Risks, Benefits, and Effects on Management for Biopsy of the Papilla in Patients with Familial Adenomatous Polyposis.

Abstract: Background & Aims Ampullary and duodenal cancer are the leading causes of death in patients with familial adenomatous polyposis (FAP) after colectomy has been performed. Risk of duodenal cancer is determined based on Spigelman stage (SS) of duodenal polyposis. Guidelines recommend endoscopic surveillance of the duodenum and visualization of the papilla to stage duodenal polyposis. There is no consensus on whether biopsies should be routinely collected from duodenal papilla and findings included in SS. Additionally, there are no data on the risk of pancreatitis after biopsy collection from papilla of patients with FAP. We studied the incidence of pancreatitis after biopsy of the papilla in patients with FAP and effects of biopsy findings on SS. Methods We identified consecutive patients with FAP at a single center from January 2011 through December 2018 with > 1 endoscopy with biopsy of the papilla. Patients with history of foregut surgery were excluded. We identified 273 patients with FAP who had biopsies collected from papilla over 792 EGDs, with 1–8 independent exams with biopsy per patient. We collected demographic, endoscopic, and histology data from patients and calculated SS with vs without biopsy findings. Post-procedural pancreatitis was defined by 2 of the following: abdominal pain, lipase level 3-fold the upper limit of normal, or radiography findings consistent with pancreatitis within 7 days of esophagogastroduodenoscopy (EGD). Results Pancreatitis developed in 2 patients (0.73%): 1 after biopsy of a normal-appearing papilla and 1 after biopsy of an abnormal appearing papilla. Inclusions of biopsy data increased SS in 36 patients (13.2%), with consideration of prophylactic duodenectomy for 3.3%. Conclusions Pancreatitis after biopsy of the duodenal papilla is rare. Histology data obtained from biopsy of the papilla in patients with FAP can change SS and affect patient management.