The World Health Organization reports that 180,000 people die of burn injury around the world each year (1). The American Burn Association considers the major cause of death from burn injury in the United States to be uncontrolled infection leading to sepsis and consequent organ failure (2).
Major clinical practice guidelines recommended that enteral nutrition (EN) should be initiated within 12–24 hours of injury for patients who receive a major burn, defined as greater than 20% Total Body Surface Area (TBSA) (3 , 4). These clinical practice guideline recommendations are based on expert consensus and supported by laboratory studies demonstrating that early EN preserves gut immune function and reduces infectious complications (3 , 4). Although large observational studies support the presence of an association between early EN and reduced infectious complications (5), the only peer-reviewed systematic review of randomized controlled trials (RCTs) on this topic was unable to conduct a meta-analysis of any outcomes and failed to document any clinical benefits (6).
Audits of clinical practice consistently show that 20–40% of eligible patients do not have adequate nutrition support commenced within 12–24 hours of a major burn injury (5 , 7 , 8). If systematic reviews of RCTs confirmed patient-centered benefits arising from a reduction in infectious complications, guideline recommendations could be strengthened from grade B to grade A, and more clinicians may be persuaded to address this evidence-practice gap (9 , 10).
The purpose of this systematic review was to identify, appraise, and synthesize the most current evidence from RCTs conducted in patients with major burn injuries to determine whether early EN can alter patient-centered outcomes. Any form of nutrition support “except” early EN was accepted as an appropriate comparator.
MATERIALS AND METHODS
This systematic review and meta-analysis were conducted and reported in compliance with established methodological guidelines (11). Criteria used to define the study intervention (12), search statements, and analytic plan (13) have been previously published.
Study selection, risk of bias appraisal, and data abstraction were undertaken by at least two authors. Disagreements were settled by obtaining an opinion of a third author. Majority decisions prevailed.
Medline (www.PubMed.org), Embase (www.EMBASE.com), and the China National Knowledge Infrastructure (www.cnki.com.cn) were searched using appropriate statements and terms (12 , 13). Complete details are reported in the online-only supplement (Supplemental Digital Content 1, http://links.lww.com/CCM/E34).
Experts were contacted, and reference lists of published reviews and guidelines were hand searched. The close out date was May 1, 2018.
All RCTs comparing early EN to any other intervention published in any language were retrieved in full text and screened for inclusion. Early EN was defined as a “standard” EN formula provided via any feeding tube route within 24 hours of injury or admission to an ICU or burns unit. A “standard” EN formula was considered to be any formula “not” supplemented with additional glutamine, arginine, or other immune enhancing ingredients. The comparison group was defined pragmatically and was accepted to include any form of nutrition support “except” early EN.
RCTs reporting mortality conducted in adult populations with major burn injuries were eligible for inclusion and were reviewed in detail. A major burn was defined as thermal, chemical, or electrical injury to greater than 20% of TBSA (3 , 4).
Risk of Bias
All included trials were appraised on the reporting of three key methodological criteria: 1) the maintenance of allocation concealment; 2) the use of any form of blinding and; 3) the completeness of patient follow-up. Major methodological flaws leading to a recognized high risk of bias were defined a priori as clear failure to maintain allocation concealment (14) and excessive (> 10%) loss to follow-up (15).
The primary outcome of interest was mortality. Gastrointestinal hemorrhage, sepsis, pneumonia, renal failure, and duration of hospital stay were evaluated as secondary outcomes.
Analysis was conducted using a fixed effects model (16) with the odds ratio (OR) metric (17). The OR metric was calculated using the Mantel-Haenszel method unless data was sparse, in which case the Peto method was used (14 , 18). The underlying assumption behind the fixed effects model was assessed with a formal chi-square test of heterogeneity (16) and was quantified using the I 2 metric (19). Important heterogeneity was defined as a p value for the test of heterogeneity (p heterogeneity) less than 0.10 or I 2 greater than 50% (20).
Analysis was conducted using RevMan Version 5.3.5 for Windows (The Cochrane Collaboration, Oxford, United Kingdom, 2014). A two-tailed p value of less than 0.05 was accepted to indicate statistical significance, whereas a two-tailed p value of less than 0.10 was accepted to indicate a trend toward statistical significance.
Heterogeneity and Stratified Analysis. If important heterogeneity was detected, the following a priori identified potential sources of heterogeneity were investigated via stratified analysis: 1) methodological quality; 2) severity of burn injury; 3) intervention timing and dose; 4) cointerventions and comparison intervention received; and 5) measurement and timing of outcomes (21).
Sensitivity Analysis Excluding Trials With High Risk of Bias. To assess the robustness of the primary analysis, a sensitivity analysis was conducted by excluding all studies that were identified as having a high risk of bias.
Literature Search, Study Selection, and Risk of Bias
The primary electronic literature search identified 6,530 abstracts of potentially eligible articles. Contact with academic and industry experts, review of retrieved abstracts and hand searching of reference lists of published guidelines, and systematic reviews resulted in the retrieval of 958 full-text articles for detailed screening. Of the 958 full-text articles screened in detail, 64 RCTs were identified that appeared to address issues potentially related to the primary study question. Seven of these 64 RCTs were deemed eligible for inclusion in the meta-analysis. The study selection flow is reported in Figure 1. The online-only supplement (Supplemental Digital Content 1, http://links.lww.com/CCM/E34) contains additional details about trials that were excluded (eTable 1, Supplemental Digital Content 1, http://links.lww.com/CCM/E34).
The seven included RCTs enrolled a total of 527 participants (22–28). Five RCTs compared early EN with later nutrition support (23–25 , 27 , 28), whereas two RCTs compared early EN with early PN (22 , 26). Additional details of the study populations and study interventions are reported in Table 1.
One RCT explicitly reported the process used to maintain allocation concealment (24), one RCT failed to maintain allocation concealment by alternating patients to treatment arms (22), and the remaining five were unclear with regard to the processes used to assign patients to treatment groups. Six of the seven included studies reported mortality on 100% of randomized patients, with only one of the seven included studies documenting excessive (> 10%) loss to follow-up (24). No studies reported using any form of blinding. With a relatively uniform distribution on both sides of the line of unity, the funnel plot did not reveal any obvious publication bias (eFig. 1, Supplemental Digital Content 1, http://links.lww.com/CCM/E34).
Seven RCTs were included in the primary analysis of mortality. Four clinical trials explicitly reported hospital discharge mortality (23–25 , 28), whereas three were not explicit about the follow-up times. In these three studies, the longest period of follow-up reported for other study outcomes was 7 days (22), 14 days (26), and until at least until 50% of burn injury had healed (27).
Compared with all other types of nutrition support, commencing EN within 24 hours of major burn injury or ICU or burn unit admission resulted in a significant reduction in mortality (OR, 0.36; 95% CI, 0.18–0.72; p = 0.003; I 2 = 0%) (Fig. 2).
Three trials reported gastrointestinal hemorrhage as an outcome (22 , 27 , 28). Only one trial provided an objective definition for diagnosing a gastrointestinal hemorrhage: signs of blood in nasogastric tube aspirate, vomitus, or stool (27).
Patients who were randomized to receive early EN were significantly less likely to develop a gastrointestinal hemorrhage (OR, 0.23; 95% CI, 0.11–0.49; p = 0.0001; I 2 = 0%) (Fig. 3).
Three trials reported sepsis as an outcome (22 , 25 , 28). There was a significant reduction in the onset of sepsis in patients who received early EN (OR, 0.23; 95% CI, 0.11–0.48; p < 0.0001; I 2 = 0%) (Fig. 4).
Three trials reported pneumonia rates (22 , 25 , 28). Patients who received early EN were significantly less likely to develop pneumonia (OR, 0.41; 95% CI, 0.21–0.81; p = 0.01; I 2 = 63%) (eFig. 2, Supplemental Digital Content 1, http://links.lww.com/CCM/E34). However, important heterogeneity was present within this group of trials (p heterogeneity= 0.07; I 2 = 63%).
Three trials reported renal failure as an outcome (22 , 25 , 28). Patients who were randomized to receive early EN were significantly less likely to develop renal failure (OR, 0.27; 95% CI, 0.09–0.82; p = 0.02; I 2 = 32%) (eFig. 3, Supplemental Digital Content 1, http://links.lww.com/CCM/E34).
Duration of Hospital Stay
Three trials reported duration of hospital stay as an outcome (23 , 24 , 28). Receiving early EN resulted in a significant reduction in hospital stay (–15.31 d; 95% CI, –20.43 to –10.20; p ≤ 0.00001; I 2 = 0%) (eFig. 4, Supplemental Digital Content 1, http://links.lww.com/CCM/E34).
Stratified Analysis to Investigate Heterogeneity
Important heterogeneity was present when pneumonia was considered as an outcome (p heterogeneity = 0.07; I 2 = 63%). Due to the limited number of trials (three) in this meta-analysis, an investigation of possible sources of heterogeneity could not be undertaken according to the preplanned criteria.
Sensitivity Analysis Excluding Trials With High Risk of Bias
After removal of the two clinical trials identified as having a high risk of bias (22 , 24), sensitivity analysis confirmed our primary findings: commencing EN within 24 hours of major burn injury or ICU or burn unit admission resulted in a significant reduction in mortality (OR, 0.35; 95% CI, 0.13–0.97; p = 0.04; I 2 = 28%).
Post Hoc Sensitivity Analysis Based on Comparator Group Intervention
Although two RCTs compared early EN with early PN (22 , 26), the remaining five RCTs compared early EN with later oral or enteral intake (23–25 , 27 , 28). Early EN significantly reduced mortality in comparison to early PN (OR, 0.32; 95% CI, 0.12–0.86; p = 0.02; I 2 = 0) and early EN also significantly reduced mortality in comparison to later nutrition support (OR, 0.42; 95% CI, 0.18–0.99; p = 0.047; I 2 = 0).
Our literature search identified seven RCTs conducted in 527 major burn injury patients that compared standard EN, commenced within 24 hours of burn injury, to other forms of nutrition support. The primary analysis demonstrated that early EN significantly reduced mortality. Sensitivity analysis, conducted by excluding trials found to be at high risk of bias, confirmed the presence of a significant reduction in mortality attributable to early EN. Analyses of secondary outcomes demonstrated that early EN significantly reduced gastrointestinal hemorrhage, sepsis, pneumonia, renal failure, and duration of hospital stay.
Maintaining Gut Barrier Function
The improvements in clinical outcomes demonstrated in this meta-analysis are entirely consistent with the physiologic rationale cited to support clinical recommendations for early EN made by major clinical practice guidelines: early EN maintains the gut’s physical barrier immune function and thus reduces the sequelae arising from bacterial translocation (3 , 4).
The gastrointestinal tract is accepted to be one of the most sensitive organs to compromised blood flow (29). Even in healthy individuals, only 1 hour of moderate exercise results in splanchnic hypoperfusion that compromises epithelial cell integrity leading to increased permeability and activation of neutrophils (30). Soon after injury, patients with major burns are known to develop gastric and duodenal mucosal changes consistent with ischemic injury that may progress to ulcerative erosions giving rise to occult blood in stools or even life-threatening bleeds (31). Furthermore, these ischemic intestinal changes compromise the gut’s physical barrier immune function, allowing the translocation of bacteria from the gut to the mesenteric lymph and portal circulation (32). In addition to increasing infectious complications, this overflow of gut bacteria activates WBCs and tissue-specific macrophages (e.g., Kupffer cells, lung resident macrophages) resulting in an inflammatory cascade that initiates consequent sequential organ failures and drives the clinical expression of sepsis (2 , 33–35).
Large observational studies conducted in adult patients with a major burn injury demonstrate that patients who receive early EN are 40% less likely to experience a gastrointestinal hemorrhage (36). Fluid resuscitated animal models show that, independent of cardiac output, early EN results in a significant increase in blood flow to both the small and large intestine after a major burn (37). The preservation of blood flow by early EN maintains the gut’s physical barrier immune function, which significantly reduces quantifiable bacterial translocation (32). Early EN also reduces detectable endotoxin in the blood, blunts the excessive cortisol response to burn injury, reduces tumor necrosis factor alpha, and enhances the host’s ability to kill translocating bacteria (32 , 37 , 38).
Maintenance of the gut’s physical barrier immune function by early EN is a plausible mechanism of action to explain the major clinical effects detected in our meta-analysis: gut integrity is preserved leading to fewer gastrointestinal bleeds, less infectious complications, a reduction in consequent organ failures, and a reduction in the onset of sepsis. The cumulative benefit of these effects improves patient survival and reduces hospital length of stay.
Strengths and Limitations
The results of our primary analysis were robust, with sensitivity analysis excluding two trials identified as having a high risk of bias, confirming the presence of a significant mortality reduction. Although six of seven included RCTs explicitly reported mortality for 100% of enrolled patients, only one included trial specifically reported how allocation concealment was maintained.
Although the analysis of our primary outcome (mortality) was based on meta-analysis of data generated from seven clinical trials recruiting 527 patients with major burn injury, there was little consistency between trials in the reporting of secondary outcomes. For example, only three trials recruiting 359 patients reported the onset of gastrointestinal hemorrhage, three trials with 247 patients reported the onset of sepsis and three trials recruiting 110 patients report hospital stay. Thus, analyses based on these secondary outcomes contain more uncertainty than analysis of the primary outcome, mortality.
Our literature search was not restricted to English language scientific indexing databases. Searching the China National Knowledge Infrastructure indexing database identified three eligible RCTs that were not indexed on English language databases. Furthermore, our search identified an additional 13 RCTs that compared early EN to other forms of nutrition support, but these RCTs could not be included because patient outcomes were not reported (eTable 1, Supplemental Digital Content 1, http://links.lww.com/CCM/E34). If a study focuses primarily on physiology, patient outcomes must be reported to ensure improvements in physiology are aligned with improvements in patient outcomes (40 , 41). Although small studies may lack power to detect effects on patient outcomes, if patient outcomes are reported, power can be improved by including these studies in subsequent meta-analyses (14).
Our analysis did detect the presence of important heterogeneity when pneumonia was considered as an outcome. Due to the low number of clinical trials reporting pneumonia, preplanned stratified analyses could not be undertaken to identify the source of this heterogeneity. None of the included clinical trials reported explicit criteria used to diagnose pneumonia.
Caveats and Recommendations for Future Research
Based on the limitations identified through the conduct of this systematic review and meta-analysis, we feel a number of caveats and recommendations are warranted: 1) Because only one of the seven included RCTs explicitly reported how allocation concealment was maintained, we strongly recommend authors of RCTs conducted in this field should ensure they provide sufficient details of the randomization process such that readers are confident that allocation concealment was maintained. Allocation concealment is the single most important design method that protects an RCT from being classified as high risk of bias (14 , 39); 2) Because our search strategy identified 13 clinical trials that were on-topic but were excluded because they did not report mortality, we strongly recommend that future researchers in this field should ensure mortality is reported for all enrolled patients, with minimum acceptable follow-up until hospital discharge; 3) In addition to documenting mortality, we recommend future trials report key secondary outcomes important to patients with major burn injury (gastrointestinal hemorrhage, sepsis, acute renal failure, quality of life) and also report outcomes and complications specific to the interventions under study (lean body mass, physical function, line infections, diarrhea); 4) The use of explicit diagnostic criteria and a blinded adjudication committee may reduce the presence of heterogeneity between studies when pneumonia, or other subjective outcomes, are considered important (42) and; 5) We strongly recommend that future systematic reviews conducted on this topic should search major non-English language indexing databases to maximize the identification of informative publications.
Globally, 180,000 people die of burn injury each year (1). Current major clinical practice guidelines recommend commencing early EN if patients have a major burn injury (3 , 4). Our meta-analysis documents a significant and robust reduction in mortality attributable to standard EN commenced within 24 hours of major burn injury. These findings may result in current clinical practice recommendations being strengthened from grade B to grade A (9), which could translate to a meaningful improvement in practice that will have an impact on the lives of many thousands of people.
2. Greenhalgh DG, Saffle JR, Holmes JH, et al. American Burn Association consensus conference to define sepsis and infection in burns. J Burn Care Res 2007; 28:776–790
3. Rousseau AF, Losser MR, Ichai C, et al. ESPEN endorsed recommendations: Nutritional therapy in major burns. Clin Nutr 2013; 32:497–502
4. ISBI Practice Guidelines Committee; Steering Subcommittee; Advisory Subcommittee: ISBI Practice Guidelines for burn care. Burns 2016; 42:953–1021
5. Mosier MJ, Pham TN, Klein MB, et al. Early enteral nutrition
in burns: Compliance with guidelines and associated outcomes in a multicenter study. J Burn Care Res 2011; 32:104–109
6. Wasiak J, Cleland H, Jeffery R. Early versus delayed enteral nutrition support for burn injuries. Cochrane Database Syst Rev 2006; 19:CD005489
7. Kurmis R, Heath K, Ooi S, et al. A prospective multi-center audit of nutrition support parameters following burn injury. J Burn Care Res 2015; 36:471–477
8. Miquet Romero LM, Rodríguez Garcell R, Barreto Penié J, et al. Current status of nutritional care provision to burnt patients: Processes audit of a burnt patients department from a tertiary hospital. Nutr Hosp 2008; 23:354–365
9. Bischoff SC, Singer P, Koller M, et al. Standard operating procedures for ESPEN guidelines and consensus papers. Clin Nutr 2015; 34:1043–1051
10. Neumann I, Alonso-Coello P, Vandvik PO, et al. Do clinicians want recommendations? A multicenter study comparing evidence summaries with and without GRADE recommendations. J Clin Epidemiol 2018; 99:33–40
11. Moher D, Liberati A, Tetzlaff J, et al; PRISMA Group: Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Med 2009; 6:e1000097
12. Doig GS, Heighes PT, Simpson F, et al. Early enteral nutrition
, provided within 24 h of injury or intensive care unit admission, significantly reduces mortality in critically ill
patients: A meta-analysis
of randomised controlled trials. Intensive Care Med 2009; 35:2018–2027
13. Doig GS, Heighes PT, Simpson F, et al. Early enteral nutrition
reduces mortality in trauma patients requiring intensive care: A meta-analysis
of randomised controlled trials. Injury 2011; 42:50–56
14. Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. 2011The Cochrane Collaboration, . Available at: http://handbook.cochrane.org
15. Graf J, Doig GS, Cook DJ, et al. Randomized, controlled clinical trials in sepsis: Has methodological quality improved over time? Crit Care Med 2002; 30:461–472
16. Villar J, Mackey ME, Carroli G, et al. Meta-analyses in systematic reviews of randomized controlled trials in perinatal medicine: Comparison of fixed and random effects models. Stat Med 2001; 20:3635–3647
17. Deeks JJ. Issues in the selection of a summary statistic for meta-analysis
of clinical trials with binary outcomes. Stat Med 2002; 21:1575–1600
18. Bradburn MJ, Deeks JJ, Berlin JA, et al. Much ado about nothing: A comparison of the performance of meta-analytical methods with rare events. Stat Med 2007; 26:53–77
19. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis
. Stat Med 2002; 21:1539–1558
20. Hatala R, Keitz S, Wyer P, et al; Evidence-Based Medicine Teaching Tips Working Group: Tips for learners of evidence-based medicine: 4. Assessing heterogeneity of primary studies in systematic reviews and whether to combine their results. CMAJ 2005; 172:661–665
21. Glasziou PP, Sanders SL. Investigating causes of heterogeneity in systematic reviews. Stat Med 2002; 21:1503–1511
22. Lam NN, Tien NG, Khoa CM. Early enteral feeding for burned patients–an effective method which should be encouraged in developing countries. Burns 2008; 34:192–196
23. Chiarelli A, Enzi G, Casadei A, et al. Very early nutrition supplementation in burned patients. Am J Clin Nutr 1990; 51:1035–1039
24. Peck MD, Kessler M, Cairns BA, et al. Early enteral nutrition
does not decrease hypermetabolism associated with burn injury. J Trauma 2004; 57:1143–1148–1149; discussion 1148–
25. Vicic VK, Radman M, Kovacic V. Early initiation of enteral nutrition improves outcomes in burn disease. Asia Pac J Clin Nutr 2013; 22:543–547
26. Chen ZY, Wang SL, Yu B, et al. Effect of early intestinal nutrition on intestinal permeability in burn patients. Practical Journal of Medicine 2007; 23:833–835
27. He WH, Wang T. Effect of cimetidine and early enteral nutrition
on preventing burn stress ulcer. Sichuan Medical Journal 2001; 22:942–943
28. Wu ZT. Effect of enteral nutrition on serum protein expression and the role of body tissue repari in severely burned patients. China Medicine and Pharmacy 2013; 3:7–12
29. Garcia dL, Ortiz LC, Sanchez SM. Guidelines for specialized nutritional and metabolic support in the critically-ill patient: Update. Consensus SEMICYUC-SENPE: Critically-ill burnt patient. Nutr Hosp 2011; 26(Suppl 2):59–62
30. van Wijck K, Lenaerts K, van Loon LJ, et al. Exercise-induced splanchnic hypoperfusion results in gut dysfunction in healthy men. PLoS One 2011; 6:e22366
31. Czaja AJ, McAlhany JC, Pruitt BA Jr. Acute gastroduodenal disease after thermal injury. An endoscopic evaluation of incidence and natural history. N Engl J Med 1974; 291:925–929
32. Gianotti L, Alexander JW, Nelson JL, et al. Role of early enteral feeding and acute starvation on postburn bacterial translocation and host defense: Prospective, randomized trials. Crit Care Med 1994; 22:265–272
33. McClave SA, Heyland DK. The physiologic response and associated clinical benefits from provision of early enteral nutrition
. Nutr Clin Pract 2009; 24:305–315
34. Moore FA, Moore EE. The evolving rationale for early enteral nutrition
based on paradigms of multiple organ failure: A personal journey. Nutr Clin Pract 2009; 24:297–304
35. Magnotti LJ, Xu DZ, Lu Q, et al. Gut-derived mesenteric lymph: A link between burn and lung injury. Arch Surg 1999; 134:1333–1340; discussion 1340–1341
36. Raff T, Germann G, Hartmann B. The value of early enteral nutrition
in the prophylaxis of stress ulceration in the severely burned patient. Burns 1997; 23:313–318
37. Inoue S, Lukes S, Alexander JW, et al. Increased gut blood flow with early enteral feeding in burned guinea pigs. J Burn Care Rehabil 1989; 10:300–308
38. Peng YZ, Yuan ZQ, Xiao GX. Effects of early enteral feeding on the prevention of enterogenic infection in severely burned patients. Burns 2001; 27:145–149
39. Schulz KF, Grimes DA. Allocation concealment in randomised trials: Defending against deciphering. Lancet 2002; 359:614–618
40. Shaughnessy AF, Slawson DC. What happened to the valid POEMs? A survey of review articles on the treatment of type 2 diabetes. BMJ 2003; 327:266
41. Fleming TR, DeMets DL. Surrogate end points in clinical trials: Are we being misled? Ann Intern Med 1996; 125:605–613
42. Tian F, Heighes PT, Allingstrup MJ, et al. Early enteral nutrition
provided within 24 hours of ICU admission: A meta-analysis
of randomized controlled trials. Crit Care Med 2018; 46:1049–1056