P188 Is there a threshold volume to optimise adverse events following drainage of PFC with LAMS?

Introduction

Lumen apposing metal stents (LAMS) are increasingly used for the endoscopic drainage of symptomatic pancreatic fluid collections (PFC). High technical and clinical success rates are reported. However, there is risk of adverse events (AE) including maldeployment and bleeding at the time of placement, and delayed AE of bleeding and stent embedment. Recent studies have reported an overall AE rate of between 13 and 24%. Whether or not there is a relationship between volume and AE is unclear. A recent consensus group report suggested that performance of 5-10 procedures under supervision might represent the minimum number to obtain competency. On the other hand, a multi-centre Italian study among 516 patients reported 15 cases as the minimum number associated with the lowest risk for AE. Whether or not this number is generalizable to other health care settings and whether the association differs between immediate and delayed AE is unknown. The aim of this study was to investigate the relationship between institutional procedural volume and AE and determine if any of the proposed volume thresholds were associated with AE.

Methods

Data from a large retrospective multicentre study involving 18 units from the UK & Ireland was retrieved. The study was performed with the aim of investigating the technical and clinical success of LAMS for PFC and the incidence of immediate and delayed AE and their associated risk factors. Data on LAMS placed for drainage of PFC in adults (>18 years of age) between 2015 and 2019 was collected. Initial univariable analyses was performed with subsequent multivariable logistic regression. All clinically relevant variables were included as variables and the volume cut-offs of ≤ 4, ≤ 10, <15 were investigated.

Results

A total of 1018 patients underwent EUS guided drainage of PFC (WON 52.9%, PC 47.1%) with LAMS between October 2015 and January 2020. Full follow up was available in 952. One patient with a sedation related AE was excluded. Immediate LAMS related immediate AE occurred in 3.7% (38/1017) and delayed AE in 17.5% (167/952). Immediate AE occurred in 3% (23/769) procedures in the ≥15 group and in 6.1% (15/248) in the <15 group (p=0.04) On multivariable logistic regression <15 procedures was associated with an increased risk of immediate AEs (OR, 2.21; 95% CI, 1.11-4.43; p=0.03). The Cut-offs of ≤ 4 and ≤ 10 were not associated with immediate AE and there was no volume relationship identified to delayed AE.

Conclusions

A minimum institutional volume of 15 procedures is associated with reduced risk for immediate AE, there was no relationship to delayed AE. These findings have important implications for training and service development.