Effects of nutritional interventions on nutritional status in patients with gastric cancer: A systematic review and meta-analysis of randomized controlled trials

Background & aims: Nutritional interventions may improve quality of life, morbidity and mortality in gastric cancer (GC) patients. A growing number of randomized controlled trials (RCTs) evaluated different nutritional strategies - oral nutritional supplements (ONS), enteral nutrition (EN), enteral immunonutrition (EIN), parenteral nutrition (PN) and nutritional counselling - in GC patients. This systematic review and meta-analysis aims to assess the effects of these nutritional interventions on nutritional status of GC patients undergoing gastrectomy and/or chemotherapy. Methods: A systematic literature search was performed in Pubmed, Web of Science and Scopus databases from inception to March 2020, based on fixed inclusion and exclusion criteria. Effect sizes were estimated with mean difference (MD) or standard mean difference (SMD) with 95% confidence interval (CI) and heterogeneity was assessed by measuring inconsistency (I2) based on chi-squared test. Pooled analyses and quality assessment were performed with Review Manager 5.3. Results: A total of 25 RCTs were identified, including ONS (n = 7), EN (n = 6), PN (n = 4), EIN (n = 5) and nutrition counselling (n = 3) interventions. Ten RCTs with 1838 patients were deemed eligible for pooled analyses. Body weight loss was found lower in ONS group versus control group (MD 0.77; 95% CI −0.02–1.56; p = 0.05). PN and EIN studies did not assess body weight, while all nutrition counselling studies did not show significant differences (p > 0.05). Twenty-three out of 25 studies evaluated serum protein levels - albumin (ALB) and/or prealbumin (PA) and/or transferrin (TF). ALB levels did not significantly differ (p > 0.05) in 4 ONS studies. Significant improvements of PA levels from baseline to postoperative day (POD) ≥ 7 were shown in EN compared with PN groups (MD 19.90; 95% CI 10.09–29.70; p < 0.0001). Compared with EN, EIN interventions showed no significant improvements of ALB, PA and TF levels (p > 0.05