P467 - GASTROINTESTINAL COMPLICATION RISK IN PEG PATIENTS: A TWO-ARM BAYESIAN META-ANALYSIS OF ENTERAL AND PARENTERAL NUTRITION

P467

GASTROINTESTINAL COMPLICATION RISK IN PEG PATIENTS: A TWO-ARM BAYESIAN META-ANALYSIS OF ENTERAL AND PARENTERAL NUTRITION

H. N. Arslan^1,*, G. Bozkul², S. Şenol Çelik¹

¹School of Nursing, Koç University, Istanbul, ²Faculty of Health Sciences, Tarsus University, Mersin, Türkiye

Rationale: PEG is widely used for long-term nutrition in patients unable to maintain adequate oral intake. However, the gastrointestinal (GI) complication risk associated with enteral nutrition (EN) versus parenteral nutrition (PN) in this population remains uncertain. Addressing this gap is crucial to optimize clinical decision-making in PEG care.

Methods: A two-arm Bayesian meta-analysis was conducted to compare GI complication risks between EN and PN among PEG patients. Six studies reporting binary GI outcomes were identified. A fixed-effect Bayesian model using the gemtc and rjags packages in R was applied. Odds ratios (OR) with 95% credible intervals (CrI) were calculated. Model convergence was evaluated using trace and density plots. Treatment ranking was assessed via SUCRA values and cumulative rankograms.

Results: A total of 3,379 patients were included (EN: 2,046; PN: 1,333). EN was associated with a significantly increased GI complication risk compared to PN (OR: 2.2; 95% CrI: 1.1–5.7). PN ranked highest in safety with a SUCRA score of 0.981. Model diagnostics confirmed convergence and stability, supporting the reliability of the findings.

Conclusion: This is the first two-arm Bayesian meta-analysis comparing GI risks of EN vs. PN in PEG patients. Despite the limited number of arms, Bayesian modelling provided consistent ranking and credible estimation where traditional methods may fall short. Findings suggest PN may be safer in high-risk PEG patients, underlining the importance of individualized nutrition planning.

Disclosure of Interest: None declared