目的: 通过回顾北京大学第三医院近年来不同胆肠吻合术后行内镜逆行胰胆管造影术(endoscopic retrograde cholangiopancreatography，ERCP)患者的临床资料并再次阅读内镜图片，分析不同类型胆肠吻合术后特别是胃肠改道术后ERCP的特点、安全性和有效性。方法: 收集2005年1月至2021年12月在北京大学第三医院内镜中心因胆系疾病接受ERCP的胆肠吻合术后患者的临床资料及内镜图片，分为胆管-十二指肠吻合术(choledochoduodenostomy，CDD)、Roux-en-Y胆肠吻合术(Roux-en-Y hepaticojejunostomy，RYHJ)和Whipple术后三组，以及ERCP成功和失败两组，进行组间比较统计分析。结果: 共纳入89例患者，ERCP操作132例次，患者年龄9~80岁，中位年龄57岁，包括CDD组4例，RYHJ组30例，Whipple组54例，胆管-回盲部吻合1例。ERCP距手术的时间分别为30 (1~40)、2.75 (0.5~14)、2 (0.3~19)和10年，手术后出现胆系疾病症状的时间分别为240 (3~360)、12 (1~156)、22 (0~216)和60个月。50%的CDD经局部麻醉操作即可成功，RYHJ(96.7%)和Whipple(100%)基本均为全身麻醉，组间差异有统计学意义(P < 0.001)。CDD组首次进镜成功率为100.0%，RYHJ组和Whipple组分别为40.0%和77.8%，更换内镜后分别提升至43.3%和83.3%，不同术式间进镜成功率的比较差异有统计学意义(P < 0.001)。CDD、RYHJ、Whipple组插管成功率分别为100.0%、53.8%、86.7%，组间差异有统计学意义(P=0.031)。CDD、RYHJ、Whipple组ERCP成功率分别为100.0%、33.3%、78.8%，组间差异有统计学意义(P < 0.001)。术后并发症发生率为23.9%(21/88)，分别为感染(14.8%)、胰腺炎(9.2%)、出血(3.4%)、穿孔(2.3%)。胆肠吻合术后ERCP的常见原因分别为吻合口狭窄(50.0%，良性狭窄39.3%，恶性狭窄10.7%)、胆管结石(37.5%)、反流性胆管炎(12.5%)。吻合方式为预判ERCP成功与否的唯一因素(OR=7, 95%CI：2.591~18.912，P < 0.001)。结论: 伴胃肠改道的胆肠吻合术后ERCP需在全身麻醉下进行，具有良好的安全性和有效性，RYHJ的成功率显著低于Whipple术后，手术方式是决定ERCP能否成功的唯一影响因素。

Objective: To distinguish clinical features, safety and efficiency of endoscopic retrograde cholangiopancreatography (ERCP) in patients after bilioenteric anstomosis based on retrospectively analyzed clinical data and endoscopy procedures. Methods: Data extracted from patients after bilioenteric anstomosis due to biliary disease treated with ERCP from January 2005 to December 2021 in the Department of Gastroenterology, Peking University Third Hospital were retrospectively analyzed. Clinical data and endoscopic pictures were reevaluated and analyzed. The patients were divided into three groups, including the patients with choledochoduodenostomy (CDD), Roux-en-Y hepaticojejunostomy (RYHJ) and Whipple. Differences between ERCP success and failure were conducted. Results: In the study, 89 cases with 132 ERCP procedures were involved, 9-80 years old, median 57 years old, containing 4 CDD, 30 RYHJ, 54 Whipple and 1 bile duct ileocecal anastomosis patients; The time between ERCP and surgery were 30 (1-40), 2.75 (0.5-14), 2 (0.3-19), and 10 years, respectively; The time between surgery and symptom were 240 (3-360), 12 (1-156), 22 (0-216), and 60 months, respectively. Fifty percent of CDD could succeed only under local anaesthesia, RYHJ (96.7%) and Whipple (100.0%) needed under general anaesthesia (P < 0.001). Successful first entry rates of CDD, RYHJ and Whipple were 100.0%, 40.0% and 77.8%, respectively. After changing the endoscopy type, successful entry rate could increase to 43.3% of RYHJ and 83.3% of Whipple. The successful entry rate of different anastomotic methods was significant (P < 0.001). The cannulation success rates of CDD, RYHJ and Whipple were 100.0%, 53.8% and 86.7% respectively, with significant difference between the groups (P=0.031). ERCP success rates of CDD, RYHJ and Whipple were 100.0%, 33.3% and 78.8% respectively, with significant difference between the groups (P < 0.001). Complications were found in 23.9% (21/88) patients, including infection (14.8%), pancreatitis (9.2%), bleeding (3.4%), and perforation (2.3%) ranked by incidence. Causes of ERCP in post bilioenteric anstomosis were anastomotic stenosis (50.0%, benign 39.3%, malignant 10.7%), choledocholithiasis (37.5%) and reflux cholangitis (12.5%). Anastomotic method was the only predicting factor of ERCP success in patients after bilioenteric anstomosis (OR=7, 95%CI: 2.591-18.912, P < 0.001). Conclusion: ERCP in post bilioenteric anstomosis patients with gastrointestinal reconstruction need general anaesthe-sia, with good safety and efficiency. The successful rate of RYHJ was significantly lower than Whipple. Anastomotic method was the only predicting factor of ERCP success.