Objectives: Protein is a fundamental component of critical care nutrition, but there has been uncertainty about the optimal amount. We undertook this systematic review and meta-analysis to examine the relationship between delivered protein and mortality in randomised controlled trials (RCTs) of nutritional interventions involving critically ill adults. Secondary outcomes included the effect of protein dose on lengths of stay, mechanical ventilation and incidence of infections.
Methods: We reviewed the relevant English-language literature published between 1966 and 2015 and identified RCTs comparing different strategies of nutritional support lasting at least 48 hours in critically ill adults. Articles were included if mortality was reported and the difference in delivered protein between interventions was significant (P < 0.05). We calculated summary estimates for mortality as odds ratios (ORs) with 95% confidence intervals (CIs) using a random-effects estimator, and we used meta-regression to assess the effect of delivered protein on mortality.
Results: From 3016 assessed records, 357 full-text articles were reviewed and 14 studies, investigating various interventions and routes of nutrition and comprising 3238 patients, were included. The mean protein delivered was 42.95 g/day (SD, 20.45 g/day) or 0.67 g//kg/day (SD, 0.38 g/kg/day) in patients receiving less protein, and 67.15 g/day (SD, 28.47 g/day) or 1.02 g/kg/day (SD, 0.42 g/kg/day) in the higher protein group. Provision of less protein did not influence mortality risk (pooled OR, 0.935; 95% CI, 0.716 -1.219; P = 0.618; I2 = 48.2%). Meta-regression analysis did not show a relationship between mean daily protein delivered and mortality (P = 0.433; I2 = 50.18%). There were no differences between groups in any secondary outcomes.
Conclusions: Delivery of varying amounts of nutritional protein was not associated with any effect on mortality.