Review: timing or administration route of nutritional support does not affect death or disability in head injury
Q In patients with head injury, does timing or route of nutritional support affect morbidity and mortality?
Cochrane Injuries Group specialised register, Cochrane Central Register of Controlled Trials, Medline, EMBASE/Excerpta Medica, Web of Science, CAB Abstracts, CINAHL, National Research Register, ZETOC, and reference lists of identified studies and reviews.
Study selection and assessment:
randomised controlled trials (RCTs) evaluating timing (early v delayed) or administration routes (parenteral v enteral or gastric v jejunal enteral) of nutritional support in patients with acute traumatic brain injury of any severity. 11 RCTs (n = 534) met the inclusion criteria. 7 RCTs evaluated timing, and 7 RCTs evaluated administration route. Quality assessment of individual trials was based on adequacy of allocation concealment, which was unclear in 7 trials and inadequate in 4 trials.
all cause mortality, disability (Glasgow Coma Scale dichotomised such that death, persistent vegetative state, or severe disability = poor outcome and moderate disability or good recovery = good outcome), length of hospital stay, and infection.
Of 7 RCTs evaluating timing of nutritional support, patients receiving early or delayed support did not differ for all cause mortality (7 RCTs) or poor outcome with respect to disability (3 RCTs) (table). Of 7 RCTs comparing parenteral with enteral nutrition, groups did not differ for mortality (5 RCTs) or poor outcome (2 RCTs) (table). Results on length of hospital stay and infection rates could not be combined in a meta-analysis and were not statistically compared between groups.
In patients with head injury, no definite benefit exists for early or delayed nutritional support or for parenteral or enteral nutrition in terms of death and disability.
- John McArthur, RN, MHSc
The review by Perel et al attempts to determine the optimal method of delivery and timing of nutritional support in patients with head injury. The review itself is well done but is limited by the lack of well designed studies meeting criteria. Data on death and disability were not available for all included trials, and a possibility of bias exists in the findings of improved outcomes (in terms of death and disability) with early feeding. Although the pooled results suggest a trend towards better outcomes with parenteral nutrition, the precision of the point estimates for mortality was low. Other methodological limitations included inadequate or unclear allocation concealment and lack of intention to treat analysis. Moderate heterogeneity was evident among studies in timing of nutritional intervention and administration modality. Head injury prompts increased nutritional demands through alteration of metabolic responses, and concerns include a higher incidence of aspiration pneumonia when patients with head injury are fed enterally rather than parenterally. However, the review showed no difference between these methods in overall infections and mortality. Furthermore, no RCT data are available to support the use of supplemental total parenteral nutrition in patients with an intact gastrointestinal tract.
Much work still needs to be done to determine whether early or delayed nutritional support or parenteral or enteral nutrition is best for patients with head injury.
For correspondence: Dr P Perel, London School of Hygiene and Tropical Medicine, London, UK.
Source of funding: not stated.