記住我
Abstract
Background: Endoscopic retrograde cholangiopancreatography (ERCP) for patients with periampullary diverticulum (PAD) remains a challenge. This study aims to investigate the factors and techniques related to successful and safe ERCP in patients with PAD.
Methods: We enrolled patients who underwent ERCP in a large tertiary center. The difficult cannulation rate, technical success rate, clinical success rate, and adverse events (AEs) rate were compared between patients with or without PAD. Three independent logistic regression models were established to identify factors and techniques associated with difficult cannulation, clinical success, and AEs.
Results: Five thousand five hundred and ninety patients were included, of which 705 (12.6%) were diagnosed with PAD. Patients with PAD had a significantly higher difficult cannulation rate compared with patients without PAD (10.6% vs 8.0%, P < 0.0001), but the rates of technical success (clinical success (95.2% vs 95.2%, P = 0.951), and AEs (16.5% vs 14.4%, P = 0.156) were similar. Type I PAD (odds ratio [OR] = 2.114, 95% confidence interval [CI]:1.05-5.25) and ERCP indication for pancreatic diseases (OR = 1.196, 95%CI: 1.053-1.261) were independently associated with difficult cannulation. Small endoscopic sphincterotomy (EST) with balloon dilatation (OR = 1.581, 95%CI: 1.044-2.393) was independently associated with clinical success. Somatostatin injection showed no preventive effect on post-ERCP pancreatitis (OR = 1.144, 95%CI: 1.044-1.254). Moreover, the auxiliary cannulation techniques were safe for PAD patients.
Conclusions: PAD did not affect ERCP outcomes. However, the choice of techniques and AE prophylactic measures should be more specific, especially for patients with type I PAD.
Keywords: Adverse events (AEs), clinical success rate, difficult cannulation, endoscopic retrograde cholangiopancreatography (ERCP), periampullary diverticulum (PAD)
How to cite this article:Authors Chuanchao Xia, Liqi Sun and Lisi Peng have contributed equally to this work and share first authorship.
Authors Zhendong Jin and Haojie Huang share senior authorship
See accompanying Editorial
Introduction 
Endoscopic retrograde cholangiopancreatography (ERCP) is a common endoscopic procedure, which is widely used in the diagnosis and treatment of cholangiopancreatic diseases. However, ERCP is a technical challenge and is associated with serious adverse events (AEs). Possible ERCP-related AEs (AEs) include acute pancreatitis, hemorrhage, perforation, cholangitis, and acute cholecystitis. Among them, post-ERCP pancreatitis (PEP) is the most common, with an incidence of 9.7% and a mortality rate of 0.7%.[1] Many high-risk factors have been reported to be associated with these AEs, including female gender, papilla anatomic variation, difficult cannulation, pancreatic deep wire pass, needle-knife fistulotomy, Oddi sphincter dysfunction, and biliary sphincterotomy, etc.[2] To prevent post-ERCP AEs, many prophylactic measures that have a significant effect have been described. These measures mainly include rectal administration of indomethacin, pancreatic stent placement, and intravenous fluid hydration.[3] But some precautions remain controversial, such as the use of somatostatin. Therefore, the development of post-ERCP AEs can be the result of multiple factors, related either to the patient or the technique.
Periampullary diverticulum (PAD) is a duodenal outpouching arising from the ampulla of Vater. PAD is a common clinical situation and the clinical discovery rates vary from 6% to 31.7%.[4],[5] The presence of PAD may cause technical problems during ERCP and AEs after ERCP. However, the influence of PAD on technical success rate and AEs rate is contradictory.[6],[7] Moreover, there is no consensus about the preferred techniques for PAD patients to receive ERCP.[8] Although the aforementioned prophylactic measures have a certain effect on the prevention of AEs after ERCP in patients without PAD, the impact of prophylactic measures on AEs rate in PAD patients is not clear due to the changes in anatomy and technical application.
Here, we conducted a large sample retrospective study to explore the impact of PAD on ERCP outcomes. The effects of varied indications, ERCP techniques, and prophylactic measures on ERCP outcomes in PAD patients were also evaluated.
Methods Patients
From January 2017 to December 2020, the patients who underwent ERCP in the department of gastroenterology, Changhai Hospital, were analyzed. The study was approved by the Institutional Review Board of Changhai Hospital (Shanghai Changhai Hospital Ethics Committee). Patients who received ERCP in our center were enrolled according to the inclusion criteria, and were divided into those with and without PAD. PAD was observed near the papilla through a duodenoscope. The exclusion criteria were as follows: (a) patients with papilla-related variables other than PAD (minor papilla, long axis papilla, etc.); (b) repeated ERCPs with prior endoscopic sphincterotomy (EST); (c) surgically altered anatomy, such as Billroth I, II and Roux-en-Y; (d) papillary fistula or stone impaction; (e) papillary carcinoma or adenoma; (f) ERCP performed by trainees.
ERCP procedure
All the ERCP procedures included in this study were performed by endoscopists in our center, using standard side-view duodenoscope (JF-260, TJF-240, or TJF-260; Olympus, Tokyo, Japan) in the left lateral position, after fasting overnight. All the endoscopists were experienced and perform more than 150 ERCP operations per year. The patients were sedated and vital signs were routinely monitored during ERCP procedures.
After the duodenoscope reached the papilla and examined carefully, selective cannulation was performed using the guidewire-assisted technique initially. In most cases, EST was routinely performed. In some cholangitis patients with poor general condition, only biliary stent placement or nasobiliary placement was performed under the guidance of guidewire, and in some cases with small bile duct stones (<8 mm), only balloon dilation (EPBD) was performed as an alternative to EST. There were two types of EST techniques, large EST and small EST (the scope of papilla cutting ≤ 5 mm). When cannulation was difficult (prolonged cannulation time and repetitive unintended guidewire insertion into the pancreatic duct), either precut biliary sphincterotomy, needle-knife fistulotomy or trans-pancreatic sphincterotomy were performed to facilitate cannulation according to the expertize of the endoscopists. These techniques are defined as auxiliary cannulation techniques in our study.
Amylase and blood routine examinations were obtained within a short time (<3 h) after ERCP, and one day after the procedures, to determine the occurrence of PEP or infection. Intravenous injection of somatostatin (0.25 mg/h) lasted for 1-3 days in patients with elevated amylase level and abdominal pain after ERCP.
Definition of parameters
PAD was divided into three types according to the position of the major duodenal papilla: type I, inside the diverticulum; type II, in the margin of the diverticulum; and type III, near the diverticulum [Figure 1].[9] Difficult cannulation was defined as follows: (a) failure of selective bile duct cannulation with standard ERCP technique within 10 min; (b) more than 5 cannulation attempts; or (c) failure to gain access to the major papilla.[10] Technical success was defined as a successful selective cannulation. Clinical success was defined as achieving certain clinical aim (e.g., stone removal, biliary/pancreatic duct drainage) after successful selective cannulation. Difficult stones were defined as large diameter (>1.5 cm), massive, or unusually shaped (barrel-shaped) stones. AEs were divided into two grades: mild AEs and severe AEs. For mild AEs, no more intervention was required. Only fasting, clinical observation or drug treatment were needed. For severe AEs, further endoscopic or surgical interventions and critical care were needed. According to the above principles, solely hyperamylasemia was included in the mild pancreatitis group.
Figure 1: Three different types of periampullary diverticulum. (a) Type I. inside the diverticulum; (b) Type II, in the margin of the diverticulum; and (c) Type III, near the diverticulum. (Red arrow: diverticulum; Yellow arrow: papilla)The indications of ERCP were as follows: (a) bile duct stone with or without mild cholangitis; (b) severe cholangitis or pancreatitis requiring emergency ERCP; (c) malignant biliary strictures; (d) benign biliary strictures; (e) indeterminate biliary strictures; (f) pancreatic diseases requiring pancreatic duct drainage (pancreaticobiliary maljunction, pancreas divisum, chronic pancreatitis, and intraductal papillary mucinous neoplasm).
The procedure techniques utilized in ERCP after cannulation were as follows: (a) single wire-guidance; (b) full EST with balloon dilatation; (c) full EST without balloon dilatation; (d) small EST with balloon dilatation; (e) small EST without balloon dilatation; (f) single balloon dilatation. The auxiliary techniques to facilitate cannulation included needle-knife fistulotomy, trans-pancreatic sphincterotomy and precut biliary sphincterotomy.
The prophylactic measures were divided into the following: (a) without any measures; (b) pancreatic stent alone; (c) rectal administration of indomethacin alone; (d) somatostatin injection alone; (e) (b)+(c); (f) (c)+(d);(g) (b)+(d); (h) (b)+(c)+(d).
Statistical analysis
Mean ± standard deviation (SD) and median (Interquartile range (IQR)) were used to describe the deviation of the data. Frequency (percentage) was used to describe the classification of variables. The differences between groups were compared by student's t-test, Chi-square test and Fisher's exact test. Variables of potential significance (p < 0.10) were entered into the logistic regression model to identify independent predictors of difficult cannulation, clinical success, and AEs. All models were adjusted for age and sex, and the results were presented as odds ratio (OR) with 95% confidence interval (CI). P values less than 0.05 were considered to indicate significance. Data were analyzed with IBM SPSS Statistics version 22.0 (IBM Corp., Armonk, NY, USA).
Results Patients' characteristics
From January 2017 to December 2020, a total of 5590 patients were included in this study. Of the 5590 patients, 705 (12.6%) were confirmed to have PADs. The detailed patient selection flowchart is shown in [Figure 2]. Of the 705 patients, 59 (8.4%) had type I PAD, 101 (14.3%) had type II PAD, and 545 (77.3%) had type III PAD. A single diverticulum was found in 649 (92.1%) patients with PAD, and 56 (7.9%) patients had two or more diverticula.
In the overall study cohort of 5590 patients, 2127 (38.1%) patients were females and 3463 (61.9%) patients were males. The median age was 57 years (45-69).
The impact of PAD on ERCP outcomes
The differences in variables between PAD group and non-PAD group were compared and are presented in [Table 1]. The gender distribution was not significantly different in the two groups, with a male percentage of 60% vs 62.2% (P = 0.254). However, patients with PAD tended to be older than patients without PAD (median, 68 years vs 56 years, P < 0.0001).
According to the criteria of difficult cannulation, the difficult cannulation rate was significantly higher in the PAD group than in the non-PAD group (10.6% vs 8.0%, P < 0.0001). However, different difficult cannulation rates did not affect the ERCP outcomes. Both groups had high technical (95.6% vs 95.2%) and clinical success rates (95.2% vs 95.2%), and were not significantly different between the two groups (P = 0.648 and 0.951).
The overall AEs rate was 14.7% and the majority AEs were mild pancreatitis (14.1%). In PAD group, the AEs rate was 16.5%, which was numerically higher than that in the non-PAD group (14.4%), but there was no significant difference (P = 0.156). Further analysis of AE subtypes showed that there was no difference between mild pancreatitis (15.9% vs 13.8%, P = 0.136), severe pancreatitis (0.3% vs 0.3%, P = 1.00), perforation (0.1% vs 0.04%, P = 0.33), bleeding (0% vs 0.2%, P = 0.61), and cholangitis (0.1% vs 0.2%, P = 1.00). Thirty one patients (0.55%) developed severe AEs (15 severe pancreatitis, 3 perforations, 7 bleedings and 6 cholangitis) and three patients (0.43%) developed severe AEs (2 severe pancreatitis and 1 perforation) in the PAD group. There was no significant difference in the severe AEs rate between the two groups (P = 0.824).
Factors associated with difficult cannulation in patients with PAD
A total of 75 (10.6%) PAD patients faced difficult cannulation during ERCP. A logistic regression model was established to identify factors associated with difficult cannulation. Age, gender, types of PAD, indication, and difficult stones were included in univariate analysis. The differences in parameters between the difficult cannulation group and the smooth cannulation group were compared and are presented in [Table 2].
Table 2: Univariate and multivariable analysis of risk factors for difficult cannulationBased on univariate analysis, type I PAD (χ2 = 6.374, P = 0.012), indications for indeterminate biliary stricture (χ2 = 3.269, P = 0.07), and pancreatic diseases (χ2 = 25.546, P < 0.0001) were closely correlated with difficult cannulation. Whereas, type III PAD (χ2 = 5.687, P = 0.02) and indications for bile duct stones (χ2 = 34.136, P < 0.0001) were closely correlated with smooth cannulation. Moreover, patients with difficult stones (χ2 = 7.746, P = 0.005) were correlated with smooth cannulation. These factors were included in the multivariate logistic regression analysis.
The multivariate logistic regression analysis results are shown in [Table 2]. Type I PAD (OR = 2.114, 95%CI: 1.05-5.25) and indications for pancreatic diseases (OR = 1.196, 95%CI: 1.053-1.261) were independently associated with difficult cannulation, after age and gender were adjusted. On the other hand the indication for bile duct stones was independently associated with smooth cannulation (OR = 0.279, 95%CI: 0.13-0.598).
Preferred factors and techniques facilitating clinical success
Of the 705 PAD patients, 34 (4.8%) experienced clinical failure. Age, gender, types of PAD, and techniques were included in the univariate analysis. Type I PAD (χ2 = 4.015, P = 0.045) and large EST without balloon dilatation (χ2 = 372.307, P < 0.0001) were correlated with clinical failure. Small EST with (χ2 = 19.787, P < 0.0001) or without (χ2 = 10.783, P = 0.0001) balloon dilatations were correlated with clinical success [Table 3]. After age and gender were adjusted, the multivariate logistic regression analysis showed that large EST without balloon dilatation (OR = 0.225, 95%CI: 0.147-0.343) was independently associated with clinical failure. Small EST with balloon dilatation (OR = 1.581, 95%CI: 1.044-2.393) was independently associated with clinical success [Table 3].
Table 3: Univariate and multivariable analysis of factors for clinical successFactors associated with adverse events in patients with PAD
Of the 705 patients with PAD, 116 (16.5%) developed AEs (112 mild pancreatitis, 2 severe pancreatitis, 1 perforation, and 1 cholangitis) after ERCPs. Three patients had severe AEs (2 severe pancreatitis and 1 perforation). Age, gender, types of PAD, indications, techniques, and preventive measures were included in univariate analysis [Table 4].
Table 4: Univariate and multivariable analysis of risk factors for adverse eventsBased on univariate analysis, older age (t = 4.213, P = 0.04), male sex (χ2 = 4.650, P = 0.031), type I PAD (χ2 = 5.327, P = 0.021), difficult stones (χ2 = 4.109, P = 0.043), protection of somatostatin injection only (χ2 = 10.184, P = 0.001), and rectal indomethacin + somatostatin injection (χ2 = 6.915, P = 0.009) were factors correlated with AEs. Whereas, the utilization of balloon dilatation only (χ2 = 5.451, P = 0.02) was correlated with the absence of AEs. These factors were included in multivariate logistic regression analysis, except for the “somatostatin injection only” factor (due to the small number of cases included).
Based on multivariate logistic regression analysis, older age (OR = 1.019, 95%CI: 1.003-1.035), male sex (OR = 1.687, 95%CI: 1.085-2.622), type I PAD (OR = 1.979, 95%CI: 1.061-3.691), difficult stones (OR = 1.743, 95%CI: 1.019-2.982), and utilization of rectal indomethacin+ somatostatin injection as preventive measure (OR = 1.144, 95%CI: 1.044-1.254) were independently associated with AEs. Utilization of balloon dilatation only (OR = 0.78, 95%CI: 0.613-0.993) was an independent protective factor for AEs. We confirmed that the use of rectal indomethacin alone as a prophylactic measure was not associated with AEs (P = 0.33), while somatostatin injection alone (P = 0.001) and rectal indomethacin + somatostatin injection (P = 0.009) were risk factors of AEs.
In our study, only three included PAD patients showed severe AEs. It was impossible to build a multivariate logistic regression model. Therefore, the analysis of factors associated with severe AEs was not included in our study.
The impact of auxiliary cannulation techniques on ERCP outcomes
It is worth mentioning that in order to facilitate canulation, needle-knife fistulotomy, trans-pancreatic biliary sphincterotomy, and precut techniques were used in our study cohort when difficult canulations occurred. In total, 21 PAD patients underwent needle-knife fistulotomy, 13 patients underwent trans-pancreatic biliary sphincterotomy, and 28 patients underwent precut biliary sphincterotomy. Successful cannulation was achieved in 20 out of 21 (95.2%), 11 of 13 (84.6%), and 26 of 28 (92.9%) aforementioned PAD patients, respectively. Moreover, these techniques were not correlated with the development of AEs in our study cohort (P = 0.128, P = 0.785, and P = 0.838, respectively).
Discussion PAD is a common clinical anatomic abnormality, with clinical detection rates ranging from 6% to 31.7%.[4],[5],[11] PAD is associated with a higher incidence of biliary stones, cholangitis, and biliary pancreatitis.[12] The first choice for the treatment of these diseases is ERCP. Therefore, the way to deal with PAD patients is important for endoscopists.
Numerous previous studies have reported the influence of PAD on technical success and AEs of ERCP. However, the results varied. A meta-analysis by Jayaraj et al.[6] concluded that the success rate of ERCP in patients with PAD was lower than that in patients without PAD, and the AEs rate was not significantly different between the two groups. A more recent meta-analysis by Mu et al.[13] concluded that the presence of PAD did not increase the overall cannulation failure rate post-year 2000. PAD also affected the occurrence of early pancreatitis, perforation, and bleeding. These two meta-analyses reached the opposite conclusions about the influence of PAD. Moreover, the subgroup analysis of different PAD types was not included in the two meta-analyses. Therefore, it is still necessary to discuss the influence of different types of PAD on ERCP outcomes. In our study, we found that the presence of PAD was associated with difficult cannulation, especially in patients with type I PAD. However, the increase of cannulation difficulty did not affect the technical success rate and clinical success rate.
This result may be due to the utilization of innovative cannulation techniques. Several novel instruments and techniques had been described to facilitate cannulation, including endoclips,[4] endobiliary forceps,[14] biopsy forceps,[15] and cap-assisted forward viewing endoscopy[16] etc., However, these novel instruments and techniques were only described in case reports. The preferred techniques were still unknown. In our center, the most commonly used auxiliary techniques were precut biliary sphincterotomy, including needle-knife fistulotomy and trans-pancreaticbiliary sphincterotomy. Precut biliary sphincterotomy was considered an independent risk factor for PEP[17] and hemorrhage.[18] However, recent studies have shown that if performed by experienced endoscopists, precut biliary sphincterotomy is safe, with a high successful cannulation rate even in high-risk individuals.[19] The data on the impact of auxiliary cannulation techniques in PAD patients was sparse. In 1998, Fogel et al.[20] utilized precut biliary sphincterotomy to cannulate papilla with PAD. Successful cannulation was achieved in 7 out of 8 patients (93.9%) with PAD. The AEs rate was as high as 25%. Park et al.[21] reported the efficacy and safety of needle-knife fistulotomy in patients with PAD. Successful cannulation rate was 93.9% and AEs rate was comparable to that of the patients without PAD. In our study, we identified that the auxiliary cannulation techniques were effective and safe to assist cannulation for PAD patients. Despite the small sample size, we still believe that the precut techniques are suitable options for cannulation in PAD patients, especially in patients with type I PAD.
The goal of ERCP is to achieve clinical success. The procedure techniques and PAD types had a great impact on the clinical success rate. Traditionally, EST and balloon dilation were routine techniques that contributed to clinical success. However, preference of the technique remained unknown. In 2004, Tham et al.[22] concluded that EST for PAD patients was equally effective and safe as for patients without PAD. Chen et al.[23] and Xu et al.[24] reported that EST was less effective than balloon dilation and small EST plus balloon dilation in patients with PAD, but were equally safe. A multicenter prospective study by Zulli et al.[25] concluded that large balloon dilation was a safe and effective technique for patients with PAD. However, no comparative group was included in this study. In our study, small EST with balloon dilation was more effective than large EST without balloon dilation for PAD patients. The possible reason was that large EST was technically difficult because it was often used for complex biliary tract conditions. Moreover, balloon dilatation alone was an independent protective factor for AEs. In patients with type I PAD, small EST was also identified to be a safe procedural technique. Therefore, we believe that small EST and balloon dilatation are the preferred techniques to perform ERCP in patients with PAD.
The effect of PAD subtypes on ERCP outcomes was rarely reported. Yue et al.[5] identified that type I PAD had a low rate of successful cannulation. The influence of PAD on AEs was not discussed in this study. In our study, we also found that type I PAD was associated with difficult cannulation, lower clinical success (not independently) and AEs. The result may be due to the blockage of papillary openings by the diverticulum wall in type I PAD. To cannulate the papilla, continuous air inflation is required to fill the diverticulum cavity and expose the opening of the papilla.[5] As a result, the endoscopists tend to be conservative due to the thin duodenal wall inflated by air. Therefore, experienced endoscopists are the preferred ones to perform ERCP for patients with type I PAD.
Interestingly, we found that bile duct stones and difficult stones (not independently) were associated with higher successful cannulation rate. This may be due to the fact that expanded bile duct was easier to be cannulated. The conclusion was contrary to a study conducted by Tabak et al.[12] in a smaller study cohort. We are the first to identify that ERCP indication for difficult stones is an independent risk factor for AEs. These results should be verified in further studies.
The factors associated with the development of AEs are complicated. PEP is the most common AE and needs close attention. Many prophylactic measures have been described to prevent PEP. Rectal indomethacin, pancreatic stent, intravenous fluid hydration, and somatostatin injection are the most commonly used prophylactic measures.[26] The preventive effect of these measures on PAD patients were rarely evaluated. In our center, rectal indomethacin is routinely used in most patients before ERCP. Intravenous fluid hydration is not performed. During ERCP, if a PAD is identified, we tend to place a pancreatic stent. After ERCP, the PAD patients are generally given somatostatin injection. After taking these measures, the AEs rate for PAD patients undergoing ERCP was similar to that in patients without PAD. However, we found that somatostatin injection did not prevent PEP in PAD patients. Somatostatin and its analog octreotide inhibiting pancreatic enzyme secretion were thought to be protective factors of PEP.[1] However, it may not benefit everyone.[3] The best technique is to inject a large dose of somatostatin one hour before ERCP and then maintain the injection for at least 10 hours after the operation.[27] The conclusion of our study may be due to selection bias. Another major cause for bias was that the bolus somatostatin injection before ERCP was not performed in our center due to medical insurance policy. Therefore, the role of somatostatin to prevent PEP in PAD patients' needs to be further evaluated through well-designed studies.
Our study had limitations worthy of discussion. First, it was a retrospective review of a cohort of patients in a single center, where selection bias may exist. Second, the incidence of type I PAD was low. Third, severe AEs were still rare although our overall sample size was reasonable.
Conclusion This study was a retrospective study with a large sample size. Although ERCP with difficult cannulation was more common in patients with PAD, technical, success, and AEs rates were similar to those in patients without PAD. Type I PAD and ERCP indication for pancreatic diseases were independently associated with difficult cannulation. The ERCP indication for bile duct stone was independently associated with smooth cannulation. Precut biliary sphincterotomy was safe to facilitate cannulation. After successful cannulation, small EST and balloon dilatation were preferred procedural techniques to large EST for ERCP to achieve clinical success. To avoid AEs, experienced endoscopists are required for PAD patients with old age, male gender, type I PAD, and difficult stones. Balloon dilatation was a much safer procedural technique to perform ERCP. Somatostatin injection after ERCP had no effect on the prevention of PEP in our study cohort.
Acknowledgments
This study was funded by the National Natural Science Foundation of China (to Z.D.J), No. 82170657 and the National Outstanding Youth Science Fund Project of National Natural Science Foundation of China (to H.J.H.), No. 82022008.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References 
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/sjg.sjg_311_22
留言 (0)