Comparative analysis of early versus late feeding after an EGD-based intervention: meta-analysis

The right time to restart feeding after EGD-based interventions is not well established. However, impaired nutrition from prolonged fasting may lead to unfavorable clinical outcomes. Herein, we performed a systematic review and meta-analysis to study the impact of early feeding (within 24 hours) versus late feeding (>24 hours) in patients undergoing acute therapeutic EGD-based interventions.

Multiple databases, including MEDLINE, Scopus, and Embase, were searched (in May 2022) using specific terms for studies evaluating the outcomes of early versus late feeding after EGD-based interventions. Outcomes of interest were early recurrent bleed (<7 days), late recurrent bleed (>7 days), mortality rates, length of hospital stay, and rate of blood transfusion. Standard meta-analysis methods were used using the random-effects model. I ²% heterogeneity was used to assess the heterogeneity. The Grading of Recommendations Assessment, Development and Evaluation Working Group approach was used to assess the certainty of evidence.

Eight studies (813 patients) were included in the final analysis. The "early feeding" cohort included 411 patients (31% women) with a mean age of 58 years, and the "late feeding" cohort included 402 patients (26.4% women) with a mean age of 57 years. Four studies (283 patients) evaluated patients with band ligation/sclerotherapy in acute variceal bleeding (224 patients), 2 studies (309 patients) with endoscopic treatment of peptic ulcer bleeding, and 2 studies in endoscopic mucosal dissection for gastric mucosal epithelial neoplasia. The pooled risk ratio (RR) of early recurrent bleeding in early feeding versus late feeding was 1.6 (95% confidence interval [CI], .7-3.7; I ² = 0%, P = .2). Similarly, the RR of late recurrent bleeding was .9 (95% CI, .4-2.3; I ² = 0%, P = .9). The pooled RR of total recurrent bleeding was 1.2 (95% CI, .7-2.2; I ² = 0%, P =.3). The pooled RR of mortality between the early feeding and late feeding groups was .6 (95% CI, .3-1.2; I ² = 0%, P = .16). The pooled rate of the standard difference of the mean length of hospital stay was -1.184 (95% CI, -1.5 to -.81; I ² = 92%, P = .00) between the early and late feeding groups. The pooled rate of the mean difference in blood transfusion between early and late feeding groups was .1 (95% CI, -.4 to .41; I ² = 77.6%, P = .96).

Our meta-analysis demonstrates early enteral feeding within 24 hours does not appear to have a higher risk of recurrent bleeding and mortality than delayed enteral feeding in patients undergoing EGD-based therapeutic interventions. However, early feeding is associated with a shorter length of hospital stay compared with late feeding.