A retrospective study of outcomes and the validation of the Sheffield Gastrostomy Score in PEGs, RIGs and PIGs

There are limited studies comparing the safety and effectiveness of Radiologically Assisted Gastrostomies (RAGs) against Percutaneous Endoscopic Gastrostomies (PEGs). The Sheffield Gastrostomy Score (SGS) can be used to help predict 30-day mortality, more information is needed on its validity in RAGs. Our aim is to compare mortality between RAGs (Radiologically Inserted Gastrostomies (RIGs) and Per-oral Image Guided Gastrostomies (PIGs)) with PEGs and validate the SGS.

Data on gastrostomies newly inserted in three hospitals from 2016-2019 were retrospectively collected. Demographics, indication, insertion date, date of death, inpatient status and blood tests (albumin, CRP and eGFR) were recorded.

1977 gastrostomies were performed: Gastrostomy mortality at 7 days was 1.3% and at 30 days was 6%. There was a 5% 30-day mortality for PEGs, 5.5% RIGs, 7.2% PIGs (p = 0.215). Factors increasing 30 day mortality were age ≥60 years (p = 0.039), albumin <35 g/L (p = 0.005), albumin <25 g/L (p < 0.001) and CRP ≥10 mg/L (p < 0.001). For patients who died within 30 days; 0.6% had an SGS of 0, 3.7% = 1, 10.2% = 2 and 25.5% = 3, with similar trends for RAGs and PEGs. ROC curves showed the area under the curve for all gastrostomies, RAGs and PEGs as 0.743, 0.738, 0.787 respectively.

There was no significant difference between 30-day mortality for PEGs, RIGs and PIGs. Factors predicting risk include age ≥60 years, albumin <35 g/L, albumin <25 g/L and CRP ≥10 mg/L. The SGS has been validated in this study for PEGs and for the first time in RAGs as well..