Introduction Percutaneous endoscopic gastrostomy (PEG) insertion is a valuable means of providing enteral nutrition. Previous studies have identified risk factors for poor outcomes and a scoring system devised. Further studies have shown other variables might be important including CRP and neutrophil-lymphocyte ratio (NLR). This study aimed to externally validate the score and analyse for additional variables associated with 30 day mortality.

Methods Retrospective review of new PEG insertions in our institute between July 2009 to October 2015. Patient demographics, indication for PEG insertion, bioprofile, ASA grade and survival were recorded. The Sheffield Gastrostomy score (SGS) and the NLR were then calculated for each patient. Univariate analysis was performed and receiver operating characteristic (ROC) curves constructed to assess the discriminative ability of the scoring system.

Results 171 PEG insertions (median age 73 years, 91 males) were included. Indication for PEG was Stroke (50), Cancer (30), Neurological (24), Dementia (5), Other (58) and not stated (4). Univariate analysis showed that 30 day mortality was associated with ASA grade 3–4 (p = 0.026) and NLR (p = 0.012) and a trend towards age > 70 (p = 0.085) and low albumin levels (p = 0.058). For each increasing SGS gradation mortality rose, with 5.3% of those scoring zero dying compared to 38.5% scoring three. The area under the ROC curve was 0.617 (95% confidence interval 0.534–0.695, p = 0.04) showing reasonable discrimination.

Conclusion The SGS displayed reasonable predictive ability in this external sample but identified further predictors (ASA grade and NLR). This suggests that the SGS requires remodelling with additional variables identified from a large, multicentre prospective study.

Disclosure of Interest None Declared