To our knowledge, this is the first study to carry out the concept “unplanned repeated ERCP.” In this retrospective cohort, the majority of unplanned repeated ERCP for CBDS patients were postoperative unsatisfactory serum bilirubin reduction, with symptoms of cholangitis or not. Irregular papilla, post-ERCP fever and abdominal pain, post-/pre-ERCP serum TBil levels ⩾1.2 times, and post-/pre-ERCP serum ALT levels ⩾1.2 times were identified as independent risk factors for unplanned repeated ERCP. CBDS patients with unplanned repeated ERCP may exhibit a higher propensity for CBDS recurrence and biliary infection in the long term, without statistical significance.

In our study, the main reasons for unplanned repeated ERCP in patients with CBDS were unrelieved biliary obstruction or cholangitis. The threshold of interval time between the initial and unplanned repeated ERCP procedure was determined based on multiple factors, including the natural course of CBDS, the incidence of CBDS-related or procedure-related complications, the real experience of CBDS patients’ hospitalization in our center, and previous reports of unplanned reoperations. Recurrent biliary stone is one of the most common complications after stone removal, but there has not been an accurate definition for the recurrence time of CBDS. In recent clinical studies, 3 months was often selected as the criterion of the time interval between initial stone clearance and biliary stone recurrence.^9,10 So, to distinguish unplanned repeated ERCP demand from conventional complications, the time interval between the initial and repeated ERCP should be determined to be less than 3 months. In addition, sometimes temporary insertion of biliary plastic stents was performed in patients with CBDS for better biliary drainage. The usual replacement interval of biliary plastic stents is 3 months, varying depending on the etiology and expertise of endoscopists. ¹¹ Kubesch et al. ¹² previously reported that the median plastic stent patency was 41 days for scheduled stent exchange and was significantly shorter (median 34 days) in cases with one plastic stent. Moreover, in surgery-related research, unplanned re-operation within 24 h, 48 h, or 30 days following the major surgery has been reported as the outcome of interest.^{13bibr14-17562848251342342bibr15-17562848251342342–16} Therefore, it is reasonable and clinically valuable to analyze unplanned repeated ERCP within 1 month in our study.

In our cohort, the incidence of unplanned repeated ERCP within 1 month in CBDS patients was about 2.95%, which is comparable with those previously reported in low-risk surgical fields. Lim et al. ¹⁷ found that the unplanned reoperation rate was 1.9% in 6749 cases of outpatient plastic surgery. And unplanned reoperation rate was about 4% in inpatient spinal surgery ¹⁸ and up to 14.2% in major head and neck surgery. ¹³ These suggest that ERCP procedures are relatively minimally invasive and safe therapeutic options with controllable risks.

In our analysis of risk factors for repeated ERCP, the irregular papilla, a dilated CBD, CBDS size, and location in the superior CBD were significantly associated with increased risks of unplanned repeated ERCP within 1 month. Papilla morphology can affect the difficulty of selective biliary cannulation, which is the milestone of the ERCP procedure. The irregular papilla, including small, protruding or pendulous, creased or ridged papilla, increased biliary cannulation difficulty, and post-ERCP infections or pancreatitis,^19,20 which may require repeated ERCP intervention. Then, CBD dilatation ⩾10 mm and intrahepatic CBDS were independent risk factors for CBDS recurrence, as previous studies reported.^{21 bibr22-17562848251342342–23} Common causes of CBD dilatation include cholelithiasis, tumor in the ampullary region, congenital factor, history of cholecystectomy, etc. ²⁴ As benign or malignant biliary stricture was excluded in this study, CBD dilatation was mainly related to the stones and partially represented the severity of biliary obstruction. After ERCP, the dilated CBD appears to have increased enterobiliary reflux and infection. ¹⁰ In addition, unintended guidewire passage into the pancreatic duct was always identified as a risk factor of post-ERCP pancreatitis, ²⁵ but did not increase the risk of repeated ERCP within 1 month. Interestingly, the basket was used more for stone extraction in the repeated ERCP group versus the unrepeated group. However, several studies indicated that there was no significant difference in stone clearance and complications between the basket and balloon catheters.^26,27 We speculate this discrepancy may be affected by the size of CBDSs and endoscopists’ experience. In multivariate regression analysis, post-ERCP fever, abdominal pain, and elevated serum TBil or ALT over 1.2 times that of pre-ERCP were independent predictors for unplanned repeated ERCP. Tierney et al. ²⁸ reported that patients developing post-ERCP cholangitis were more likely to have higher baseline TBil and ALT, which was not observed in our study. Compared with pre-ERCP serum liver function parameters, post-ERCP serum TBil and ALT levels were reduced less or even elevated in the repeated ERCP group than those in the unrepeated ERCP group. To better quantify the risk and improve utility, we categorized liver function indicators and found that if the post-ERCP TBil level decreased by less than 20% of the pre-ERCP TBil level, it is necessary to be cautious about repeated ERCP with a sensitivity of over 80%, while if the post-ERCP TBil level increased by more than 20% of the pre-ERCP TBil, this will be a strong indicator for repeated ERCP demand with the specificity of over 80%.

The total incidence of adverse events (bleeding: 0.9%, CBDS recurrence: 6.0%, biliary infection: 5.2%) in our study was comparable with previously reported, in which the general incidence rates of ERCP-related bleeding, CBDS recurrence, acute cholangitis, and cholecystitis were 0.3%–9.6%, 4%–25%, 0.5%–3.0%, and 0.5%–5.2%, respectively.^23,29 Generally, multiple ERCP sessions may increase the risks of procedure-related complications (bleeding, perforation, infection, pancreatitis). However, in our study, no significant increase in the risk of post-procedural bleeding or biliary infection was found in the repeated ERCP group, and no perforation or post-ERCP pancreatitis occurred. In terms of the impact of unplanned repeated ERCP on clinical outcomes, our data revealed that unplanned repeated ERCP, although not an independent factor, is partially relevant with increased likelihood of long-term adverse events (CBDS recurrence, acute cholangitis, or cholecystitis), especially in the patients who had gallbladder without stones or underwent cholecystectomy. We also investigated that cholecystectomy or cholecystolithiasis did not affect adverse outcomes (e.g., CBDS recurrence) in the long term. We predict the possible reason is: The common cause of CBDS is stone passage from the gallbladder. If this condition is excluded, other multiple factors may play a greater role in recurrent stone or biliary infection, including ERCP-related impairment of the biliary tract, anatomy, motility, or metabolic disorders.^{10,23,30 bibr31-17562848251342342–32} Notably, besides unplanned repeated ERCP, biliary stent placement presented a significant association with poor prognosis in patients with a resected or stone-free gallbladder. A meta-analysis identified biliary stent placement as one of the strong risk factors of recurrent CBDS following ERCP, ³⁰ which is partially consistent with our results. Long-term stent placement and CBD dilation may promote bacterial proliferation and stent-stone complex formation.^23,33 But for patients with cholecystolithiasis, it is the morphology of CBD or CBDS that plays the predominant role in long-term adverse events, specifically irregular papilla, CBD dilation, or CBDS diameter ⩾0.65 cm. These could provide the gastroenterologists with different surveillance strategies based on individual gallbladder status.

Our findings provide novel and valuable insights into perioperative management and quality control of ERCP procedures. Preoperative risk identification and intraoperative planning are key factors in preventing unplanned repeated ERCP. Postoperative surveillance, including abdominal symptoms or fever, unsatisfactory TBil reduction, will facilitate early recognition and interventions for repeated ERCP. In addition, like many surgical fields, the rate of unplanned repeated ERCP is also an outcome measure to determine a center’s ability to provide safe and effective care and cannot be understated. The relatively low reoperation rate makes this metric difficult to study on only one institution, urgently demanding a multi-center or national database establishment.

This study is the first to assess unplanned repeated operations in ERCP procedures. We found multiple factors associated with repeated ERCP risk based on clinical characteristics, ERCP technique parameters, and laboratory tests. These will help endoscopists to counsel patients on reoperation risk reduction and individualize surveillance in the perioperative period. Then, we performed follow-up over 3 years to evaluate the impact of unplanned repeated ERCP on clinical prognosis. There are also some limitations in our study. First, this is a single-center, retrospective, observational study. Despite the follow-up via telephone or hospital visits, CBDS recurrence may be underestimated to some degree, as cholelithiasis is sometimes asymptomatic, so that they usually do not perform radiographic examination routinely. The results may be more convincing if further multicenter, prospective studies are conducted. Second, the limited sample size in the repeated ERCP group may partially cause deviations in results or miss some potentially valuable risk factors. During subgroup analysis, the decrease in the sample size per subgroup may also reduce statistical power. Larger sample size studies are required to confirm these findings. Third, we cannot retrospectively identify the component of stones, which may affect the pattern of stone recurrence and disease progression.