- Click here for explanation of classification scheme.

To perform a meta-analysis addressing whether enteral nutrition with immune-enhancing feeds benefits critically ill patients after trauma, sepsis, or major surgery.

·   Medline search (1967-1998)

·   Original articles

·   Published in English

·   Search terms: human, enteral nutrition, arginine, nucleotides, omega-3 fatty acids, immunonutrition, Impact, Immun-Aid

·   Published and unpublished articles from both manufacturers

·   Target population consisted of critically ill patients requiring enteral nutrition via a tube

·   Only trials with institutional approval and informed consent were included

·   Randomized trials that compared critically ill patients who received enteral  nutrition with a feed enriched with arginine with or without glutamine, nucleotides, and omega-3 fatty acids with pt receiving a standard enteral feed preparation and outcome measures of mortality, infection rate, days of mechanical ventilation,  ICU length of stay, hospital length of stay, and days with diarrhea.

·   Trials that do not have experimental and control groups matched in the beginning

·   Trials that patients overlap into larger studies

Descriptive and outcome data were extracted independently from the papers by the same two reviewers.  Abstraction differences were resolved by consensus between the two reviewers.  Access to the original data sets (4-6) allowed determination of whether the primary diagnosis was medical, surgical, or trauma based and identification of individual patients who died.

Outcome measures:

·   Mortality

·   Infection

·   Ventilator days

·   ICU stay

·   Hospital stay

·   Diarrhea days

·   Calorie intake

·   Nitrogen intake

 

Meta-analysis performed on an intent-to-treat basis.

 

Infection and mortality were treated as binary variables. 

 

Meta-analysis results were expressed in terms of a relative risk (RR) for the treatment group vs. control group, such that RR < 1 favored treatment group and RR > 1 favored control group.

 

Ventilator days, ICU LOS, hospital LOS, diarrhea days, calorie intake, and nitrogen intake were treated as continuous variables. 

 

The outcome measure was the number of days for the treatment group minus the number of days for the control group, such that a negative number favored the treatment group and a positive number favored the control group.

 

Subgroup analyses were performed on each of the medical, surgical, and trauma groups, for each outcome variable. 

 

Standard heterogeneity tests were conducted, and funnel plots of the primary outcome variables were created.

 

To correct for confounding effects of differences in overall mortality or in early death rate when overall mortality was the same on a comparison of time-dependent variables such as number of infections, ventilator days, and length of stay, the meta-analysis was repeated censored for no survivors.

 

· 15 trials met the inclusion criteria, one excluded due to significant differences in exp and control group at the start of trial, 2 excluded because pt were included in larger trials leaving 12 studies.

· 1557 pt enrolled, of which 1482 were included in an intent-to-treat analysis

 

· Interobserver correlation co-efficient for the quality assessment by two reviewers was r² = 0.85

· No overall effect of feeding on mortality and no heterogeneity between the studies.

· Immunonutrition has significant reduction in relative risk (RR) of acquiring infection in overall group and surgical subgroup, no heterogeneity between studies.

· Ventilator days—significant reduction in overall and trauma subgroup and remained so after censoring for mortality.  No heterogeneity between trials.

· ICU LOS—no significant reduction due to treatment.

· Hospital LOS—no heterogeneity between trials.  Significantly reduced in overall group, surgical subgroup, and medical subgroup.  Difference still significant after censoring for mortality.

· Diarrhea days—No increase in side effects of feeding were reported in pt receiving immunonutrition.

· Calories—No significant difference between groups after censoring for mortality.

· Nitrogen-Treatment group received 0.04g/kg/day more N than control group.  For the subgroup of trauma pts, the treatment group received 0.03 g/kg/day more than control group.  These results remained significant when censored for mortality.  There was no sig differences in N intake for either the surgical pt or the medical pt.

· Acute Physiology and Chronic Health II Score—No significant differences between treatment and control groups overall.  However, in the subgroup of medical pt, the APACHE II score was significantly higher by 1.2 points in the treatment groups. 

 

Trials varied in size, quality, and pt case mix, but all showed some positive effect on immunonutrition, which was maintained in the meta-analyses and could not be explained by differences in mortality or N intake between groups.  There was no evidence of any detrimental effect of immunonutrition.  Results of this analysis do not support that immune-enhancing feeds be reserved for the sickest pt.  The benefit was most marked in the surgical group.

Tests for heterogeneity were applied and duplicate analysis was avoided.