An Umbrella Review and Meta-analysis of Interventions, Excluding Enteral and Parenteral Nutrition, Initiated in the Hospital for Adults with or at Risk of Malnutrition

• Published in: The American journal of clinical nutrition Vol. 118; no. 3; pp. 672 - 696
• Main Authors: Wong, Alvin, Huang, Yingxiao, Sowa, Przemyslaw M., Banks, Merrilyn D., Bauer, Judith D.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.09.2023; American Society for Clinical Nutrition, Inc
• Subjects: Adult; Bias; Clinical outcomes; Clinical trials; Complications; Enteral Nutrition; hospitalization; Hospitals; Humans; Malnutrition; Malnutrition - therapy; Meta-analysis; Mortality; Nutrition; nutritional intervention; Parenteral nutrition; Parenteral Nutrition - methods; Quality assessment; Quality control; Quality of Life; Risk analysis; Search engines; Sequential analysis; Systematic Reviews as Topic; trial sequential analysis; umbrella review; trial sequential analysis; RCT; umbrella review; TSA; hospitalization; ROB; meta-analysis; EN; SMD; AMSTAR-2; QOL; malnutrition; LOS; RD; MD; nutritional intervention; SRMAs; GRADE; PN
• DOI: 10.1016/j.ajcnut.2023.07.003
• ISSN: 0002-9165; 1938-3207

Multiple systematic reviews and meta-analyses (SRMAs) on various nutritional interventions in hospitalized patients with or at risk of malnutrition are available, but disagreements among findings raise questions about their validity in guiding practice. We conducted an umbrella review (a systematic review of systematic reviews in which all appropriate studies included in SRMAs are combined) to assess the quality of reviews, identify the types of interventions available (excluding enteral and parenteral nutrition), and re-analyze the effectiveness of interventions. The databases MEDLINE/PubMed, CINAHL, Embase, The Cochrane Library, and Google Scholar were searched. AMSTAR-2 was used for quality assessment and GRADE for certainty of evidence. Updated meta-analyses with risk of bias (ROB) by Cochrane ROB 2.0 were performed. Pooled effects were reported as relative risk (RR), with zero-events and publication bias adjustments, and trial sequential analysis (TSA) performed for mortality, readmissions, complications, length of stay, and quality of life. A total of 66 randomized controlled trials were cited by the 19 SRMAs included in this umbrella review, and their data extracted and analyzed. Most clinical outcomes were discordant with variable effect sizes in both directions. In trials with low ROB, interventions targeting nutritional intake reduce mortality at 30 d (15 studies, n: 4156, RR: 0.72, 95% CI: 0.55, 0.94, P: 0.02, I2: 6%, Certainty: High), 6 mo (27 studies, n: 6387, RR: 0.81, 95% CI: 0.71, 0.92, P = 0.001, I2: 4%, Certainty: Moderate), and 12 mo (27 studies, n: 6387, RR: 0.80, 95% CI: 0.67, 0.95, P: 0.01, I2: 33%, Certainty: Moderate), with TSA verifying an adequate sample size and robustness of the meta-analysis. Existing evidence is sufficient to show that nutritional intervention is effective for mortality outcomes at 30 d, 6 mo, and 12 mo. Future clinical trials should focus on the effect of nutritional interventions on other clinical outcomes. The protocol is registered on PROSPERO (CRD42022341031).