Twenty-six trials (2,496 participants) were included. Fourteen trials had appropriate methods of randomisation, 14 reported allocation concealment, 12 were described as double-blind. Only five trials explained their methods, six had a minimum follow-up of 30 days and 18 reported protocol violations. Eighteen trials scored 3 or more on the Jadad scale and were considered to have a lower risk of bias.
An immune-modulating nutrition diet reduced the risk of infectious complications (RR 0.64, 95% CI 0.55 to 0.74; 26 trials; little heterogeneity Ι²=0%) and noninfectious complications (RR 0.82, 95% CI 0.71 to 0.95; 20 trials; little heterogeneity Ι²=0%) compared with a standard diet. Mean length of hospital stay was significantly shorter by 1.88 days (95% CI 0.84 to 2.91 days; 20 trials; high heterogeneity Ι²=85%) for the immune-modulating nutrition diet group compared to a standard diet. There was no evidence of a difference between diets for mortality.
Results for infectious complications were statistically significant and similar to the main result in all three subgroups based on the timing of the immune-modulating nutrition diet initiation (pre-, peri- or postoperative). Noninfectious complications were only statistically significant in favour of immune-modulating nutrition for diets started in the postoperative period. Length of stay was statistically significant for peri- and postoperative diet initiation.