This review assessed the effects of specialised nutritional support with the IMPACT formula on post-operative morbidity in patients undergoing elective surgery. Overall, IMPACT was shown to reduce post-operative infections and length of hospital stay. The main conclusion seems well supported and robust, but there was insufficient evidence to support administration and dosage recommendations.

This review assessed the effects of pre-, peri- and post-operative specialised nutritional support with the IMPACT formula on post-operative morbidity, compared with standard post-operative nutritional support, in patients undergoing elective surgery.

MEDLINE was searched for clinical trials in humans published between January 1985 and December 2003 using appropriate keywords (listed in article). Reference lists and relevant reviews were searched for additional citations. Investigators identified by the manufacturer of the IMPACT nutritional support formula were contacted for unpublished data. The language of publication was restricted to English, German, French, Spanish, Portuguese, Japanese or Chinese.

Eligible studies were restricted to randomised clinical trials of patients undergoing major elective surgery. The included populations were mostly patients undergoing gastrointestinal surgery. The intervention of interest was IMPACT (a nutritional support product containing arginine, omega-3 fatty acids and nucleotides), given before and/or after elective surgery. IMPACT was administered directly into the gastrointestinal tract or given orally. The control groups received one of the following: standard formulae, isocaloric and isonitrogenous support, or a low fat and low calorie solution delivered intravenously. Relevant outcomes included post-operative infectious complications, mortality, length of hospital stay and costs of hospital care.

The authors did not state how the papers were selected for the review, or how many reviewers performed the selection.

Two reviewers assessed the validity of the studies using the validated Jadad 5-point scale (appropriate randomisation, blinding, withdrawals and drop-outs).

The primary outcome data were extracted as the number of patients with one or more post-operative-acquired infections, length of hospital stay and hospital mortality. Data on the numbers of events in each group were used to derive relative risks (RRs) for dichotomous outcomes; means and standard deviations were used to derive the mean differences.

The authors did not state how the data were extracted for the review, or how many reviewers performed the data extraction.

Meta-analysis was carried out on the calculated RR for infection using the Mantel-Haenszel random-effects method. Length of stay data were summarised as a 'pooled simple difference', based on the calculated effect sizes for analysis. Subgroup analyses were carried out for trials administering the intervention pre-, peri- or post-operatively.

Sensitivity analyses were carried out to assess the impact of publication status and trial validity.

Formal tests of heterogeneity were carried out and partially reported. However, no further information on the tests used was provided.

This review included 17 randomised controlled trials (three unpublished) of 2,305 patients.

Eight studies were blinded. Ten studies scored 3 or more on the Jadad scale (acceptable quality) and 7 trials were judged as being of poor quality.

Post-operative infectious complications (16 studies) : the use of IMPACT was associated with a statistically significant decrease in the occurrence of post-operative infectious complications (RR 0.49, 95% confidence interval, CI: 0.42, 0.58, p<0.0001). There was no evidence of statistically significant heterogeneity (p=0.95). The beneficial effects of IMPACT were seen regardless of administration timing (pre-, peri- or post-operatively).

Six categories of post-operative morbidity were assessed. The following showed statistically significant benefits from IMPACT supplementation: wound infections, abdominal abscesses, pneumonia, urinary tract infections and anastomotic leaks. The numbers of septic episodes were reduced, but this was not statistically significant. There was no evidence of statistically significant heterogeneity for any of these subanalyses (p=0.81 to p=0.99).

Length of hospital stay (16 studies): patients receiving IMPACT were found to have significantly shorter hospital stays by a mean of 3.1 days (95% CI: -3.9, -2.3); the effect size was -0.66 (95% CI: -0.86, -0.45, p<0.0001). There was evidence of statistically significant heterogeneity (p=0.0001). The beneficial effects of IMPACT were seen regardless of administration timing (pre-, peri- or post-operatively).

Mortality (17 studies): no improvement in mortality rates was found when IMPACT was used, although rates overall were low (RR 0.72, 95% CI: -1.31, 0.39, p=0.28). There was no evidence of statistically significant heterogeneity (p=0.98).

Publication bias was assessed through sensitivity analyses based on publication status. Using only the published studies (14), IMPACT supplementation continued to reduce rates of post-operative infections and the length of hospital stay. When trial validity was incorporated into a sensitivity analysis, IMPACT supplementation continued to statistically significantly reduce rates of post-operative infection and the length of hospital stay.

IMPACT supplementation is associated with reduced infectious complications and shorter length of hospital stay, regardless of whether it is given pre-, peri- or post-operatively. The most beneficial outcomes were found when IMPACT was given during the pre-operative period of 5 to 7 days before surgery, at a suggested dosage of between 0.5 and 1.0 L/day.

The review question was clearly stated, with generally appropriate inclusion and exclusion criteria. No reason was given for the language restrictions applied, although a broad range was covered. The search terms were comprehensive, but only searching MEDLINE restricts the available pool of research and potentially eligible trials may have been omitted. There were insufficient details on the study selection, validity assessment and data extraction phases of this review to be confident that the chances of error or bias were suitably controlled. Details of the included studies did not present patient age ranges or dosage of the intervention formula, making it difficult to be sure the results are generalisable.

Meta-analyses were carried out to assess the impact of the intervention on various key outcomes. It was unclear from the paper why effect sizes were used to summarise primary studies instead of weighted mean differences, or how the 'pooled simple difference' was obtained. Further analyses appropriately assessed publication status and validity effects. Heterogeneity was present for studies assessing length of hospital stay, but this was not addressed. Heterogeneity was not reported as being assessed for many of the subsequent subgroup or sensitivity analyses.

The authors' conclusions that, overall, IMPACT is a beneficial intervention, follow from the evidence presented and, based on the numerous subgroup and sensitivity analyses, this effect seems fairly robust. However, further recommendations were made about the timing of IMPACT administration, and dosage, which do not seem to have been based on the evidence reported. No direct comparisons were reported or carried out for pre- versus peri- versus post-operative administration of IMPACT at various dosage levels, which would have been required to support the conclusions.

Practice: The authors stated that IMPACT supplementation should be administered for 5 to 7 days before surgery at a dosage of 0.5 to 1 L/day; where pre-operative support is not possible the formula should be given post-operatively.

Research: The authors did not state any implications for further research.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.