Trial of the Route of Early Nutritional Support in Critically Ill Adults

Harvey et al for the CALORIES Trial Investigators. NEJM 2014; 371:1673-1684 doi:10.1056/NEJMoa1409860

Clinical Question

  • In critically ill patients, does parenteral nutrition compared with enteral nutrition improve mortality?


  • Open, multi-centre, parallel-group, randomised control trial
  • Intention to treat analysis
  • Patients assigned to enteral or parenteral nutrition using a telephone randomisation system in a 1:1 ratio
    • Allocation balanced by ICU, age, surgical status, and presence/absence of severe malnutrition
  • 90% powered to detect an absolute risk reduction in 30 day mortality of 6.4%
    • calculated 2400 patients required


  • 33 adult general ICUs in England
  • Data collected between June 2011 and March 2014


  • Inclusion: Adults >18 years of age; expected to require nutritional support for > 2 days, starting within 36 hours of admission
  • Exclusion: Not suitable for either enteral or parenteral nutrition; received nutritional support in last 7 days; presence of gastrostomy/jejunostomy; pregnancy
  • 2400 adult patients with unplanned admissions to ICU


  • Parenteral nutrition via a dedicated lumen of a central venous catheter


  • Enteral nutrition via a nasogastric or nasojejunal tube
In both groups, nutrition product used was determined by pre-existing local practice. Energy targets were 25kcal/kg/day in both groups, with the aim of reaching target within 48-72 hours. Nutrition was initiated within 36 hours of admission, and continued exclusively for 5 days unless: the patient was transitioned to exclusive oral feeding, was discharged from ICU or died


  • Primary outcome: All-cause mortality at 30 days. No significant difference
    • 33.1% parenteral group vs. 34.2% enteral group
  • Secondary outcomes:
    • Episodes of vomiting and “clinically significant” hypoglycaemia were greater in the enteral feeding group. “Serious” hypoglycaemia was not different (0.4% vs 0.3%)
    • No statistical difference in the 16 other secondary outcome measures, including:
      • Duration of organ support
      • SOFA scores
      • Treated infectious and non-infectious complications (including aspiration events or pneumonia)
      • ICU and hospital length of stay
      • The duration of survival, and mortality at the time of discharge from the ICU, from the hospital and at 90 days
  • Serious adverse effects were low and similar in both groups (no p values)
    • regurgitation or aspiration (0.2% in parenteral group vs 0.3% in enteral group)
    • pneumothorax (0.1% in both groups)

Authors’ Conclusions

  • Early nutritional support through the parenteral route is neither more harmful nor more beneficial than such support through the enteral route


  • High quality RCT – reported in line with CONSORT criteria
  • Eligibility criteria defined appropriate to the population
  • Randomisation with minimisation to control for sensible key variables
  • Baseline characteristics were well matched with minimal exclusion criteria, meaning that results are applicable to a large percentage of the ICU population
  • Included calories from non-nutritional sources (e.g propofol) in calculations of total calories
  • Minimal loss to follow up
  • Overall mortality is a feasible percentage for ICU population


  • Not blinded – but this is was pragmatic trial and the primary endpoint is not subject to significant bias
  • Crossover occurred in 6.8% in parenteral group and 3.4% in enteral group
  • Caloric targets were not met in the majority of patients in either group, although the total caloric intake was similar (21.3 kcal/kg in parenteral group vs 18.5 kcal/kg in the enteral group). One of the potential advantages of parenteral feeding has been the effective delivery of calorific requirements particularly when early enteral feed cannot be established. If this cannot be achieved in a trial setting, then it is unlikely to be occurring in the context of our daily practice

Non-protocolised selection of nutrition products, monitoring, and titration, reflecting the variety of practice at each participating site. This is not a weakness in the study design as it needed to be generalisable and applicable to the participating ICU’s.

The Bottom Line

  • Early nutritional support through the parenteral route is neither more harmful nor more beneficial than through the enteral route. Enteral feeding does increase episodes of vomiting and hypoglycaemia but with no evidence of harm or nosocomial infection.
  • Daily calorific targets were rarely achieved in this study (< 40% in both groups). This reflects a possible deficiency in our feeding protocols and adherence within UK and highlights the need for this to be addressed.

External Links


Summary author: Dr Nick Jenkins and Dr Mike Ward-Jones on behalf of Portsmouth ICU Journal Club (edited by @stevemathieu75)
Summary date: 6th December 2014
Peer-review editor: @davidslessor

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.