TY - JOUR
T1 - Prediction and management of bleeding during endoscopic necrosectomy for pancreatic walled-off necrosis
T2 - results of a large retrospective cohort at a tertiary referral center
AU - Holmes, Ian
AU - Shinn, Brianna
AU - Mitsuhashi, Shuji
AU - Boortalary, Tina
AU - Bashir, Muhammad
AU - Kowalski, Thomas
AU - Loren, David
AU - Kumar, Anand
AU - Schlachterman, Alexander
AU - Chiang, Austin
N1 - Publisher Copyright:
© 2022 American Society for Gastrointestinal Endoscopy
PY - 2022/3
Y1 - 2022/3
N2 - Background and Aims: Lumen-apposing metal stents (LAMSs) provide an endoscopic method for management of walled-off necrosis (WON) and a gateway for the performance of endoscopic necrosectomy (EN). However, bleeding may occur in up to 20% of EN procedures. Predictive factors for bleeding in this patient population are unknown, and there is no agreed-on algorithm for the management of bleeding. The aim of this study was to evaluate preprocedural risk factors for bleeding in patients undergoing endoscopic drainage or EN for WON. Methods: A retrospective cohort of patients undergoing EN for WON was reviewed. Demographics, comorbidities, concurrent medications, and etiology of pancreatitis were recorded. Pre-, peri-, and postprocedural clinical variables were compared using the χ2 test and independent t test. Results: Between June 2014 and October 2020, 536 ENs were performed in 151 patients. Intraprocedural bleeding occurred during 28 procedures (5.2%) in 18 patients (11.9%). Endoscopic hemostasis was attempted in 8 patients (10 procedures). Eight patients (10 procedures) in total were treated by interventional radiology (IR). Thrombocytopenia (P =.006) and cirrhosis (P =.049) were associated with intraprocedural bleeding, although thrombocytopenia was present in only 1 patient. Identification of a vessel within the cavity endoscopically was also associated with bleeding (P <.001). On multivariate analysis, identification of a vessel within the cavity endoscopically remained a strong predictor of bleeding (P <.001), whereas cirrhosis was no longer significant. Patients who required IR for hemostasis were transfused with significantly more blood before the procedure than patients who did not (3.4 units vs.67 units, P =.002). Conclusions: EN for WON was associated with a 5.2% per-procedure risk of bleeding and an 11.9% per-patient bleeding risk. Identification of a vessel within the cavity during endoscopy is predictive of bleeding during EN. Patients who require more transfusions before endoscopy may require earlier intervention by IR.
AB - Background and Aims: Lumen-apposing metal stents (LAMSs) provide an endoscopic method for management of walled-off necrosis (WON) and a gateway for the performance of endoscopic necrosectomy (EN). However, bleeding may occur in up to 20% of EN procedures. Predictive factors for bleeding in this patient population are unknown, and there is no agreed-on algorithm for the management of bleeding. The aim of this study was to evaluate preprocedural risk factors for bleeding in patients undergoing endoscopic drainage or EN for WON. Methods: A retrospective cohort of patients undergoing EN for WON was reviewed. Demographics, comorbidities, concurrent medications, and etiology of pancreatitis were recorded. Pre-, peri-, and postprocedural clinical variables were compared using the χ2 test and independent t test. Results: Between June 2014 and October 2020, 536 ENs were performed in 151 patients. Intraprocedural bleeding occurred during 28 procedures (5.2%) in 18 patients (11.9%). Endoscopic hemostasis was attempted in 8 patients (10 procedures). Eight patients (10 procedures) in total were treated by interventional radiology (IR). Thrombocytopenia (P =.006) and cirrhosis (P =.049) were associated with intraprocedural bleeding, although thrombocytopenia was present in only 1 patient. Identification of a vessel within the cavity endoscopically was also associated with bleeding (P <.001). On multivariate analysis, identification of a vessel within the cavity endoscopically remained a strong predictor of bleeding (P <.001), whereas cirrhosis was no longer significant. Patients who required IR for hemostasis were transfused with significantly more blood before the procedure than patients who did not (3.4 units vs.67 units, P =.002). Conclusions: EN for WON was associated with a 5.2% per-procedure risk of bleeding and an 11.9% per-patient bleeding risk. Identification of a vessel within the cavity during endoscopy is predictive of bleeding during EN. Patients who require more transfusions before endoscopy may require earlier intervention by IR.
KW - Drainage/methods
KW - Endoscopy/methods
KW - Humans
KW - Necrosis/etiology
KW - Pancreatitis, Acute Necrotizing/etiology
KW - Retrospective Studies
KW - Stents/adverse effects
KW - Tertiary Care Centers
KW - Treatment Outcome
UR - http://www.scopus.com/inward/record.url?scp=85121149492&partnerID=8YFLogxK
U2 - 10.1016/j.gie.2021.10.015
DO - 10.1016/j.gie.2021.10.015
M3 - Article
C2 - 34678298
AN - SCOPUS:85121149492
SN - 0016-5107
VL - 95
SP - 482
EP - 488
JO - Gastrointestinal Endoscopy
JF - Gastrointestinal Endoscopy
IS - 3
ER -