Evidence in Intensive Care - Gastro-intestinal

Systematic review and meta-analyses in blue.
Other high impact trials in red.

Stress ulcer prophylaxis
Crit Care Med 2010 38(11):2222-2228

SUP reduces GI bleeding if not enterally fed (EN).
No difference in GI bleeding if EN.
VAP and mortality increased if SUP combined with EN.

EN improves mucosal blood flow and immunity.
Some studies suggest EN better at acid suppression than H2s.
SUP core part of ventilator bundle in UK and USA. (SUP, head elevation, daily sedation hold, VTE prophylaxis.
USA adds chlorhex mouth care to UK bundle.

Recommendations for SUP in ventilator bundle come from meta-analysis of studies pre 1990s.
European bundle has no SUP:
(ICM 2010;36:773-780)
No ventilator tubing changes unless indicated.
Hand hygiene.
Trained staff.
Sedation holds and weaning protocols.
Oral chlorhex.



Early versus late parenteral nutrition in critically ill adults.
EPaNIC. Casaer et al. N Engl J Med 29 June 2011;

Prospective , randomised, controlled, parallel-group, multicentre trial. 4640 patients in 7 ICUs.
TPN within 48 h or from day 8. Protocol of early enteral nutrition followed in both groups.
Mortality - no difference.
ICU and hospital LOS shorter in late group.
Fewer infections in late group.
LOV and LORRT was shorter in the late group.
Reasons unclear.
Trial supports
late use of TPN.

Early TPN vs late enteral or parenteral nutrition. Prospective observational.
JPEN 2011;35:160-168

Meta-analysis in ICM 2005;31:12-23 concluded early TPN better than late enteral (mortality reduction) but this controversial.
This study concluded early TPN results in greater intake of calories and protein but no outcome improvements.
Supports current recommendations to maximise enteral nutrition before starting TPN.
Benefit or harm from early TPN remains unclear - need decent trial - now done - see above EPaNIC trial

Early enteral feeding
ICM 2009;35(12):2018-2027

Early EN defined as within 24h.
Significant reduction in mortality and pneumonia.

Number of patients small and trial heterogenous.
Data not robust enough to drive practice.
Did not include trial in JAMA 2009 which showed only trend to reduced mortality.


Initial trophic vs full enteral feeding in patients with acute lung injury: the EDEN randomized trial.
The National Heart, Lung, and Blood Institute ARDS Clinical Trials Network. JAMA 2012; 307: 795–803.

Background - trophic feeding better than no feeding and may be better than full feeding - need ref.

Aim was 20kcal/kg compared to 25-30kcal/day.
Actually got 80% of calculated requirement compared to 25%
No difference in LOV, mortality, organ failure, infections of LOS ICU.
Less vomiting, constipation, prokinetic agents, antidiarrhoeal drugs, gastric volumes in trophic group.
More insulin in higher feed group.
Suggests there is a wide window of nutritional intake that has little effect on outcome.


Enteral nutrition compared to TPN is associated with fewer infectious complications in SAP.
Arch Surg 2008;143:1111-1117

Is there such a thing as bacterial translocation from gut to blood stream? It happens in lab rats but never proven in humans. If it does happen it could lead to septicaemia or endotoxaemia or distant septic emboli but why should it lead to pancreatic or peripancreatic infection (particularly of the non-perfused portion of the necrotic pancreas)?

Nutrition in pancreatitis
The surgeon 8 2010 105 – 110
NJ compared to PN reduces infectious complications, inflammation, organ failure and radiological changes of pancreatitis. No mortality difference.
NJ feeding safe if started within 6h.
NG equivalent to NJ.
Probiotics not beneficial and may increase mortality.

IV glutamine should be used in PN for pancreatitis.
No evidence for prokinetics in pancreatitis but there is in critically ill in general so should be used in SAP.
Antioxidants may be of benefit.

EN vs PN in pancreatitis
Cochrane review 2010

EN reduces mortality, infections, MOF and need for surgery compared to PN.

No of patients small.
EN should be started early and PN only used after several days of efforts to establish PN.

Glycaemic control:

Hypoglycemia and risk of death in critically ill patients.
The NICE-SUGAR study investigators. N Engl J Med 2012;367:1108–1118.

Original NICE-SUGAR study showed 6.5% increase in 90 day mortality with tight glucose control.
This post hoc analysis demonstrates that moderate and severe hypoglycaemia are associated with increased mortality
Authors recommend targeting blood glucose at 8 - 10 mmol/l.

GKI infusions
J clin pharm 2009 49:758

GKI suppresses circulating levels and myocardial uptake of FFAs.
Exogenous glucose provides energy for the myocardium.
Potassium promotes electrical stability.
In sepsis the heart develops insulin resistance due to pyruvate dehydrogenase de-activation. Insulin reactivates this to allow optimal glucose and lactate oxidisation thus improving inotropy.

No evidence of benefit in MI or cardiac surgery.
No evidence either way in sepsis or shock.

Large trials not warranted unless small trials suggest potential benefit.


Child-Pugh grading of cirrhosis and risk of variceal bleeding.

Grade A 5-6
Grade B 7-9
Grade C >10
>8 is high risk for bleeding

Biliruben, albumin, PT, ascites, encephalopathy

Early use of TIPS in patients with cirrhosis and variceal bleeding.
García-Pagán JC et al. N Engl J Med 2010;362:2370–2379.

Patients with cirrhosis with acute variceal bleeding with high risk of treatment failure have reduced mortality and treatment failure with no increase in adverse events with early TIPS.
How should this guide practice?
There should only be 1 endoscopy and treatment with any evidence of rebleeding prompting consideration of TIPS.
Those who have active bleeding at endoscopy with a Child-Pugh grade B or grade C with no active bleeding should be considered for TIPS / discussed with a TIPS centre.
We do not know if it is beneficial for those with a score of >13 (high morbidity and mortality).
Cardiac dysfunction is associated with morbidity and mortality with TIPS so all should undergo assessment of LV and RV and PAP. RV overload or ventricular failure are relative CIs.

Glucocorticoids plus N-acetylcysteine in severe alcoholic hepatitis
Nguyen-Khac E et al. N Engl J Med 2011; 365: 1781–1789.

Steroids improve outcome in alcoholic hepatitis in some sub groups. A fall in biliruben should encourage ongoing steroid therapy while they should be stopped if there is no fall.
This study showed combining prednisalone and NAC showed improved 1 month survival but no improvement at 6 months.
Enough rationale to support 5 days of NAC in acute alcoholic hepatitis.

Transfusion strategies for acute upper GI bleeding.
Villanueva et al. NEJM 2013;368:11-21
Restrictive (HB >7) vs liberal (HB >9) transfusion strategy in acute GI bleeding.
Mortality and LOS better in restrictive group.
Effect greatest in those with less severe liver disease (no difference in severe Child-Pugh disease).