Gut dysfunction (GDF) is a common problem in critically ill patients. It is the leading cause of multiple organ dysfunction syndrome/failure (MODS/MOF) and a significant cause of mortality and morbidity in critically ill patients [1]. In addition to this, the treatment of acute and critical illness can exacerbate GDF. Commonly used ICU interventions such as intravenous fluid resuscitation, early aggressive enteral feeding and vasopressor therapy are key factors leading to a secondary gut injury. In critical illness, intravenous fluid is the mainstay of early management for hemo-dynamic instability. It is vital to resuscitate a patient before commencing vasopressor therapy, particularly to delay the onset of an ischemic insult commonly occurring in hemodynamically unstable patients [2]. On the flip side, over-resuscitation can lead to bowel oedema leading to an ileus, while under-resuscitation with persistent splanchnic and peripheral vasoconstriction can trigger intestinal mucosal ischemia [3]. Although, enteral nutrition is the preferred approach to meet nutritional and modest fluid requirements in these patients, the delivery of early but aggressive enteral nutrition (EN) in hemodynamically unstable patients can precipitate the development of severe GDF, potentially leading to non-occlusive mesenteric ischemia which increases the chance of sepsis, multi-organ failure and mortality [4]. Intravenous fluid and enteral nutrition are two sides of the same coin and play a crucial role in determining the outcome of GDF if used wisely. However, very few studies have evaluated the role of these two modalities, thus making it difficult to understand their relationship with relevance to the severity of GDF. The aim was to review the evidence of the impact of intravenous fluid resuscitation and enteral nutrition individually on determinants of gut function and the implications in clinical practice.
Electronic databases (MEDLINE and EMBASE) were searched using keywords on ‘gastrointestinal dysfunction in adult intensive care unit (ICU) /surgical patients on enteral feeding and intravenous fluids. The databases screened for all publications from the earliest available until 16th October 2018 (Appendix A).
Randomised controlled trials were searched by applying the keywords. Any additional studies on the impact of ‘intravenous fluid’ and ‘enteral feeding’ were included in the screening for the systematic review and meta-analysis. The search identification, screening and selection were conducted by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow chart (Fig1) [5]. The study selection criteria were as follows.
The inclusion criteria were:
The studies were excluded if they were:
non-ICU or non-surgical patients
paediatric population
animal studies
published in non-English languages
conducted on healthy volunteers
non-randomized trials (intravenous fluid therapy and enteral feeding)
not relevant to either of the interventions planned to study pattern of feeding (bolus vs continuous), comparative feed compositions (standard vs immune-enriched), related routes of feeding (nasogastric vs nasojejunal or jejunal) and studies addressing medications (e.g. prokinetic therapy).
Study characteristics included baseline demographic data such as author, publication year, study setting (ICU or surgical ward), admission diagnosis, study population, the total number of patients, fluid or enteral feeding interventions applied to experimental and control groups. The effect of fluid therapy and enteral feeding on GDF was analysed by separating the severity of GDF outcomes: 1)
The methodological quality of included randomised controlled trials was assessed according to the Cochrane recommendations (The Nordic Cochrane Centre, The Cochrane Collaboration, 2008) [7]. These included systematic differences between groups (selection bias and performance bias), blinding of study participants and assessors, sequence allocation and concealment of allocated groups, the validity of findings and data withdrawal, incomplete outcome data (attrition and detection bias), and differences between data reporting or unreported data. The risk of bias assessment was presented according to the Cochrane collaboration recommendations. The overall quality of the study was graded as
All data were presented as the number of episodes of GDF in patients. Data analysis and interpretation were performed using Revman 5.3 (Revman, Version 5.3 for Windows; Copenhagen, Denmark: the Nordic Cochrane Centre, The Cochrane Collaboration, 2008) [7]. The nature of the analysis was not suitable for a pooled data analysis. Within each class of interventions (intravenous fluid and enteral feeding), a meta-analysis of GDF events was performed. Quantitative data meta-analysis was performed with at least two studies reporting on GDF as the primary or secondary outcome. Studies that did not have GDF as a primary or secondary outcome were excluded from the meta-analyses (Fig 1).
Heterogeneity was assessed by using I2 and classified as < 25% - low ; 25 – 50% - moderate and > 75% as high heterogeneity (heterogeneity and subgroup analysis in Cochrane consumers and communication group reviews) [8]. Regardless of the presence or absence of heterogeneity, a random-effects model was used to provide the most conservative estimate. Pooled effects for classes of interventions were calculated as weighted mean difference (MD) with 95% confidence interval (CI). P-value < 0.05 was considered statistically significant for all analyses. Ethical approval was not necessary for a review of published trials.
A total of 103 studies including intravenous fluids (n = 46) and enteral feeding (n = 57) were eligible for inclusion in the systematic review, of which 43 (n = 22 intravenous fluid; n = 21 enteral feeding) studies were included in the final meta-analyses.
In studies on
Study Characteristics of 'good' quality studies on the impact of intravenous fluid therapy on gut dysfunction included in the systematic review
Author | Year | Study Population | Study | Study | Study | Admission diagnosis | Experimental | Intravenous | Control | Intravenous | Dindo-Clavien |
---|---|---|---|---|---|---|---|---|---|---|---|
Brandstrup9 | 2003 | elective colorectal resection | surgery | RCT | 141 | postsurgical | 69 | restricted | 72 | standard | |
Holte10 | 2007 | elective surgery | surgery | RCT | 32 | elective colorectal surgery | 16 | restricted | 16 | liberal | |
Holte11 | 2007 | post-surgery | surgery | RCT | 48 | knee arthroplasty | 24 | restricted | 24 | liberal | |
Gonsalez-Fajardo12 | 2009 | post-surgery | surgery | RCT | 40 | vascular surgery transperitoneal aorto-iliac | 20 | restricted | 20 | standard | |
Yates13 | 2013 | elective surgery | surgery | RCT | 206 | elective colorectal surgery | 104 | starch | 98 | crystalloid | |
Ghodraty14 | 2017 | post-surgery | surgery | RCT | 91 | abdominal surgery | 46 | HES | 45 | ringers lactate | |
Gómez-Izquierdo15 | 2017 | post-surgery | surgery | RCT | 128 | colorectal surgery | 4 | GDFT | 64 | control |
Abbreviations: HES- hydroxyethyl starch; GDFT-goal-directed fluid therapy; RCT-randomised controlled trial. * Appendix C
Study Characteristics of 'fair ' quality studies on the impact of intravenous fluid therapy on gut dysfunction.
Author | Year | Study Population | Study | Study | Study | Admission | Experimental | Intravenous | Control | Intravenous | Dindo-Clavien |
---|---|---|---|---|---|---|---|---|---|---|---|
Gan16 | 2002 | major elective general, urologic, or gynaecologic surgery | surgery | RCT | 100 | postsurgical | 50 | GDFT | 50 | Standard | |
Moretti1 7 | 2003 | Major elective cardiac surgery | surgery | RCT | 90 | postsurgical | 30 - HetaStarch normal saline; 30 Heta Starch Balanced | ||||
Nisanevich18 | 2005 | elective intraabdominal surgery | surgery | RCT | 157 | postsurgical | 77 | Restrictive | 75 | Liberal | |
Kabon19 | 2005 | open colonic resection | surgery | RCT | 253 | ICU surgical | 124 | Small volume | 129 | Large Volume | |
Lopes20 | 2007 | High-risk surgery | surgery | RCT | 33 | ICU surgical | 17 | GDFT | 16 | Control | |
Vermuelen21 | 2009 | elective major abdominal surgical procedures | surgical | RCT | 62 | surgical | 30 | Restricted | 32 | Standard | |
Mayer22 | 2010 | major abdominal surgery | surgery | RCT | 60 | ICU surgical | 30 | GDFT | 30 | Standard | |
SAFE 23 | 2011 | ICU | ICU | RCT | 1218 | ICU | 603 | Colloid | 615 | Crystalloid | |
Guidet24 | 2012 | severe sepsis | ICU | RCT | 196 | ICU | 100 | Colloid | 96 | Crystalloid | |
Perner25 | 2012 | severe sepsis | ICU | RCT | 798 | ICU | 398 | Colloid | 400 | Crystalloid | |
Reddy26 | 2016 | critically ill | ICU | RCT | 69 | critically ill | 35 | plasmalyte | 34 | saline |
Abbreviations: GDFT-goal-directed fluid therapy; ICU -intensive care unit; S-ICU -surgical ICU; RCT-randomised controlled trial; * Appendix C
Study Characteristics of 'poor ' quality studies on the impact of intravenous fluid therapy on gut dysfunction.
Author | Year | Study Population | Study | Study | Study | Admission diagnosis | Experimental | Intravenous | Control | Intravenous fluid | Dindo-Clavien | |||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Prein27 | 1990 | post-surgery | surgery | RCT | 18 | modified Whipple's | 6- ringers' lactate; 6– starch; 6-albumin | I | II | III | IV | |||
Salim28 | 1991 | elective surgery | surgery | RCT | 130 | Hartmann's procedure +/- cholecystectomy | 71 | early oral | 59 | conventional intravenous | ||||
Yogendran29 | 1995 | elective surgery | surgical | RCT | 200 | surgical | 100 | Low-infusion | 100 | High infusion | ||||
Wilkes30 | 2001 | elective, open surgical | surgical | RCT | 47 | surgical | 23 | Balanced | 24 | Saline | ||||
Lobo 31 | 2002 | post-surgery | surgery | RCT | 20 | colorectal surgery | 10 | restricted | 10 | liberal | ||||
Conway32 | 2002 | major bowel surgery | surgical | RCT | 57 | surgical | 28 | GDFT | 39 | Standard | ||||
Venn33 | 2002 | hip fracture surgery | surgical | RCT | 90 | surgical | 29CON- VF ; CVP guided FT- 3 1 ; Doppler-guided FT- 30 | |||||||
SAFE34 | 2004 | ICU | ICU | RCT | 6997 | ICU | 3497 | Colloid | 3500 | Crystalloid | ||||
Parker35 | 2004 | hip fracture surgery | surgical | RCT | 396 | surgical | 198 | Colloid | 198 | Crystalloid | ||||
Noblett36 | 2005 | elective colorectal resection | surgical | RCT | 108 | surgical | 54 | GDFT | 54 | Standard | ||||
Wakeling37 | 2005 | large bowel surgery | surgical | RCT | 128 | surgical | 64 | GDFT | 64 | Standard | ||||
Mackay38 | 2006 | elective colorectal surgery | surgical | RCT | 80 | surgical | 41 | Restricted | 39 | Standard | ||||
En-quiang38 | 2009 | critically ill | S-ICU | RCT | 76 | severe acute pancreatitis | 30 | controlled fluid expansion | 30 | rapid fluid expansion | ||||
Senagore40 | 2009 | laparoscopic | surgical | RCT | 64 | surgical | 21 GDFT/LR; 21 GDFT/HS;2 22 standard | |||||||
Futier41 | 2010 | major abdominal surgery | surgery | RCT | 70 | postsurgical | 36 | Restricted-GDFT | 34 | Conservative GDFT | ||||
Benes42 | 2010 | elective intraabdominal | surgery | RCT | 120 | ICU surgical | 60 | GDFT | 60 | Standard | ||||
Pillai43 | 2011 | post-surgery | surgery | RCT | 66 | radical cystectomy | 34 | intervention | 32 | control | ||||
Du44 | 2011 | critically ill | ICU | RCT | 41 | severe acute pancreatitis | 20 | starch | 21 | ringers' lactate | ||||
James45 | 2011 | Blunt and penetrating | surgical | RCT | 109 | surgical | Penetrating trauma- HES 36 ; SAL 31 Blunt trauma- HES20 ; SAL 22 | |||||||
Challand46 | 2012 | major elective colorectal | surgical | RCT | 179 | surgical | 90 | GDFT | 89 | Standard | ||||
Myberg47 | 2012 | ICU | ICU | RCT | 7000 | ICU | 3500 | Colloid | 3500 | Crystalloid | ||||
Srinivasa48 | 2012 | elective colectomy | surgical | RCT | 85 | surgical | 37 | GDFT Restricted | 37 | Restricted | ||||
Zheng49 | 2013 | post-surgery | surgery | RCT | 60 | gastrointestinal surgery | 30 | GDFT | 30 | control | ||||
Scheeren50 | 2013 | High-risk surgery | ICU | RCT | 52 | ICU | 26 | GDFT | 26 | Control | ||||
Pestana51 | 2014 | post-surgery | S-ICU | RCT | 142 | abdominal surgery | 70 | GDFT | 72 | control | ||||
Pearse52 | 2014 | Major Gastrointestinal | surgery | RCT | 734 | surgical | 368 | GDFT | 366 | Standard | ||||
Peng53 | 2014 | elective surgery | surgery | RCT | 80 | orthopaedic surgery | 40 | GDFT | 40 | standard | ||||
Reisinger54 | 2017 | elective colorectal resection | surgery | RCT | 58 | postsurgical | 27 | GDFT | 31 | Standard | ||||
Abbreviations: : HES- hydroxyethyl starch ; HS- hetastach; SAL- saline; LR- lactate ringers; GDFT-goal-directed fluid therapy; ICU - intensive care unit ; S-ICU -surgical ICU; CON-IVF- conventional intravenous fluid therapy; CVP- central venous pressure; FT-fluid therapy RCT-randomised controlled trial; * Appendix C
In studies on
Study Characteristics of studies on the impact of enteral feeding on gut dysfunction included in the systematic review
Author | Year | Study Population | Study Setting | Study type | Study patients | Admission diagnosis | Experimental | Control | Dindo-Clavien Classification# | Quality Grading* | |||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
I | II | III | IV | ||||||||||
Hoover55 | 1980 | surgical | surgical | RCT | 48 | surgical | 26 EF | 22IVF | Poor | ||||
Adams56 | 1986 | ICU surgical | ICU | RCT | 46 | multiple trauma | 23 (EN) | 23(PN) | Poor | ||||
Moore57 | 1986 | major abdo trauma | surgical | RCT | 59 | surgical | 29 (EN) | 30 (PN) | Poor | ||||
Bower58 | 1986 | surgical | surgery | RCT | 20 | GI/pancreato-biliary surgery | 10 (EN-JeJ) | 10 (PN) | Poor | ||||
Hamoui59 | 1989 | surgical | surgical | RCT | 19 | major GI surgery | 11EN | 8PN | Poor | ||||
Von Meyenfeldt60 | 1992 | surgical | surgery | RCT | 101 | GI/colon cancer | 50 (EN) | 51(PN) | Poor | ||||
Montecalvo61 | 1992 | surgical | surgical | RCT | 38 | surgical | 19 NG | 19 NJ | Poor | ||||
Dunham62 | 1994 | critically ill | ICU | RCT | 37 | trauma | 12 (EN)+ 15 (PN) + 10(EN+PN) | Poor | |||||
Borzotta63 | 1994 | trauma | surgical trauma | RCT | 48 | trauma | 27 (EN) | 21(PN) | Poor | ||||
Daly64 | 1995 | surgical | surgical | RCT | 60 | surgical | 18ENSD; 12SD-IP; 19 ENSD-IP-OP; 11 EN-IP | Poor | |||||
Carr65 | 1996 | post-surgical | surgery | RCT | 28 | intestinal resection | 14(EEN) | 14(CEN) | Poor | ||||
Beier-Holgersen66 | 1996 | post-surgical | surgery | RCT | 60 | major abdominal surgery | 30(EEN) | 30 (placebo) | Poor | ||||
Baigrie67 | 1996 | post-surgical | surgery | RCT | 97 | oesophagectomy/gastrectomy | 50 (EN) | 47(PN) | Poor | ||||
VanBerge68 | 1997 | post-surgical | surgery | RCT | 57 | pancreatoduodenectomy | 30 (CON) | 27(CYC) | Poor | ||||
Kalfarentzos69 | 1997 | critically ill | ICU | RCT | 38 | Severe acute pancreatitis | 18(EN) | 20 (PN) | |||||
Heslin70 | 1997 | surgical | surgery | RCT | 195 | upper GI malignancy | 97 (EN) | 98(IVF) | Poor | ||||
Reynolds71 | 1997 | major upper GI surgery | surgical | RCT | 67 | surgical | 33 (EN) | 34(PN) | Poor | ||||
Stewart72 | 1998 | elective surgical | surgery | RCT | 80 | colorectal resections | 40 (EOF) | 40 (COF) | Poor | ||||
Windsor73 | 1998 | surgical | surgical | RCT | 34 | acute pancreatitis | 16 EN | 18PN | Poor | ||||
Singh74 | 1998 | surgical | surgical | RCT | 43 | surgical | 22JEJ | 21IVF | Poor | ||||
Braga75 | 1998 | surgical | surgical | RCT | 166 | surgical | 55 STD-EN; 55 -STD-EN enriched; 56 TPN | Poor | |||||
Taylor76 | 1999 | critically ill | ICU | RCT | 82 | head injury | 41TRO | 41 EN | Fair | ||||
Pupelis77 | 2000 | critically ill | S-ICU | RCT | 60 | severe pancreatitis/peritonitis | 30 (JEN) | 30 (Control) | Poor | ||||
Minard78 | 2000 | critically ill | ICU | RCT | 27 | head injury/trauma | 12(EEN) | 15(DEN) | Poor | ||||
Powell79 | 2000 | critically ill | ICU | RCT | 27 | severe acute pancreatitis | 13 (EN) | 14(NBM) | Poor | ||||
Kearns80 | 2000 | critically ill | ICU | RCT | 44 | critically ill | 23 G | 21 SI | Poor | ||||
Bozzetti81 | 2001 | elective surgery | surgery | RCT | 317 | GIcancer | 159(EN) | 158(PN) | Poor | ||||
Braga82 | 2001 | surgical | surgery | RCT | 257 | GIcancer | 126(EEN) | 131(PN) | Poor | ||||
Montejo83 | 2002 | critically ill | ICU | RCT | 101 | critically ill | 50 (JEN) | 51(GEN) | Poor | ||||
Davies84 | 2002 | critically ill | ICU | RCT | 73 | critically ill | 34 (NJ) | 39 (NG) | Poor | ||||
Bertolini85 | 2003 | critically ill | ICU | RCT | 39 | Sepsis | 18(EN) | 17 (PN) | Poor | ||||
Kompan86 | 2004 | critically ill | ICU | RCT | 52 | multiple trauma | 27(EEN) | 21(DEN) | Poor | ||||
Malhotra87 | 2004 | post-surgical | surgery | RCT | 164 | perforated gut and peritonitis | 83 (EN) | 81(NBM) | Poor | ||||
Kumar88 | 2006 | Surgical | surgical | RCT | 31 | surgical | 15 NG | 16 NJ | Poor | ||||
Nguyen89 | 2007 | critically ill | ICU | RCT | 31 | critically ill | 23 (NJ) | 28 (NJ) | Poor | ||||
Han-Guerts90 | 2007 | post-surgical | surgery | RCT | 150 | oesophagectomy | 71 (ND) | 79 (JEJ) | Poor | ||||
Descahy91 | 2008 | critically ill | ICU | RCT | 100 | ICU | 50EEN | 50CEN | Poor | ||||
Tien92 | 2009 | critically ill | ICU | RCT | 200 | ICU | 98TRO | 102 EN | Poor | ||||
Barlow93 | 2011 | Surgical | surgery | RCT | 121 | upper GI malignancy | 64 (EN) | 57(NBM+IVF) | Poor | ||||
Altintas94 | 2011 | critically ill | ICU | RCT | 71 | ICU | 30 (EN) | 41 (PN) | Poor | ||||
Rice95 | 2011 | Surgical | surgical | RCT | 247 | surgical | EN 123 | 124 IVF | Poor | ||||
Davies96 | 2013 | critically ill | ICU | RCT | 181 | ICU | 91 NJ | 89 NG | Poor | ||||
Zhu97 | 2013 | post-surgical | surgery | RCT | 68 | pancreaticoduodenectomy | 34(JT) | 34(NJT) | Poor | ||||
Sun98 | 2013 | critically ill | S-ICU | RCT | 60 | severe acute pancreatitis | 30(EEN) | 30(DEN) | Poor | ||||
Kadamani99 | 2014 | critically ill | ICU | RCT# | 15 | critically ill | 15 (CON) | 15 (BOL) | Poor | ||||
Boelens100 | 2014 | elective surgical | surgery | RCT | 123 | rectal surgery | 61(EEN) | 62(EPN) | Poor | ||||
Harvey101 | 2014 | critically ill | ICU | RCT | 2388 | critically ill | 1197(EN) | 1191(PN) | Poor | ||||
Ma102 | 2015 | acute surgical | surgery | RCT | 35 | acute pancreatitis | 17 (NTF) | 18(NPO) | Poor | ||||
Bing Li103 | 2015 | post-surgical | surgery | RCT | 400 | gastrectomy | 200(EEN) | 200 (PN) | Poor | ||||
Taylor104 | 2016 | critically ill | ICU | RCT | 50 | critically ill | 25 (NJ) | 25(NG + ProK) | Poor | ||||
Ozen105 | 2016 | critically ill | ICU | RCT | 51 | critically ill | 26(no-GRV's) | 25(GRV's) | Poor | ||||
Van Barneveld106 | 2016 | elective surgical | surgery | RCT | 123 | rectal ca malignancy | 61 (EEN) | 62(EPN) | Good | ||||
Malik107 | 2016 | critically ill | ICU | RCT | 60 | critically ill | 30 (EF) | 30 (placebo) | Poor | ||||
Fan108 | 2016 | critically ill | ICU | RCT | 80 | Severe TBI | 40 (EN) | 40 (PN) | Poor | ||||
Stimac109 | 2016 | acute pancreatitis | pancreatitis | RCT | 214 | acute pancreatitis | 107 EN | 107 IVF | Poor | ||||
Hongyin110 | 2017 | acute surgical | surgery | RCT | 161 | acute pancreatitis | 83 (APD)/61 EN) | 78(non-APD)/68(EN) | Poor | ||||
Reigner111 | 2018 | critically ill | ICU | RCT | 2410 | shock | 1202(EN) | 1208(PN) | Fair |
Impact of intravenous fluid therapy on variables of gut dysfunction
Symptoms of GDF § | Interventional | Control | Odds Ratio [95% CI]* | P Trend | I2 (%)# |
---|---|---|---|---|---|
Nausea | 88/ 274 | 90/278 | 0.98 (0.67, 1.44) | 0.92 | 0 |
Vomiting | 62/462 | 94/447 | 0.51 (0.28, 0.94) | 45 | |
Ileus | 66/832 | 80/828 | 0.83 (0.52, 1.32) | 0.42 | 23 |
GI bleed | 15/592 | 10/587 | 1.48 (0.66, 3.35) | 0.34 | 0 |
Anastomotic leak | 44/833 | 43/867 | 1.03 (0.54, 1.96) | 0.93 | 31 |
Perforation | 7/238 | 6/234 | 1.05 (0.36, 3.09) | 0.92 | 0 |
Intestinal obstruction | 5/451 | 11/445 | 0.53 (0.20, 1.45) | 0.22 | 0 |
a: restricted, goal-directed, low-infusions or a controlled-expansion fluid therapy given as crystalloid fluid (normal saline or plasmalyte) or colloid fluid (hydroxyethyl starch)
b: standard, liberal, conventional, high-infusions or rapid-expansion fluid regimes given as crystalloid fluids (ringers lactate, plasmalyte and saline).
*CI - Confidence interval used; Significant P values (<0.05) are shown in bold; #I2 - heterogeneity between studies expressed as percentages; § GDF - gut dysfunction
Impact of enteral feeding on variables of gut dysfunction as classified by feeding categories
Symptoms of GDF§ | Intervention Enteral | Control Parenteral | Odds [95% Ratio CI]* | P Trend | I2 [%]# |
---|---|---|---|---|---|
Vomiting | 605/2388 | 350/2598 | 2.02 (1.74, 2.35) | 0 | |
Diarrhoea | 190/1508 | 421/1515 | 1.75 (0.39, 7.86) | 0.46 | 92 |
Abdominal distension | 123/1386 | 90/1390 | 1.51 (0.93, 2.45) | 0.10 | 28 |
Ileus | 52/347 | 65/347 | 0.97 (0.34, 2.76) | 0.96 | 58 |
Anastomotic leak | 28/540 | 54/545 | 0.54 (0.31, 0.95) | 14 | |
Intestinal ischaemia | 33/2493 | 16/2495 | 1.87 (0.72, 4.87) | 0.20 | 42 |
Peritonitis | 5/265 | 18/268 | 0.31 (0.11, 0.87) | 0 | |
Early | Delayed | ||||
Vomiting | 3/56 | 19/54 | 0.11 (0.03, 0.41) | 0 | |
Diarrhoea | 27/39 | 23/40 | 2.45 (0.26, 22.75) | 0.43 | 69 |
Abdominal Distension | 12/66 | 21/69 | 0.51 (0.22, 1.91) | 0.12 | 0 |
Enteral | NBM | ||||
Vomiting | 21/220 | 22/219 | 0.72 (0.18, 2.90) | 0.65 | 0 |
Abdominal Distension | 66/242 | 48/240 | 1.40 (0.75, 2.64) | 0.29 | 33 |
GI bleed | 2/133 | 2/133 | 0.99 (0.17, 5.86) | 0.99 | 0 |
Anastomotic leak | 12/244 | 24/236 | 0.46 (0.22, 0.95) | 0 |
*CI - Confidence interval used; Significant P values (<0.05) are shown in bold; #I2 - heterogeneity between studies expressed as percentages; § GDF - gut dysfunction
The quality of studies was graded based on the Cochrane Quality assessment tool for randomised controlled trials for intravenous fluid (Tables 1-3) and enteral feeding (Table 4) studies (Appendix C and D). All studies met the criteria for randomisation and allocation concealment, but a wide variability existed between studies for other domains (blinding of participants and personnel, blinding of outcome assessment and assessor, incomplete outcome data and selective reporting). In the intravenous fluid group, quality assessment for 7 studies [9, 10, 11, 12, 13, 14, 15] (15%) scored
Twenty-two randomised controlled trials [9,10,13, 14, 15, 16,18, 19, 20, 21, 22,26,28,31,41, 42, 43,49,51, 52, 53, 54] evaluated
Twenty-one randomised controlled trials’ [60,63,67, 70,71, 72, 73,75,78,81,85, 86, 87,91,93,94,100,101,106, 108,111] enrolled 18,543 patients of which, 50% (n = 9260) patients were randomised to the enteral nutrition groups. The remaining half (n = 9283) were randomised to the parenteral nutrition group, delayed enteral feeding or nil-by-mouth group.
In the EN group, a significant increase in vomiting episodes was observed compared to in parenteral nutrition group (
Four randomised controlled trials’ enrolled 324 patients, of which 50% of patients were allocated to the early enteral nutrition group and the other half to the delayed/conventional enteral nutrition group. A significant decrease in the vomiting episodes was observed in the early enteral nutrition group compared to delayed/conventional enteral nutrition group (
Six randomised controlled trials’ enrolled 1667 patients, of which 50% was randomised to the intervention group. There was a tendency of reduced anastomotic leaks in patients receiving enteral feeding (
The results of the meta-analysis demonstrate that restricted/goal-directed fluid therapy regardless of the type of fluid reduces mild GDF (vomiting) but not other complications associated with GDF. Enteral feeding, on the other hand, significantly increased vomiting episodes compared to parenteral nutrition but ‘early’ enteral nutrition significantly reduced the incidence of vomiting compared to delayed feeding. Enteral feeding was likely to reduce severe gut complications such as anastomotic leak and peritonitis compared with parenteral nutrition or an NBM status. Other mild to moderate variables of GDF (i.e. nausea, abdominal distension, ileus or diarrhoea) and moderate to severe complications (i.e. GI bleed, perforation, intestinal obstruction or intestinal ischaemia) were not associated with significant changes in outcomes. The results suggest that although the beneficial effects of restricted/goal-directed intravenous fluids and enteral feeding are essential to reduce some form of GDF, the impact is not prevalent for other variables of GDF (e.g. ileus and intestinal ischemia) associated with poor clinical outcomes. This may reflect the paucity of high-quality literature on the interaction between intravenous fluid (resuscitation) and enteral feeding as a combined therapy on the impact of GDF. The role of these two modalities in combination should be regarded as an important aspect in identifying the impact on the severity of GDF in acute surgical and critically ill patients.
Nevertheless, the use of trophic enteral feeding has been suggested in haemodynamically unstable patients to maintain gut integrity [4]. Authors have argued that enteral nutrition comes with its risks such as aspiration, pneumonia, intestinal obstruction, necrosis and pneumonitis intestinalis. However, the present study demonstrated no such differences for any of these complications. For gastrointestinal complications, a significant reduction in anastomotic leaks in the enteral nutrition group suggesting its benefits irrespective of the feeding route was observed. It is common practice in some areas, particularly intensive care, to commence patients on parenteral nutrition with anastomotic leaks before a trial of enteral nutrition. However, it should be acknowledged that a correct assessment for an enteral nutrition challenge can be countered in patients on parenteral nutrition with significant complications (e.g. anastomotic leaks), hence lowering the threshold of initiating enteral nutrition. Barlow et al. [93] found a lower incidence (2 vs 7) of anastomotic leaks in the early enteral nutrition group. They attributed a three-day shorter length of stay and reduced postoperative complications from installing early enteral nutrition. A similar effect was confirmed by a Cochrane review [115] in which enteral nutrition reduced the risk of anastomotic leaks from 27% in the standard group to 13% in early enteral group. These results affirmed with the present findings. It is hypothesised that enteral nutrition may improve perfusion at the anastomosis site, which promotes mucosal wound healing and prevents further leaks.
In comparison, Lewis et al. (2009) did not support this finding and observed mortality of 50% in the intervention group (enteral group) with anastomotic leaks [114]. However, it is likely that a smaller sample size may result in a false positive rate for mortality, thus exaggerating the magnitude of the negative result. The benefit of enteral feeding in complications such as perforation and peritonitis has been confirmed by several reports, which resonated with our findings. Early enteral feeding seems to maintain gut integrity by improving mucosal circulation and oxygen delivery that may reduce the risk of peritonitis [74, 87,113].
The present study is not without limitations:
The severity score in majority of the studies including surgical patients was low (ranging between I to III) hence the overall effect may be confounded by the clinical severity of the cohort. The majority of studies were conducted in stable postoperative patients and results may not be generalisable to a high-risk group, e.g. septic shock.
Critically ill patients are a heterogeneous group, and the effect on gut function can differ with specific sub-population. Such high-risk heterogenous patients need to be assessed in robust, well-designed, and randomised controlled trials. A possible stumbling block may be the ethical dilemma of implementing clinical trials using regimented interventions in these patients is often challenging for institutions and ethics committees.
Individualised unit protocols were variable with prescription of fluid and enteral feeding regimes possibly confounding the overall impact on GDF outcomes.
Most studies included small numbers of patients and were single-centred studies.
Postoperative morbidity manifested as GDF may be associated with the type of surgical procedure or manipulation of the bowel during surgery which may be associated with inducing a surgical stress response. However, this is expected to be low in our study, considering that the majority of the cohort included stable postoperative patients.
The majority of our studies found no differences between long-term endpoints (mortality and length of stay) but the occurrence of GDF was excluded from primary endpoints.
Most importantly, it was difficult to define or classify gut dysfunction because, until now, there is no valid, objective or a reliable scoring system to assess gut function in intensive care patients [125]. This suggests the need to develop a novel scoring tool to address this concern in future trials. Due to fewer studies on the effect of intravenous fluids and enteral nutrition on GDF, our meta-analyses may have been underpowered to see significant outcomes on GDF. Overall, studies on intravenous fluid remain mostly inconclusive, and potentially the impact of intravenous fluids may project variable outcomes when applied to a homogenous cohort instead of heterogeneous patient groups.
Further, inconclusive results from large-scale fluid and enteral feeding trials raise the suspicion that GDF may be the missing link, which perhaps may be associated with long-term outcomes. This dimension is often ignored when evaluating endpoints. To observe a difference in the key outcome, we first need to understand the combined effects of intravenous fluids and enteral nutrition in influencing clinical outcomes, including GDF. It is expected that as a result of the potential interaction between these two modalities, patients receiving liberal fluid resuscitation and early aggressive feeding are more likely to be at risk of severe GDF. More work is required to understand the implications of intravenous fluids and enteral nutrition on GDF and how this may impact overall patient outcomes. Future studies should evaluate this potential interaction and assess the combined impact of these two modalities on GDF in surgical and critically ill patients.
A restricted/goal-directed fluid regime and early enteral feeding compared to parenteral or a nil-by-mouth regime may reduce the risk on mild GDF in some, but not all complications of severe GDF. Because of a preventive strategy, we need to first understand the interaction between both (intravenous fluids and enteral feeding) and their impact on the gut so its implications can be translated into clinical practice eventually. Hence, it can be hypothesised that conservative fluid resuscitation and aggressive enteral feeding may potentially be the fundamental cause of developing severe life-threatening GDF (i.e. intestinal ischemia) and complications that can delay recovery and affect clinical outcomes in acute surgical and critically ill patients. Future research should evaluate and focus on an extended conceptual framework on the cross-interaction of conservative and aggressive modes across these two interventions and its impact on various levels of severity of GDF.