistical heterogeneity. Statistical heterogeneity was assessed using the I
Assessment of reporting biases
Publication bias was to be tested using funnel plots; however, this was not appropriate given the small number of studies located.
For proportions (dichotomous outcomes), relative risk (RR) was used. Continuous data were converted to the mean difference (MD) using the inverse variance method and an overall MD calculated.
Subgroup analysis and investigation of heterogeneity
Subgroup analysis was planned based on method of feeing i.e. jejunal versus gastric, nasogastric versus naso-jejunal by calculation of RR or MD in each subgroup and examination of the 95% confidence intervals (CI). Non-overlap in intervals were planned to be taken to indicate a statistically significant difference between subgroups; however, no subgroup analysis was appropriate with the data available. An intention-to-treat analysis was to be performed where possible.
Description of studies
A total of 192 references have been identified (to Dec 2007). Independent scrutiny of the titles and abstracts identified 18 potentially relevant articles. Of the 18 articles assessed in full text form, 15 were excluded because they used: feeds other than enteral alone i.e. total parenteral nutrition (Enzi 1990; Taylor 1999; Chen 2006), explored immediate feeding only (McArdle 1984), examined mixed populations such as paediatric or adults populations, or both (Engelhardt 1994; Gottschlich 2002; Hart 2003; McDonald 1991), study designs such as case reports, case series (Andel 2001; Chiarelli 1990; Garrel 1991; Hansbrough 1993; Kaufman 1986; Koller 1994; Noordenbos 2000). The remaining three studies by Peck 2004; Peng 2001 and Wang 1997 formed the basis of the review.
In Peck 2004, 27 patients (19 men) between 18 and 50 years of age admitted within 24 hours burn injury with at least 20% injuries total body surface area (TBSA) were randomised to either early enteral feeding (within 24 hours) or delayed enteral feeding (between 24 hours and seven days). A commercial feeding formula was used with the rate of the infusion increased every six hours as tolerated until the goal rate was achieved. Patients were allowed oral intake as tolerated, stressing complex carbohydrates and protein-containing fluids and foods, minimizing fluids with simple sugars. All patients were treated according to practice guidelines set out by the American Burn Association protocols; fluid resuscitation, topical wound care with silver sulphadiazine cream during the first week, and active physical and occupational therapy. This trial reported on a series of clinical outcome measures such as number of infections, number of antibiotic days, length of stay and mortality and energy balance such as resting energy expenditure and basal energy expenditure.
In Peng 2001, 22 patients (15 men) with burns ranging from 50% to 80% TBSA were randomised to either early enteral feeding (within 24 hours) or delayed enteral feeding (after 48 hours). A commercial feeding formula was given via a nasogastric tube or orally as early as possible in the early enteral group (70-80ml/3h in the first 24 hour post-burn, 0.75 kcal/ml, and 100-150 ml/2h in the second 24 hour post-burn, 0.75-1 kcal/ml). The delayed enteral feeding group was given an orally administered liquid diet after 48 hour post-burn. The trial reported on a series of secondary outcome measures such as plasma endotoxin, tumour necrosis factor - alpha (TNF-a) and tumour necrosis factor - interleukin 8 (TNF-IL8) levels.
In Wang 1997, 21 patients (18 males) between 18 and 51 years of age with burns equal to or more than 30% TBSA were randomly assigned to either early enteral feeding (within 12 hours) and delayed enteral feeding (between 12 and 72 hours). Patients allocated to early enteral feeding received a commercial feeding formula given via the oral route or nasoduodenal tube. The trial reported on a series of metabolic and biochemical markers such as resting energy expenditure (REE) and TNF-a and TNF-IL8 levels.
Risk of bias in included studies
Randomisation procedures were not reported in any of the three studies.
Allocation concealment was unclear in all of the studies.
Patient baseline characteristics
Burn depth was not described in any of the studies, although estimates of burn size were provided by all three studies. Within the studies, patients were generally well matched for sex, age, and size of burn injury.
None of the trials described blinding of investigators or outcome assessors.
None of the studies reported any violation of the allocated treatment.
Loss to follow-up
One study (Peck 2004) reported on losses to follow-up. None of the remaining studies suffered any losses to follow-up.
Effects of interventions
1. All cause mortality at end of follow-up
Only one trial contributed results to this outcome (Peck 2004) involving nine subjects (33% of the total study population), four (28%) were randomised to early enteral nutritional support and five (38%) to late enteral nutritional support. There was no difference in all cause mortality between the two groups (RR 0.74; 95% CI 0.25 to 2.18; P=0.59).
2. Length of stay
Only one trial contributed results to this outcome (Peck 2004). There was no difference in overall length of stay (P=1.00) and number of days in intensive care (WMD 3.00 days; 95% CI -21.55 to 27.55; P=0.81) between those receiving early enteral nutritional support and those receiving late enteral nutritional support.
3. Frequency of infection
Only one trial contributed results to this outcome (Peck 2004). There was no difference in number of infections or number of antibiotic days between those receiving early enteral nutritional support and those receiving late enteral nutritional support (P=1.00).
4. Number of adverse events
Only one trial contributed results to this outcome (Peck 2004). There was no significant difference between those receiving early enteral nutritional support and those receiving late enteral nutritional support with respect to tube feeding tolerance (P=1.0), incidence and duration of diarrhoea, (WMD 4.00 days; 95% CI -12.22 to 4.22; P=0.34) aspiration events (RR 0.19; 95% CI 0.01 to 3.56, P=0.26), or the need for TPN (RR 2.32; 95% CI 0.54 to 9.95; P=0.26).
Only one trial contributed results to this outcome (Peck 2004). Data represented as an 'average change per week' showed no significant difference in body weight between those receiving early enteral nutritional support and those receiving late enteral nutritional support (-1.3kg with early enteral support versus -1.2kg with late enteral support; P=0.9).
2. Nutritional markers
Only one trial contributed results to this outcome (Peck 2004). Data represented as an 'average change per week' showed no significant difference in transthyretin (1.3mg/dL with early enteral support versus 1mg/dlL with late enteral support; P=0.5), transferrin (2.0mg/dL with early enteral support versus 5mg/dlL with late enteral support; P=0.2), and urine urea nitrogen (-383mg/24hrs with early enteral support versus 12mg/24hrs with late enteral support; P=0.2) levels in those receiving early enteral nutritional support and those receiving late enteral nutritional support.
3. Metabolic markers
Resting energy expenditure (REE) was reported in one trial (Wang 1997). The authors noted a decrease, on average of 27%, on post burn day four, eight, and 14 respectively in those receiving early enteral nutritional support than those receiving late enteral nutritional support (P<0.05). A second study by Peck 2004 adjusted REE by standardizing basal energy expenditure using the Harris-Benedict equation and reported on daily energy expenditure (DEE) instead. Result showed no significant trend for DEE in either group, although the estimated DEE in the early group tended to be higher rather than lower than the estimated mean DEE in the late group (1.56 versus 1.42, representing a difference of 0.14; P=0.23). When adjusted for %TBSA, inhalation injury and age, the trend for a higher estimated mean DEE in the early group compared with the late group persisted (mean difference 0.17; P=0.07).
4. Biochemical markers
Two studies (Peng 2001; Wang 1997) examined the effects of early enteral nutrition using biochemical markers such as plasma endotoxin and tumour necrosis factor - alpha and interleukin 8 (TNF-a, TNF-IL8) levels. However, the data from these studies could not be pooled as there were variations in the monitoring and recording of these results. One study (Peng 2001) examined changes in endotoxin and TNF levels and the other study (Wang 1997) examined the content of endotoxin and TNF in plasma.
Wang 1997 showed that serum endotoxin levels in the early nutritional support group were lower on post burn days four, eight and 14 than those who received late enteral nutritional support (P<0.05). Similar reports of endotoxin levels being lower in those receiving early nutritional support group were also noted by Peng 2001 in a series of different time points (i.e. one to six hours, seven to 12 hours, 24 hours, 36 hours, three and five days).
Markers such as TNF-a and TNF-IL8 was seen to be significantly elevated after burn injury in both studies. In Peng 2001, the serum TNF-a levels in both groups were significantly higher than in normal controls (P<0.001). The serum TNF-a levels in those receiving late enteral nutritional support continued to rise after burn injury and reached their highest level at 24 hours post-burn. Following a small decline at day three post-burn, the serum TNF-a levels increased gradually and reached another highest level at day five post-burn. The serum TNF-a level in those receiving early enteral nutritional support increased slower after burn injury (P<0.001). In Wang 1997, TNF-IL8 was lower in those receiving early enteral nutritional support during various time points (P <0.05 to 0.01).
5. Hormonal markers
Only one trial contributed results to this outcome (Wang 1997). Plasma glucagon, cortisol and urinary catecholamine levels were reportedly lower in those receiving early enteral nutritional support (P<0.05 to 0.01).
This systematic review summarises the best available evidence relating to safety and effectiveness of early versus delayed nutritional support in adult patients with burn injury. After an extensive search of the literature, we found three RCTs.
Our results highlighted the need for high level, good quality research into the safety and effectiveness of early versus delayed enteral nutritional support in adults with burn injury. There was wide variation amongst the studies in the timing of the initiation of both early and late enteral nutritional support and the type of clinical, metabolic and hormonal outcome measures used to determine effectiveness. All of the trials were prone to a number of methodological shortcomings including a lack of power to detect useful clinical differences between intervention groups. More importantly, the small sample sizes in all three studies may have precluded any definitive statement on definition, safety or frequency of adverse events.
The evidence regarding the benefit of early enteral nutritional support on standard clinical outcomes such as number of infections, length of hospital stay and mortality, remains inconclusive. Similarly, the question of whether early enteral feeding influenced or decreased hyper-metabolism, remains uncertain. This uncertainty is coupled with a number of previously published studies which showed little or no improvement in the reduction of burn hypermetabolism. For example, it was Eyer 1993 who first suggested that enteral nutrition did not attenuate the hyper-metabolic response to injury and Gottschlich 2002 who reported that although early enteral feeding reduced caloric deficits, it did not show a reduction in morbidity, mortality, hyper-metabolism, hospital stay or endocrine status. More so, Hart 2003 suggested that those who underwent late enteral feeding and burn wound excision had a decreased metabolic rate compared with the early feeding and excision group, but this effect was eliminated once delayed group patients were fed and underwent excision. Although there was a suggestion of fewer infections in the early group, no difference was reported between groups in regards to body mass index, length of stay or mortality. More recently with our included study by Peck 2004, early enteral feeding did not reduce post-burn hyper-metabolism nor did it show any difference between groups in regards to number of infections, antibiotic and ventilator days, overall length of stay and mortality.
Peck 2004 would suggest that the ambiguity surrounding this clinical phenomenon could be due to a lack of understanding surrounding the molecular mechanism of the hypermetabolic response to burn injury. It appears that approximately 60% of the increased metabolic response to burn injury is attributable to an increased protein synthesis, gluconeogenesis, urea production, and substrate cycling (Yu 1999). The remaining 40% may be attributable to Na+-K+- ATPase activity and proton leakage across the mitochondrial membrane. This understanding, coupled with recent data on the beneficial effect of beta-blockade (Herdon 2001) and oxandrolone (Wolf 2003) on reversing skeletal muscle catabolism would therefore suggest that the reversal of post-burn hyper-metabolism would require more than early enteral nutritional support alone (Peck 2004).
Additionally, the frequency of monitoring the metabolic, hormonal and biochemical markers differed between the studies. In some studies monitoring was observed at day one, four, eight and 14 post burn injury, whilst in others monitoring was done on a weekly basis. This could have led to either early detection and possible treatment or possibly an inflated incidence in some cases leading to no treatment or insufficient time for the treatment to take effect.
The ambiguity surrounding the reduction in hyper-metabolism was still over-shadowed by gaps in knowledge surrounding when early nutritional support (i.e. timeliness of initiation of support) must be started to achieve benefit. Classic animal studies (Dominnioni 1984; Mochizuki 1984) revealed impressive findings such as decreases in the hyper-metabolic response to tissue injury, lower levels of circulating stress hormones (i.e. glucagons, cortisol and norepinephrine), increased gastrointestinal blood flow, and the ongoing preservation of intestinal mucosal integrity when animals were fed within two hours of burn injury. Although we hoped our systematic review would help clarify this issue in humans, optimal timing for the initiation of enteral nutrition support still represented an important, unresolved issue. The time to initiate feeding varied from within 24 hours (Peng 2001) and up to seven days post burn injury for late enteral nutritional support (Peck 2004). These variations suggest that early feeding may be difficult to achieve in burn injured patients given the transport and emergency and resuscitation measures required to stabilise a patient with a severe burn.
A number of complications were noted by Peck 2004 although these were not seen to be significant. This was in direct contrast to other studies in the literature that suggested early enteral nutrition support could lead to devastating complications such as bowel necrosis (Gottschlich 2002).
Overall, the evidence would suggest that a degree of vigilance be taken when commencing an early feeding regime, although this is tempered with authors such as Gottschlich 2002 who believe that the provision of enteral nutrition is an important intervention that can be started within a few hours. Peck 2004 purports if this is to be done, it should be done with the knowledge that there appears to be no obvious metabolic or clinical benefit. As a result, current wisdom would then propose that the need for ongoing future research that includes conducting large multi-centre, randomised, double-blind studies, determining precise feeding implementation times, rate of feeding progression, consistent feeding practices (i.e. polymeric versus elemental, fibre enriched, immune-enhancing) coupled with a number of key outcome measures is needed which in turn would allow for pooling of data in a meta-analysis.
Implications for practice
The results of this review suggest that early feeding may be associated with blunting the hypermetabolic response to burn injury, but the trials were small and prone to a number of methodological shortcomings. More RCTs are needed.
Implications for research
This systematic review highlights the need for further high level, good quality research into the use of early versus delayed feeding burned injured patients. Suggestions for future research include precise feeding implementation times, conducting appropriately powered studies, and selecting uniform and objective outcome measures which will allow for pooling of data in a meta-analysis.
Doris Lam for help with translation.
Data and analyses
- Top of page
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Index terms
Appendix 1. Search strategy
Cochrane Injuries Group’s Specialised Register (searched 4
(Burn* or (thermal and injur*)) and (nutrition* or feed*)
MEDLINE 1966 to December 2007
1. exp Burns/
3. (thermal adj3 injur$).ab,ti.
4. 1 or 2 or 3
5. ((alimentary canal or gastrointestinal or enteral or enteric or oral$ or sip or gastric or tube$ or method$) adj3 (nutrition$ or feed$)).ab,ti.
6. exp Enteral Nutrition/
7. exp Feeding Methods/
8. (PEG or percutaneous endoscopic gastrostomy).ab,ti.
9. ((nasogastric or gastrostomy or jejunostomy or gastric or orogastric or nasoenteric or nasojejunal or feed$) adj3 tube$).ab,ti.
10. 5 or 6 or 7 or 8 or 9
11. 4 and 10
12. (randomised or randomized or randomly or random order or random sequence or random allocation or randomly allocated or at random or controlled clinical trial$).tw,hw.
13. clinical trial.pt.
14. 12 or 13
15. exp models, animal/
16. exp Animals/
17. exp Animal Experimentation/
18. exp Disease Models, Animal/
19. exp Animals, Laboratory/
22. 20 not 21
23. 14 not 22
24. 11 and 23
EMBASE 1980 to December 2007
1. exp Burns/
3. (thermal adj3 injur$).ab,ti.
4. 1 or 2 or 3
5. exp Enteric Feeding/
6. exp Nose Feeding/
7. exp Tube Feeding/
8. ((alimentary canal or intestinal$ or intraintestinal$ or gastrointestinal or enteral or enteric or oral$ or sip or gastric or tube$ or method$) adj3 (feed$ or nutrition$)).ab,ti.
9. ((feed$ or nasogastric or orogastric or nasoenteric or nasojejunal or gastrostomy or jejunostomy or gastric) adj3 tube$).ab,ti.
10. (percutaneous endoscopic gastrostomy or PEG).ab,ti.
11. 5 or 6 or 7 or 8 or 9 or 10
12. 4 and 11
13. exp animal model/
14. Animal Experiment/
15. exp ANIMAL/
16. exp Experimental Animal/
17. 13 or 14 or 15 or 16
19. 17 not 18
20. (randomised or randomized or randomly or random order or random sequence or random allocation or randomly allocated or at random or controlled clinical trial$).tw,hw.
21. exp clinical trial/
22. 20 or 21
23. 22 not 19
24. 12 and 23
CENTRAL (The Cochrane Library, issue 4, 2007)
#1 MeSH descriptor Burns explode all trees
#3 thermal near3 injur*
#4 (#1 OR #2 OR #3)
#5 ((alimentary next canal) or gastrointestinal or enteral or enteric or oral* or sip or gastric or tube* or method*) near3 (nutrition* or feed*)
#6 ((percutaneous next endoscopic next gastrostomy) or PEG)
#7 ((nasogastric or gastrostomy or jejunostomy or gastric or orogastric or nasoenteric or nasojejunal or feed*) near3 (tube*))
#8 MeSH descriptor Enteral Nutrition explode all trees
#9 MeSH descriptor Feeding Methods explode all trees
#10 (#5 OR #6 OR #7 OR #8 OR #9)
#11 (#4 AND #10)
Zetoc (searched 4th December 2007)
nutrition* burn* delay*
Last assessed as up-to-date: 3 December 2007.
Protocol first published: Issue 4, 2005
Review first published: Issue 3, 2006
Contributions of authors
Jason Wasiak: conception of review, literature searching, study selection, review development and drafting of written submissions.
Heather Cleland: drafting of protocol, review development and drafting of written submissions.
Rachel Jeffery: literature searching, conception of background.
Declarations of interest
Medical Subject Headings (MeSH)
MeSH check words
* Indicates the major publication for the study