S1058 Association of Hospital Procedural Volume with Outcomes of Endoscopic Retrograde Cholangio-Pancreatography

Abstract

Introduction: Endoscopic Retrograde Cholangio-Pancreatography (ERCP) is well established procedure for treatment and diagnosis of various pancreato-biliary diseases. Post-ERCP pancreatitis (PEP) is the most common complication of ERCP. Data on relationship of hospital volume with outcomes of ERCP is lacking. We aimed to study outcomes of ERCP based on hospital volume. Methods: We utilized National Inpatient Sample database 2016 to identify all adult ( >=18 years) patients with primary procedure diagnosis of ERCP. We then classified patients based on hospital volume Hospitals were categorized into 3 groups based on tertiles of annual procedural volume: low (5-50 cases/year), medium (50 to 110 cases/year), and high (110 - 505 cases/year). The primary outcome was PEP, in-hospital mortality, post-ERCP bleeding and bile duct perforation. Multivariate logistic regression analyses were performed to examine the association of hospital ERCP volume and outcomes. Results: Our study included 94,055 ERCPs procedures performed across 2102 hospitals with median annual procedural volume of 70 (interquartile range: 40 to 140). Compared to low volume hospitals, medium and high-volume hospitals had younger patients, less females, more elective procedures, and more patients with pancreatic cancer and less with CBD stone (Table 1). L Therapeutic interventions including bile duct dilation, bile duct stenting, pancreatic duct dilation and pancreatic duct stenting was significantly higher performed in high and medium volume hospital compared to low volume hospitals (Table 1). The length of stay and total cost was also more in high volume hospitals. On multivariate analysis, odds ratio of PEP comparing high vs. Low, medium vs. Low and high vs medium was 1.41 (95% confidence interval 1.33 - 1.49, p< 0.001), 1.21 (95% CI 1.14-1.29, p< 0.001) and 1.16 (1.10-1.23, p< 0.001), respectively (Table 2). Similarly, high volume hospitals also had significantly higher odds of post-ERCP bleeding compared to low (OR 2.05, p< 0.001) and medium volume hospitals (OR 1.71, p< 0.001). In-hospital mortality was also higher in high and medium-volume hospitals as compared to low volume hospitals (Figure 1). Conclusion: Higher hospital ERCP volume is associated with higher rates of PEP, bleeding, and worse outcomes. This is likely resulting from high comorbidity burden and higher number of patients with pancreatic cancer requiring advanced therapeutics in high volume centers. Further studies are needed to determine to validate our findings.Table 1Figure 1.: Forrest plot and influence analysis - Functional Success of SIS and SBS in relieving malignant hilar biliary obstruction.