Outcome Measures in Critical Care Nutrition Interventional Trials: A Systematic Review

Abstract

Historically, randomized controlled trials (RCTs) in critical care have used mortality as the primary outcome, yet most show no effect on this outcome. Therefore, there has been a shift in the literature to focus on alternative outcomes. This review aimed to describe primary outcomes selected in RCTs of nutrition interventions in critical illness. Systematic search of the literature identified RCTs of nutrition interventions in critically ill adults published between January 2007 and December 2018. Primary outcomes were documented and categorized as mortality, morbidity, health service/cost‐effectiveness, or nutrition outcome. The direction of effect of the intervention on the primary outcome (positive, neutral, or negative) was extracted. Of 1163 citations identified and assessed for eligibility, 125 articles were included. However, 52 articles (42%) did not provide a sample‐size calculation, leaving 73 articles (58%) for data extraction. The primary outcomes reported were morbidity (n = 24, 32.9%); health service/cost‐effectiveness (n = 21, 28.8%); nutrition outcomes (n = 16, 21.9%); mortality (n = 11, 15.1%); and other (n = 1, 1.4%). No RCTs with mortality as the primary outcome reported a difference between intervention and control. Trials that included other primary outcomes frequently reported a difference (n = 27 of 62; 43.5%). Morbidity was the most frequently reported outcome category in RCTs that evaluated a nutrition intervention in critically ill adults, with mortality least frequent. Power calculations were only reported in 58% of included studies. Trials were more likely to show a significant result when an outcome other than mortality was the primary outcome.