Initiation of Enteral Feeding After Necrotizing Enterocolitis.
Management of necrotizing enterocolitis (NEC) consists of cessation of enteral feeding, intravenous antibiotic administration, and supportive treatment. There is no evidence-based recommendation regarding when to restart feeding after a NEC episode. We performed a systematic review and meta-analysis to examine the effect of timing of enteral feeding reinitiation on NEC recurrence.METHODS
MEDLINE, Embase, Google scholar, and Cochrane databases were searched. Human studies evaluating enteral feeding timing with a primary outcome of NEC recurrence were included. A total of 2,257 titles or abstracts were screened, and 47 full-text articles were analyzed. A systematic review and meta-analysis comparing NEC recurrence and other associated outcomes between early (<5 days after NEC diagnosis) and delayed (>5 days) initiation of enteral feeding after NEC were performed according to the PRISMA statement. The meta-analysis data were analyzed using RevMan 5.3 to estimate odds ratios (ORs) with 95% confidence intervals (CIs).RESULTS
Two retrospective observational studies met the inclusion criteria, comprising 56 cases in which enteral feeding was started early and 35 cases of delayed enteral feeding initiation. There were no randomized controlled trials (RCTs). The recurrence rates of NEC were unchanged between early (5.4%) and delayed (8.6%) enteral feeding groups (pooled OR = 0.61; 95% CI: 0.12-3.16; p = 0.56; I2 = 0%). Catheter-related sepsis (pooled OR = 0.20; 95% CI: 0.01-3.29; p = 0.26; I2 = 67%) and post-NEC stricture (pooled OR = 0.28; 95% CI: 0.07-1.18; p = 0.08; I2 = 23%) rates were not different between early and delayed enteral feeding groups.CONCLUSION
Initiating early enteral feeding, within 5 days of NEC diagnosis, is not associated with adverse outcomes, including NEC recurrence. In addition, catheter-related sepsis and post-NEC stricture rates were unchanged between early and delayed enteral feeding groups after NEC. However, the quality of the evidence from the review of literature is suboptimal. A further RCT is needed to confirm these results.