Nutrition in Intensive Care

Richard Leonard St Marys Hospital, London

Slides and summary

www.st-marys-anaesthesia.co.uk ICU Downloads

Resources
www.evidencebased.net www.criticalcarenutrition.com ACCEPT study Martin, CM et al CMAJ 2004;170:197-204 cluster RCT of nutrition algorithms intervention ICUs had lower mean hospital LOS 10% reduction in ICU mortality but p=0.1 no difference in attainment of most nutritional targets

Six simple questions

Why do we feed ICU patients? Which patients should we feed? When should we start to feed them? Which route should we feed by? How much feed should we give? What should the feed contain?

Why feed ICU patients?

few data directly compare feeding with no feeding two trials and one meta-analysis suggest worse outcomes in un(der)fed patients catabolism of critical illness causes malnutrition malnutrition closely associated with poor outcomes many ICU patients are malnourished on admission

Aims of feeding ICU patients

treat existing malnutrition minimise (but not prevent) the wasting of lean body mass that accompanies critical illness

Nutritional assessment
important to identify existing malnutrition clinical evaluation is better than tests history
weight loss, poor diet, reduced function

examination
loss of subcutaneous fat, muscle wasting, peripheral oedema, ascites

Which patients should we feed?

which patients can safely be left to resume feeding themselves?
14 days starvation - dangerous depletion of lean body mass mortality rises in ICU patients with a second week of severe under-feeding 5 days without feed increases infections but not mortality one view is therefore that 5-7 days is the limit

Which patients should we feed?

however
ACCEPT study fed all patients not likely to eat within 24 hours one meta-analysis suggests reduced infections if patients are fed within 48 hours one meta-analysis of early TPN versus delayed EN found reduced mortality with early feeding

Which patients should we feed?

all malnourished patients all patients who are unlikely to regain normal oral intake within either 2 or 5-7 days depending on your view

When should we start to feed?

early feeding usually defined as starting within the first 48 hours of admission meta-analysis suggests reduced infections if patients are fed within 48 hours meta-analysis of early TPN versus delayed EN found reduced mortality with early feeding ACCEPT study aimed to start within 24 hours of ICU admission

When should we start to feed?

surgical issues
gastric, duodenal or high small bowel anastomoses critical mesenteric ischaemia

When should we start to feed?

without undue delay once the patient is stable this will usually be within 48 hours of ICU admission

What route should we feed by?

enteral feeding is claimed to be superior because
it prevents gut mucosal atrophy it reduces bacterial translocation and multi-organ failure lipid contained in TPN appears to be immunosuppressive

Is enteral feeding really better?

mucosal atrophy occurs far less in humans TPN is associated with increased gut permeability bacterial translocation does occur in humans and may be associated with infections increased gut permeability never shown to cause translocation translocation has never been shown to be associated with multi-organ failure enteral nutrition has never been shown to prevent translocation

Outcome evidence
does EN reduce infections?
pancreatitis - probably abdominal trauma - probably (2 trials of 3) head injury - evenly balanced other conditions no clear conclusion Lipman reviewed 31 trials and found no consistent effect meta-analysis by Heyland et al found reduced infections

EN is definitely a risk factor for VAP

Outcome evidence
does route of feeding affect mortality?
Heylands meta-analysis showed no effect on mortality Doig and Simpsons more robust meta-analysis found TPN reduced mortality when TPN and EN were directly compared; TPN versus early EN showed no difference

What route should we feed by?

enteral feeding is
cheaper easier and therefore preferable in most cases

parenteral feeding is obviously necessary in some

A pragmatic approach
Woodcock and MacFie Nutrition 2001 serious doubt about viability of enteral feeding within 7 days randomised to EN or TPN EN group
no reduction in infections higher incidence of under-feeding and feed-related complications

What route should we feed by?

EN preferred for majority on pragmatic grounds alone TPN obviously necessary for some if there is serious doubt that EN can be established in a reasonable time (ACCEPT study used 1 day; others would use 2 or 5 or 7)
commence TPN maintain at least minimal EN keep trying to establish EN

Enteral feeding
underfeeding is a serious problem NJ tubes
probably do not reduce VAP probably increase proportion of target delivered

prokinetic agents of unproven efficacy PEGs are not advisable in acutely ill patients

Diarrhoea
use fibre-containing feed avoid drugs containing sorbitol and Mg exclude and treat
Clostridium difficile infection faecal impaction

consider
malabsorption (pancreatic enzymes, elemental feed) lactose intolerance (lactose-free feed) using loperamide

TPN - complications
catheter-related sepsis
no benefit from single lumen catheters

hyperchloraemic metabolic acidosis electrolyte imbalance - low Pi, K, Mg refeeding syndrome abnormal LFTs rebound hypoglycaemia on cessation deficiency of thiamine, vit K, folate

How much feed should we give?

overfeeding is
useless - upper limit to amounts of protein and energy that can be used dangerous
hyperglycaemia and increased infection uraemia hypercarbia and failure to wean hyperlipidaemia hepatic steatosis

How much should we feed?

underfeeding is also associated with malnutrition and worse outcomes

How much feed should we give?

energy - 25 kCal/kg/day (ACCP)
indirect calorimetry
gold standard no evidence of benefit shows that other methods are inaccurate, especially as patients wean

equations
eg Schofield correct for disease, activity

How much feed should we give?

nitrogen
no benefit from measuring nitrogen balance nitrogen 0.15-0.2 g/kg/day protein 1-1.25 g/kg/day severely hypercatabolic patients (eg burns) may receive up to 0.3 g nitrogen/kg/day

What should the feed contain?

carbohydrate
EN: oligo- and polysaccharides PN: concentrated glucose

lipid
EN: long and medium chain triglycerides PN: soya bean oil, glycerol, egg phosphatides

nitrogen
EN: intact proteins PN: crystalline amino acid solutions

water and electrolytes micronutrients

Nutrition in acute renal failure

essentially normal CVVHD/F has meant fluid and protein restriction are no longer necessary or appropriate

Nutrition and liver disease

chronic liver disease
energy requirement normal lipolysis increased so risk of hypertriglyceridaemia protein restriction not normally needed, but in chronic encephalopathy intake should be built up from 0.5 g protein/kg/day BCAA-enriched feed may permit normal intake in the protein-intolerant

acute liver failure

gluconeogenesis impaired, so hypoglycaemia a risk

Nutrition in respiratory failure

avoid overfeeding at all costs energy given as 50% lipid may reduce PaCO2 and improve weaning, but unproven

Nutrition in acute pancreatitis

transpyloric feeding shown to
be safe reduce infection rate probably reduce mortality

malabsorption may require elemental feeds and pancreatic enzyme supplements TPN no longer standard therapy - however, some patients do not tolerate enteral feeding

What else should the feed contain?

glutamine?
selenium? immunonutrition?

Glutamine
primary fuel for enterocytes, lymphocytes and neutrophils; also involved in signal transduction and gene expression massive release from skeletal muscle during critical illness may then become conditionally essential is not contained in most TPN preparations

Enteral glutamine
reduces villus atrophy in animals and humans reduced pneumonia and bacteraemia in two studies - multiple trauma, sepsis one much larger study (unselected ICU patients) showed no effect difficult to give adequate dose enterally probably not worth it

Parenteral glutamine
Liverpool study in ICU showed reduction in late mortality London study of all hospital TPN patients showed no benefit French trauma study showed reduced infection but no mortality effect German ICU study improved late survivial in patients fed for more than 9 days

Parenteral glutamine
glutamine becomes conditionally essential in critical illness and is not given in standard TPN parenteral supplementation appears to be beneficial in patients requiring TPN for many days

Selenium
regulates free-radical scavenging systems low levels common in normals and ICU patients several small studies inconclusive but suggest benefit one large, flawed recent study showed nonsignificant mortality benefit watch this space

Immunonutrition
omega-3 fatty acids
produce less inflammatory eicosanoids

arginine
nitric oxide precursor enhances cell-mediated immunity in animals

nucleotides
DNA/RNA precursors deficiency suppresses cell-mediated immunity

Immunonutrition
few studies in ICU populations some found reduced infection in elective surgery one unblinded study has shown reduced mortality in unselected ICU patients; benefit in least ill (CCM 2000; 28:643) another showed increased mortality on re-analysis which barely failed to reach statistical significance (CCM 1995; 23:436)

Immunonutrition
first meta-analysis (Ann Surg 1999; 229: 467)
no effect on pneumonia reduced other infections and length of hospital stay increased mortality only just missing statistical significance did not censor for death

second meta-analysis (CCM 1999; 27:2799)

reduced infection reduced length of ventilation and hospital stay no effect on mortality

Immunonutrition
third meta-analysis (JAMA 2001; 286:944)
benefit in elective surgery increased mortality in ICU patients with sepsis

large Italian RCT (ICM 2003; 29:834)

compared enteral immunonutrition with TPN stopped early because interim analysis showed increased mortality in septic patients 44.4% vs 14.3%; p=0.039

Immunonutrition
arbitrary doses random mixture of agents mutually antagonistic effects diverse case mix

individual components need proper evaluation

Why do we feed ICU patients?

to treat existing malnutrition to minimise the wasting of lean body mass that accompanies critical illness

Which patients should we feed?

all malnourished patients all patients who are unlikely to regain normal oral intake within 2 days

When should we start to feed?

without undue delay once the patient is stable within 2 days

What route should we feed by?

EN preferred for majority on pragmatic grounds alone TPN obviously necessary for some if there is serious doubt that EN can be established in 2 (or 5, 7) days
commence TPN maintain at least minimal EN keep trying to establish EN

How much feed should we give?

25 kCal/kg/day equations indirect calorimetry

What should the feed contain?

carbohydrate lipid nitrogen water and electrolytes micronutrients