Risk factors for adverse outcomes at various phases of endoscopic ultrasound-guided treatment of pancreatic fluid collections: Data from a multi-institutional consortium.

Abstract

No comprehensive study has examined short- and long-term adverse outcomes of endoscopic ultrasound (EUS)-guided treatment of pancreatic fluid collections (PFCs) including walled-off necrosis (WON) and pseudocysts. In a multi-institutional cohort of 357 patients receiving EUS-guided treatment of PFCs (228 with WON and 129 with pseudocysts), we examined PFC type-specific risk factors for procedure-related adverse events (AEs), clinical failure, and recurrence. Odds ratios (ORs) and hazard ratios (HRs) with 95% confidence intervals (CIs) were computed using the logistic and Cox regression models, respectively, adjusting for potential confounders. Adverse events were observed predominantly in WON, and risk factors were WON extension to the pelvis (OR 2.49; 95% CI 1.00-6.19) and endoscopic necrosectomy (OR 5.15; 95% CI 1.61-16.5). Risk factors for clinical failure in WON treatment included higher Charlson Comorbidity Index (OR for ≥3 vs. ≤2, 2.58; 95% CI 1.05-6.35), extension to the pelvis (OR 3.63; 95% CI 1.57-8.43), nonuse of a lumen-apposing metal stent (OR 2.88; 95% CI 1.10-7.54), and percutaneous drainage (OR 3.73; 95% CI 1.27-10.9). Patients with pseudocysts extending to the paracolic gutter and the need for more than two endoscopic/percutaneous procedures had ORs for clinical failure of 5.28 (95% CI 1.10-25.3) and 5.52 (95% CI 1.61-18.9), respectively. Pseudocysts requiring the multigateway approach were associated with a high risk of recurrence (HR 4.00; 95% CI 1.11-11.6). The adverse outcomes at various phases of EUS-guided PFC treatment may be predictable based on clinical parameters. Further research is warranted to optimize treatment strategies for high-risk patients.