With the development of endoscopic ultrasound (EUS) examination and magnetic resonance cholangiopancreatography (MRCP), ERCP has evolved into a therapeutic tool gradually. However, ERCP is challenging in patients with surgically altered anatomy. No standardized guidelines have been provided to choose the safest and most effective approach. Therefore, we retrospectively analyzed the outcomes, adverse events and related risk factors of 121 cases in 93 patients with surgically altered anatomy in our center.

ERCP procedure faces 3 challenges for these patients: reaching the target site, selective biliary cannulation, and performing diagnostic or therapeutic interventions [8]. The success rate of ERCP in patients with a reconstructive gastrointestinal tract has been reported to be 63–95% [30]. In our study, success rate of reaching the papilla or surgical anastomosis was 93.4%. Successful selective cannulation and cholangiography were achieved in 93.8% of them, which is similar to the rate of the same procedure using double-balloon enteroscopes (94%) [5]. In general, success rate of ERCP procedure in patients with surgically altered anatomy in our center was superior.

Success rates have increased with the introduction of device-assisted enteroscopes (DAEs), such as double-balloon enteroscopes, single-balloon enteroscopes and spiral enteroscopes. A previous review has reported an overall 74% ERCP success using DAEs [14]. A retrospective study by Shah et al. has shown a 63% ERCP success with DAEs [12]. Inamdar et al. reported a pooled technical, diagnostic, and procedural success rates of 80.9%, 69.4%, and 61.7% respectively with single-balloon enteroscopes [31]. Corresponding success rates were reported to be 86%, 100%, and 86% by Ali et al. with spiral enteroscopes in Roux-en-Y anatomy [20]. Newly designed features in DAEs, such as high force transmission and passive bending, made it possible to increase the success rates of ERCP procedures in these patients further [32, 33]. However, DAEs are not designed for this purpose and have not been widely available because of the need for specialized equipment and expertise. In our study, standard enteroscopes were proved effective and safe to perform in patients with surgically altered anatomy, which brought better applicability.

Despite emerging enteroscopes applied in patients with surgically altered anatomy, ERCP-related success rates are still less satisfactory than in patients with normal gastrointestinal anatomy. The procedure is considered challenging for several reasons. First, an increase of the intestine length, and sharp angulation of the bowel make it difficult to identify the afferent loop and reach the target site [10]. Existing enteroscopes tend to be too short to approach the papilla or anastomosis. Second, selective cannulation is difficult due to the unfavorable orientation of the papilla, along with limited availability of accessories, and lack of an elevator [34]. Third, adhesions and strictures left by reconstruction also impede the process. Techniques such as manual compression method, positional change, and use of a transparent cap have been proposed to increase the success rates [35].

Since no standardized procedure has been applied in these patients, a thorough understanding of postoperative anatomy and multiple training seem extremely important. Relative knowledge of the reconstruction method, biliary anastomosis, the lengths of the limbs, and the presence of adhesions and strictures is indispensable. In addition, proper selection of endoscopes is likely to determine the outcome of the ERCP procedure. Selection of endoscopes was mainly depended on patients’ anatomy of the gastrointestinal tract, such as lengths of afferent and efferent loops, degrees of angulations, and endoscopists’ operating experience. In our study, technical success was highest in patients with Billroth-II gastrectomy, while lowest in patients with Roux-en-Y gastrectomy. However, ERCP in patients with Roux-en-Y reconstruction had perfect diagnostic success and therapeutic success. Longer afferent limbs, more severe adhesion in Roux-en-Y reconstruction make it particularly difficult to reach the blind end. Therefore, important drawbacks appeared when using a conventional side-view duodenoscope in these patients. A short-type single-balloon enteroscope may be useful to address this issue. With a passive bending part, it contributed to the advancement of the scope. The balloon also helps to hold and fix the intestine, making it possible to insert deeply. Considering an increase of Roux-en-Y reconstruction due to the growing need for laparoscopic surgeries [36], special attention is required to this group. Conventional duodenoscopes and gastroscopes are recommended to patients with Billroth-I or Billroth-II gastrectomy, owing to the relatively short afferent limbs. As for Billroth-II reconstruction, it is under debate whether to choose a forward-view or a side-view enteroscope [11]. In our study, several patients with Billroth-II reconstruction changed the enteroscope midway, from a gastroscope to a duodenoscope. A forward-view gastroscope can provide better visualization, making it safer to reach the papilla. Then with a side-view duodenoscope, cannulation becomes easier due to the appropriate view to the papilla. Similar results have been raised by Park et al. The forward-view endoscope was associated with a higher afferent loop intubation rate, while the side-view endoscope with a higher selective cannulation rate [37]. Enteroscope exchange strategy is expected to be more effective for these patients.

In our present study, ERCP-related adverse events occurred in 25.6% of patients. Previous studies reported the complication rate from 3.5 to 12.4% [9, 11, 12, 16, 18, 38–40]. However, hyperamylasemia was not included in most of these studies. A comparable 23.0% ERCP-related complication rate was reported combined with hyperamylasemia [35]. Post-ERCP pancreatitis (PEP) is recognized as the most common procedure-related adverse event in conventional ERCP, while perforation in balloon-assisted ERCP [7]. A systematic review based on randomized controlled trials reported an overall 9.7% rate of PEP [41]. Other complications include bleeding, cholangitis, mucosal laceration and cardiopulmonary related diseases. Recently, the advent of DAEs has allowed lower rates of ERCP-related adverse events. It was reported in patients with Roux-en-Y reconstruction using DAEs with complication risks ranging from 0 to 19.5% [42]. A meta-analysis concluded an overall 6.5% complication rate in 461 patients using single-balloon enteroscopes [31]. Anvari et al. reported a 4% complication rate in 1523 patients with double-balloon enteroscopes [5]. However, DAEs are more technically demanded and bring a higher risk of complications. Conventional ERCP allows more practical use. In our center, no lethal complication occurred, with most adverse events treated conservatively. Conventional enteroscopes were equally effective and safe.

Our study showed that multiple cannulation attempts, endoscopic papillary balloon dilation, and biopsy in the bile duct or papilla were independent risk factors of ERCP-related adverse events. Multiple cannulation attempts mean prolonged procedure time and usually follow a pre-cut sphincterotomy, which increases the risk of PEP and bleeding. An increased rate of PEP has been seen with additional cannulation attempts in a prospective study [43]. In our study, higher rates of PEP and hyperamylasemia were likely to be associated with EPBD, which concurred with the study reported by Park et al. [35]. Previous studies have reported that compared with endoscopic sphincterotomy (EST) or endoscopic sphincterotomy with balloon dilatation (ESBD), EPBD presented higher rates of pancreatitis, while bleeding was more common in the former factors [44, 45]. The reason may due to mucosal edema of the papillary caused by procedures, leading to pancreatic outflow obstruction. In addition, biopsy in the bile duct or papilla can also increase the complication rate.

To our knowledge, few studies have been conducted to identify the risk factors of ERCP-related adverse events including multiple types of reconstruction methods and endoscopes. Compared with previous studies, our present study had a relatively larger number of patients, a more comprehensive analysis of all the reconstruction methods and endoscopes, making it more practical and applicable. However, several limitations did exist, including its retrospective design based on a single-center experience, lack of follow-up data and a control group, variations of patients, reconstruction methods and endoscopes. In the future, multi-center prospective studies are needed to validate present findings. Establishment of standardized practical guidelines and training programs are indispensable to guide ERCP procedures in patients with surgically altered anatomy.

In conclusion, ERCP procedure was feasible and safe in patients with surgically altered anatomy. For early identification of the occurrence of ERCP-related adverse events, close vigil should be kept on patients who have undergone multiple cannulation attempts, endoscopic papillary balloon dilation, and biopsy in the bile duct or papilla.