Nutritional trials using high protein strategies and long duration of support show strongest clinical effects on mortality.

There is increasing evidence from randomized-controlled trials demonstrating that nutritional support improves clinical outcomes in the population of malnourished medical inpatients. We investigated associations of trial characteristics including clinical setting, duration of intervention, individualization of nutritional support and amount of energy and protein, and effects on clinical outcomes in an updated meta-analysis.We searched Cochrane Library, MEDLINE and EMBASE, from inception to December 15, 2020. Randomized-controlled trials investigating the effect of oral and enteral nutritional support interventions, when compared to usual care, on clinical outcomes of malnourished non-critically ill medical inpatients were included. Two reviewers independently extracted data and assessed risk of bias. The primary endpoint was all cause-mortality within 12-months.We included 29 randomized-controlled trials with a total of 7,166 patients. Heterogeneity across RCTs was high, with overall moderate study quality and mostly moderate or unclear risk of bias. Overall, there was an almost 30%-reduction in mortality in patients receiving nutritional support compared to patients not receiving nutritional support (253/2960 [8.5%] vs. 336/2976 [11.3%]) with an odds ratio of 0.72 (95% CI 0.57 to 0.91, p = 0.006). The most important predictors for the effect of nutritional trials on mortality were high protein strategies (odds ratio 0.57 vs. 0.93, I2 = 86.3%, p for heterogenity = 0.007) and long-term nutritional interventions (odds ratio 0.53 vs. 0.85, I2 = 76.2%, p for heterogenity = 0.040). Nutritional support also reduced unplanned hospital readmissions and length of hospital stay.There is increasing evidence from randomized trials showing that nutritional support significantly reduces mortality, unplanned hospital readmissions and length of stay in medical inpatients at nutritional risk, despite heterogeneity and varying methodological quality among trials. Trials with high-protein strategies and long-lasting nutritional support interventions were most effective.