Diagnostic yield from gastroscopy: a UK national endoscopy database analysis

Introduction

Determine the indication, frequency, and diagnostic yield of oesophagogastroduodenoscopies (OGDs) performed in the UK; helping optimise capacity and inform future guidelines.

Methods

Cross-sectional analysis of OGDs conducted between 01/03/2019–29/02/2020 and uploaded to the UK National Endoscopy Database (which captures data from >90% UK endoscopy units). Procedures without a recorded diagnosis and those on patients aged <18 were excluded. Proportion of OGDs finding each diagnosis (normal/minor pathology, major pathology, Barrett’s Oesophagus (BO), cancer) was calculated by indication. Diagnostic yield from symptomatic and surveillance OGDs further analysed by patient age.

Results

638,484 OGDs were analysed: 66% were symptomatic, 13% therapeutic/emergency, 12% surveillance, 8% potential upper GI bleeds, 1% abnormal prior investigation. 81% reported only normal findings or minor pathology, 7.6% BO, 5.6% other major pathology, 5.2% ulcer, and 1.8% cancer. 30% of symptomatic OGDs were performed in patients <50, when compared to those performed on patients >50, these were more likely to report normal/minor pathology (95% vs 88%, p<0.01), with low cancer yield (0.1% vs 1.4%, p<0.01). 24,894 (4%) OGDs were performed for BO surveillance;19% of which did not report BO and 9% performed on patients over 80. 20% of OGDs performed as gastric ulcer follow-up were on patients <50, with only 0.2% reporting cancer.

Conclusion

Yield of major pathology in symptomatic young patients was low and justification for surveillance was unclear for some groups. Finding suggest changes to referral and surveillance recommendations could help address capacity challenges for endoscopy services