Quick Links

Recommendations Summary

CI: Timing of Enteral Nutrition and Critical Illness 2006

Click here to see the explanation of recommendation ratings (Strong, Fair, Weak, Consensus, Insufficient Evidence) and labels (Imperative or Conditional). To see more detail on the evidence from which the following recommendations were drawn, use the hyperlinks in the Supporting Evidence Section below.


  • Recommendation(s)

    CI: Timing of Enteral Nutrition

    If the critically ill patient is adequately fluid resuscitated, then EN should be started within 24 to 48 hours following injury or admission to the ICU. Early EN is associated with a reduction in infectious complications and may reduce LOS. The impact of timing of EN on mortality has not been adequately evaluated.

    Rating: Strong
    Conditional

    • Risks/Harms of Implementing This Recommendation

      A series of case studies have indicated that EN administered to patients with inadequate mesenteric perfusion may be associated with hypoxia and might promote the development of small bowel hypoxia and necrosis (Rai et al, 1996; Lawlor et al, 1998; Melis et al, 2006).

    • Conditions of Application

      Early EN is reportedly well tolerated by ICU patients. The following are guidelines for avoiding intestinal hypoxia and bowel necrosis (Rai et al, 1996; Lawlor et al, 1998; Melis et al, 2006):

      • Feeding into the small bowel
      • Administering feedings in patients adequately resuscitated and when mean arterial pressure can be sustained at 70mm Hg or more
      • Use of iso-osmolar formulations
      • Advancement of feedings when tolerance demonstrated.

      Feedings should be held when a sudden period of hypotension occurs or if there is a need for increased dosages of pressor agents or an increased need for ventilatory support. Feedings should also be held with sudden increases in nasogastric output, sudden abdominal distention or development of abdominal pain.  Further evaluation is indicated.

    • Potential Costs Associated with Application

      No obvious costs are associated with the application of this recommendation.

    • Recommendation Narrative

      The conclusion statement is supported by several RCTs in head injury, blunt trauma and surgical patients and is supported by three meta-analysis studies which demonstrate the benefit of the use of EEN to reduce infectious complications in elective surgical patients and ICU patients. Sufficient data does not exist to demonstrate a significant difference in mortality with EEN.

      Mortality

      • In a positive quality review of head injury patients (Yanagawa et al, 2002), seven trials were identified to evaluate early vs. delayed nutrition (enteral and parenteral) support
        • The relative risk for death with EEN was 0.67 (95% CI=0.41-1.07) for 284 subjects.
      • One positive quality meta-analysis (Heyland et al, 2003), two neutral quality PRCTs (Taylor et al, 1999; Eyer et al, 1993), and one negative quality PRCT (Minard et al, 2000) found no differences in mortality.
      • In one positive quality meta-analysis (Marik and Zaloga, 2001), six of the 15 studies reviewed were evaluated for mortality
        • The mortality was 8% in the early group and 11.5% in the late-fed group
        • This difference was not significant, with an RR of 0.74 (95% CI=0.37-1.48).
      • In one neutral quality clinical trial (Ibrahim et al, 2002),  no difference was noted between early vs. late feeding groups for mortality; 20% vs. 27%, P=0.334. 
      • In one neutral quality RCT (Singh et al, 1998), mortality did not differ between treatment (early fed) and control groups; 18.2% vs. 19.1%.
      • Only one neutral quality meta-analysis (Lewis et al, 2001), reported that there were fewer deaths reported with EN vs. the control group.
      • Because available RCTs may have been underpowered to show a difference and the stronger meta-analyses do not agree, further trials are needed to answer this question.

       Infectious Complications

      • In head injury patients, two neutral quality PRCTs (Graham et al,  1989; Taylor et al,  1999)support EEN vs. late EN. 
      • One positive quality prospective, randomized study (Moore and Jones, 1986) and four neutral-quality RCTs (Grahm et al, 1989; Kudsk et al, 1996; Singh et al, 1998; Taylor et al, 1999) also support the conclusion that EEN reduces septic morbidity and infections.
      • Neutral quality RCTs (Eyer et al, 1993; Ibrahim et al, 2002) conversely reported an increase in infectious complications with EEN.
      • Early nutrition support (both EN and PN) is substantiated by two positive quality reviews (Heyland et al,  2003; Yanagawa et al, 2002), one positive quality meta-analysis (Marik and Zaloga, 2001) and one neutral quality review (Lewis et al, 2001), in which similar outcomes are demonstrated when patients are fed early with either PN or EN.
      • One negative quality PRCT (Minard et al,  2000) had limitations in study design,  small sample size, lack of blinding or quasi-randomization.

      Length of Hospital Stay

      • The conclusion is supported by one positive quality RCT (Moore and Jones,  1986) and four neutral quality studies: One retrospective cost effectiveness study (Neumayer et al, 2001)and three PRCTs (Grahm et al,  1989; Kudsk et al, 1996; Taylor et al, 1999). 
      • Seven studies found either a negative result or no differences between LOS and EEN. These include one positive quality meta-analysis (Heyland et al,  2003), two neutral PRCTs (Kompan et al,  1999; Eyer et al,  1993), one neutral quality RCT (Singh et al, 1998), one neutral quality clinical trial (Ibrahim et al,  2002), and one negative quality PRCT (Minard et al,  2000).
      • The workgroup noted the differences in specific patient populations and protocol limitations.
      • The variability in the results for the clinical finding of LOS (significant heterogeneity) from one positive quality meta-analysis (Marik and Zaloga, 2001) and one positive quality systematic review (Yanagawa et al, 2002) was noted.

    • Recommendation Strength Rationale

      • In fluid-resuscitated, critically ill patients, EN started within 24 to 48 hours following injury or admission to the ICU will reduce the incidence of infectious complications
        • Conclusion statement is Grade I.
      • In fluid-resuscitated, critically ill patients,  EN started within 24 to 48 hours following injury or admission to the ICU may reduce LOS
        • Conclusion statement is Grade II.
      • The impact of timing of EN on mortality has not been adequately evaluated
        • Conclusion statement is Grade V.