ERCP-induced perforation: review and revisit after half a century

Background Endoscopic retrograde cholangiopancreatography (ERCP) is an invasive procedure. We aimed to investigate ERCP-induced perforations at our institution and conduct a comprehensive review of literature on ERCP-induced perforations (EIP) since the introduction of this procedure as a therapeutic intervention. Methods This was a case-control study, in which charts of all patients diagnosed with ERCP-induced duodenal perforation were reviewed and compared to a control group without perforation. Patient’s sociodemographic and clinical data, including ERCP procedure-related data, were gathered. Results A total of 996 ERCP procedures were performed; only 13 patients proved to have EIP. Obstructive jaundice was the most common indication for ERCP. The main predisposing factor was difficult cannulation (P = 0.003). In total, five patients required surgical treatment; the majority of them had type I perforation, whereas type IV was the most common in patients who were treated conservatively. The overall mortality rate was 15%, the surgical group had a slightly higher mortality rate. Conclusions Fifty years after the introduction of ERCP for therapy, it remains an invasive procedure that carries significant morbidity and mortality, even in skilled hands or at high- volume units. Conservative management of perforation yields favorable outcomes in selected patients.


Introduction
Endoscopic retrograde cholangiopancreatography (ERCP) is one of the most commonly utilized procedures for both diagnostic and therapeutic indications.It was first developed in clinical practice in 1968 as a diagnostic tool. 1 Cannulation of the ampulla and sphincterotomy were introduced in the early 1970s with consequent evolution of ERCP into a therapeutic procedure. 1,2Nowadays ERCP is almost exclusively indicated for the therapy of hepatobiliary disorders since other less invasive diagnostic modalities (for example: magnetic resonance cholangiopancreatography (MRCP)) which possess similar sensitivity, are widely available.
In spite of the valuable contribution of ERCP in patients' management, it is not a complication-free intervention.The ERCP-related morbidity was estimated to be 6.9%; it also carries a mortality rate of 0.33%. 3The spectrum of complications includes pancreatitis being the most common, followed by cholangitis, bleeding and perforation.The incidence of severe complications is approximately four times higher in ERCP than in other endoscopic procedures. 4e of the most feared complications is ERCP-induced perforation (EIP).Although EIP is uncommon, it is potentially fatal.Moreover, the treatment of such complication is still controversial.
This study represents the first analysis of EIPs in Jordan.Along with this analysis, we aimed to conduct a comprehensive review of literature for ERCP-induced perforations since the introduction of this procedure as a therapeutic intervention.

Study design
This was a retrospective study in which gastroenterology, surgery, anesthesiology, and intensive care data were collected from the electronic medical charts of patients who underwent ERCP during the period of January 2016 to June 2020.
The following information was extracted: patient demographics, comorbidities, indication for endoscopy, ERCP findings, presumed risk factors for difficult procedure (sphincterotomy versus precut, dilatation procedures, failed cannulation and relevant past anatomy-altering operation).In addition, clinical presentation of perforation, radiologic findings, time to diagnosis, as well as perforation type according to the Stapfer classification 5 (Type I: lateralfree wallor medial duodenal wall perforation, type II: peri-Vaterian injuries, type III: distal bile duct or pancreatic duct perforation, type IV: retroperitoneal air alone) were all documented.The treatment approach (conservative or operative), need for additional interventions (e.g.percutaneous drainage), length of hospital stay (LOS), morbidity, and mortality are also reported.
Sex and gender differences between participants were not taken into consideration, because in this study, the choice of design is little affected by implications of sex and gender.The sex of participants was defined based on self-report.

Inclusion criteria
Patients who were diagnosed to have endoscopic perforation (EP) either during ERCP or by imaging studies such as computed tomography (CT) were included.As a control group, 43 patients without ERCP-related complications were randomly selected from the same time period.The randomization process was done by selecting a control patient from every 23 patients after arranging the 966 patients chronologically by ERCP date.

REVISED Amendments from Version 1
In this revised version of the manuscript, we have addressed reviewer's comments, minor changes were made.
The alterations included specifying the term precut papillotomy.We also elaborated more on the obstructive jaundice group and stated that only two of them were diagnosed with cancer.We also have changed the term CBD Balloon dilatation to CBD stricture balloon dilatation as requested by the reviewers.The type of the stent used during ERCP was plastic, a notice about this was added on table 3 of the manuscript.In page 8, a statement was corrected from "in other types of perforations, type I, II, III" to "in other types of perforations, type II, III, IV".
The criteria that is usually applied for endoscopist to performed ERCP were included on the result section first paragraph: board certified gastroenterologist, licensed to perform endoscopy after training in an accredited structured ERCP program, we also pointed out that all ERCP perforations happened on the first attempt of the procedure as written in the result section on first paragraph.
No changes were made to the names of the authors nor the affiliations, and no new figures or tables were added.
Any further responses from the reviewers can be found at the end of the article

Ethics
The study protocol was reviewed and approved by the Institutional Review Board at King Abdullah University Hospital and by the Committee of Research on Human Subjects at the Jordan University of Science and Technology (Non-funded research No: 20200428).Patients' data were kept anonymous and confidential.All study procedures were conducted according to the World Medical Association Declaration of Helsinki.

Procedures
The management plan was dictated by the attending surgeon on call.Patients who were treated conservatively were kept nil per OS and received intravenous fluids, broad spectrum antibiotics and placement of a nasogastric tube.Failure to improve with conservative management, alarming CT findings and the presence of diffuse peritonitis were considered indications for operative treatment.Patients received percutaneous drainage for intra-abdominal collections and enteral or parenteral feeding as appropriate.

Statistical analysis
Data were managed and analyzed using JMP (version 15).Statistical summaries were presented.Q-Q plots were used to examine normality assumption for continuous variables.Afterward, significant differences in participants' characteristics were analyzed using a t-test, Fisher exact test or logistic regression.Additionally, odds ratios were used for significant results.Alpha was set at 0.05.

Results
A total of 996 ERCP procedures were performed by three experienced endoscopists (board certified gastroenterologist, licensed to perform endoscopy after training in an accredited structured ERCP program).Thirteen patients (1.34%) were diagnosed with EIP.Additionally, 43 other patients who also underwent ERCP, but with no endoscopic induced perforation were randomly selected as a control group, all patients had the procedure for the first time (Table 1).
EIP group: There were nine women (69%) and four men (31%) with a mean age of 52.6 years.Approximately two-thirds (69%) of patients had hypertension, 18.9% had diabetes mellitus while none were diagnosed to have ischemic heart disease.Only two patients (15%) had a history of abdominal surgery, both of which were laparoscopic cholecystectomy.One patient had pancreatic adenocarcinoma and another patient was known to have cholangiocarcinoma.

ERCP characteristics
Dilated common bile duct (CBD) was observed in 31% of cases and 15% had a filling defect.Precut papillotomy was performed for 85% of the EIP group, 23% had a stone extracted during ERCP.Failure of cannulation was experienced in half of patients (six patients, 46%).

Risk factors for perforation Patient's characteristics
Notably, a higher number of patients had hypertension in the EIP group compared to the non-EIP group (69% versus 47%) but fewer had diabetes mellitus (18.9% versus 22.5%) with no statistical significance (Table 1).

Indications for ERCP
The presence of obstructive jaundice and post-cholecystectomy bile leak were more common indications in the EIP compared to the non-EIP group (61% versus 53% and 7% versus 2% respectively).Choledocholithiasis (23% versus 28%), cholangitis (4% versus 0%) and gallstone pancreatitis (7% versus 11%) were more common indications in the non-EIP group.However, the observed differences in the ERCP indications did not reach statistical significance (Table 2).Obstructive jaundice proved to be secondary to cancer in 2/8 patients in the EIP group.The rest were secondary to benign strictures.

ERCP characteristics
Failure to cannulate the common bile duct showed significant association with EIP in contrast to the non-EIP (46% versus 6%; P = 0.003).Interestingly, patients who underwent CBD dilatation experienced substantially lower rates of EIP (31% versus 67%: P = 0.0214).The presence of CBD dilatation and filling defects was less common among the EIP group (31% versus 46% and 15% versus 26% respectively) but with no statistical significance.We found that the most significant risk factor and predictor of duodenal perforation during ERCP was failure of cannulation with an odds ratio of 11.4 (P = 0.003).

Diagnosis and management
The majority of patients (11 patients, 84%) were diagnosed with EIP using CT scan while the remaining two patients (16%) were diagnosed during ERCP.One of the two patients developed extensive surgical emphysema during the operation while the other was diagnosed by directly visualizing the contrast extravasation through the duodenum.
Of the patients diagnosed after the procedure, the most common symptom was abdominal pain (11 patients, 84%) with the remaining two patients (14%) presenting in the form of peritonitis.
Conservative treatment was preferred for eight patients (61%), while five patients (39%) were treated operatively; (Table 3).Operatively treated patients were of a younger age and had less comorbidities such as hypertension and diabetes, which might have influenced the choice of management.Moreover, half of the conservatively-treated group had a history of abdominal surgery, namely laparoscopic cholecystectomy, compared to only one patient (20%) in the surgical group, which could also have affected the surgeon's choice of treatment.Conservatively managed patients had more dilated CBD as opposed to the surgically treated (37% versus 20%).Papillotomy, dilatation of CBD and stone extraction were much more commonly performed in the conservatively treated group (100% versus 60%, 50% versus 0% and 37.5% versus 0%, respectively).Otherwise, there was no significant differences in ERCP characteristics between the surgical and the conservatively treated groups (Table 3).

Treatment according to perforation type
Overall, 75% of patients who had type I perforation were treated surgically whereas none of patients with type III or type IV perforation were operated.Type II perforation was divided between conservative treatment (two patients) and surgical management (one patient).

Perforation type by treatment
The majority (80%) of the surgically treated patients had type I perforation.Of those treated conservatively, three had type IV perforation, two had type II perforation, two had type I perforation and 1 had type III perforation (Table 4).

LOS, morbidity and mortality
Surgically treated patients stayed longer in the hospital with an average of 24 days compared to 17.25 days for the conservatively treated group (Table 4).Additionally, ICU admission rates and the average LOS in the ICU were both higher among the surgically treated patients (60% versus 25% and 7.94 versus three days).Major complications were noted in slightly higher rates among the surgical group.Of those treated surgically, one patient (20%) suffered from an intra-abdominal abscess requiring drainage and another patient (20%) developed pneumonia.On the other hand, one patient (12.5%) among the conservatively treated patients had an intra-abdominal abscess requiring drainage and another patient (12.5%) sustained acute kidney injury.About half of the patients in both groups received enteral or parenteral feeding at some point of their hospital stay.The overall mortality rate was 15% with a slightly higher rate in the surgical group (20% versus 12.5%).

Discussion
In this study, we investigated the incidence rate of duodenal perforation among patients undergoing ERCP and compared patients' and procedure's characteristics to a control group.Our results revealed that perforations occurred in a comparable rate (1.34%) to published series.Our study also disclosed that failure to cannulate the CBD was the main predictor of perforation during ERCP.
The reported incidence of ERCP related perforations varied between 0.11% and 2.4%. 6,7In one large analysis of 21 prospective surveys of post-ERCP complications in adults (16,855 patients) between 1977-2006, 101 patients experienced perforation (0.60%).The mortality rate in that study was 9.9%. 3In another review of 18 (mainly retrospective) studies conducted between 2000-2014, the incidence rate of EIP was 0.39%, with an overall mortality of 7.8%. 8Surprisingly, published numbers did not seem to decline significantly over the last 20 years as ERCP became more popular and widely available.
Published data revealed that the incidence rate of ERCP perforations corresponds with volume of cases at each institution, as highes-volume centers (which perform more than 1000 ERCPs per year) reported lower incidence rates (Bill et   5,6,[14][15][16] The correlation between incidence of perforation and volume of ERCP cases needs further analysis to elaborate statistical significance.These observations, however, indicate that ERCP remains an invasive procedure which may carry significant morbidity and mortality, even in skilled hands or at high volume units. "Difficult cannulation" is a term employed to describe failure to gain access into the bile duct by the conventional cannulation technique.Factors that may contribute to this difficulty include presence of periampullary diverticula, altered anatomy and bulky papilla.In many studies, this situation has been defined as a leading cause of perforation. 17The risk is highest as precut papillotomy (or sphincterotomy) is attempted by the endoscopist to overcome this challenge.In Vezakis et al.'s review, endoscopic sphincterotomy was responsible for 41% of perforations. 8Another study by Stapfer and colleagues found that cannulation of the ampulla was considered difficult by the endoscopist in 10 of 14 patients who had perforations (out of 1413 ERCPs). 5In their review of 21 prospective reports of ERCP complications, Andriulli et al., found that the overall complication rate was significantly higher whenever therapeutic interventions were utilized during ERCP. 3 Complications were also significantly higher in studies with a precut sphincterotomy rate of over 10%.Fifty years after the advent of therapeutic ERCP, difficult cannulation is still a frequently encountered condition, which mayalthough rarelypredispose to perforation and subsequent morbidity and death.
9][20] In our group of patients with perforation, surgery was required in 5/13 (38%), 4/5 of the operated patients had type I (free wall) perforation.Surgical intervention is conducted in higher risk injuries that are more likely to progress into sepsis, which explainsin partthe higher morbidity observed in this subgroup.Because duodenal or duct wall defect identification during surgery might be very challenging, especially if diagnosis is delayed (>6 hours post ERCP), 21,22 surgery can be limited to perform proper drainage and debridement of unhealthy tissue, which can reduce the risk of systemic manifestations and sepsis.Diversion surgeries (Roux-en-Y bypass) have been reported in management of large duodenal perforations. 5However, in ERCP most perforations are small, unless caused by the scope itself.
Endoscopic repair of duodenal perforations, particularly small defects(<10mm) that are recognized during ERCP has been reported. 23,246][27][28][29][30][31] With the growing body of evidence supporting the use of these techniques for the management of ERCP complications, endoscopist expertise, perforation type and diameter remain important predictors of the outcomes.
In other types of perforation, i.e. types II, III and IV, the injury is less likely to manifest as peritonitis, but rather as retroperitoneal (and to a less extent intraperitoneal) fluid or air accumulation.Conservative measures in such circumstances vary from simply restricting oral intake with parenteral nutrition, hydration and coverage with broad spectrum antibiotics to percutaneous drainage of collections under ultrasound or CT guidance.
It is critical to recognize if leakage has stopped after the endoscopy or is still ongoing, as the patient may have experienced transient fluid extravasation during the procedure due to duct or duodenal wall puncture.The presence of enlarging pockets of pus or fluid collections, especially in the retroperitoneal space, does not necessarily indicate an active leakage, which can usually be excluded utilizing CT scan with water soluble contrast (Gastrografin), or fluoroscopic imaging (meal).This may instead indicate inadequate drainage or persistent infection that mandates repositioning of the drain, upgrading the size of the drain or placing another drain.It may also be helpful to consider adding antifungals after submitting samples for cultures.
The following are literature-based guidelines which can be driven to conclude when and how to intervene in patient with suspected or proved ERCP-induced perforation: • Suspicion is raised whenever cannulation is difficult, the threshold for obtaining a post ERCP CT scan to exclude perforation has to be lowered, since early detection may improve outcomes.
• Active intraperitoneal contrast extravasation from duodenal wall on CT scan is considered a reasonable indication for prompt operative exploration.
• The main target of surgical exploration is to control sepsis by drainage of accumulated fluids.Repair of the defect, if identified, is another target with special attention to prevent occlusion of the ducts.
• The presence of free fluids inside the peritoneum or in the retroperitoneal space is an indication for drainage; this can be achieved by interventional radiology (IR) under CT scan or ultrasound guidance, or by surgery if IR service is not available or the collection is not accessible.
• Most bile duct perforations can be managed by internal biliary stents, along with drainage of any collection.External nasobiliary drainage is considered a valid alternative for internal stenting.
• Failure of non-operative measures, which can be defined as the persistence of abdominal sepsis (significant pain, tenderness and ileus), as well as fever and continuous elevation of inflammatory markers (leukocytes, CRP, ESR, among others), indicates prompt conversion into surgical approach.
• Prolonged restriction of oral intake may not be of any help if there is no evidence of an ongoing leakage.
• ERCP perforation remains an event that has to be approached by both surgeons and endoscopists.
This study had some limitations.Firstly, it was a retrospective study.Secondly, the sample size of patients with EIP was relatively small to evaluate risk factors with statistical significance.Thirdly, we assume that a small group of patients with type IV perforation may have not been detected, due to lack of abdominal or systemic manifestations.
In conclusion, fifty years after introduction of ERCP for therapy, it remains an invasive procedure with certain complications that carry significant morbidity and mortality, even in skilled hands or at high volume units.Selection of patients for ERCP must be strict, it has to be done for therapeutic indications.Difficult cannulation is still a condition that is frequently encountered, and considered the main risk factor for perforation.Early diagnosis and appropriate surgical or percutaneous drainage yield favorable outcomes.
independent or the attendant statistician of the journal.
I have read this submission.I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however, I have a number of small changes to the article, or specific, sometimes more significant revisions.

Budhi Ida Bagus
Sebelas Maret University, Surakarta, Indonesia Good written review manuscript.It might be better if you can described the criteria of endoscopist who performed the ERCP procedures (if possible).
Please add the information whether the ERCP-induced perforation cases were occurred in the first or the second procedure of ERCP?Were there any information about the duration of each ERCP procedures which associated with perforation?
).This will help in standardizing the comorbidity burden.Additionally, a sub-group analysis to see correlation with higher CCI to outcome (specifically between conservative and operative arm) can provide more information on case selection for intervention and increase generalizability of this study.

Failure of cannulation:
Since two patients (of the 13) were found to have a perforation during the procedure, it is assumed that the procedure was prematurely terminated.It should be considered to remove them from the analysis of 'failed cannulation' as that instills bias.In other words, in these two patients, perforation led to failed cannulation and not the other way around.Page #4 and table 2: Need to specify precut papillotomy vs papillotomy.Also need to look into how many patients actually received sphincterotomy (with length) and balloon sphincteroplasty (with diameter).
Page #4: I am not why the status of having HTN or DM was looked at.Was this relevant?
Page #5, Table 2.: For the obstructive jaundice, were these all due to underlying malignancy such as HOP cancer?worth elaborating more.

○
For "CBD balloon dilation", did this mean CBD stricture balloon dilation or balloon sphincteroplasty?
○ Page 6 and table 3: Please comment on stent placement (plastic/metal) in cases of EIP.
Page #6: Worth mentioning if any patient had post ERCP pancreatitis and whether that could have affected the outcome.
Table #4: I am not sure that comparing surgical and conservative groups help given that the main difference is likely the severity/type of perforation so may be comparing the outcomes of different types of perforation would have been more helpful.
Page 8: "In other types of perforations i.e types I, II and IV"...was it meant to be "type II, III and IV?

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound?Partly

Are sufficient details of methods and analysis provided to allow replication by others? Partly
If applicable, is the statistical analysis and its interpretation appropriate?Yes Are all the source data underlying the results available to ensure full reproducibility?
gathering necessary details in future work.
8-The option of treatment correlates with the type of perforation, however, the outcomes have been discussed from a different perspective, which is the mode of treatment.The question raised initially was: do more invasive treatment options carry worse outcomes?
The only approach to answer this was to review outcomes in the light treatment method.9-Yes, sir, it was meant to be type II.III and IV, The editor is contacted so that the error will be corrected in the original manuscript.
Competing Interests: No competing interests were disclosed.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias • You can publish traditional articles, null/negative results, case reports, data notes and more • The peer review process is transparent and collaborative • Your article is indexed in PubMed after passing peer review • Dedicated customer support at every stage • For pre-submission enquiries, contact research@f1000.com

©
2023 Hammad H.This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Hazem Hammad Division of Gastroenterology and Hepatology, Department of Medicine, University of Colorado, Aurora, Colorado, USA Thanks for submitting this excellent research project.
b DM, diabetes mellitus.c IHD, ischemic heart disease.d NS, not Significant.e SD, standard deviation.
a CBD, Common bile duct.b NS, not significant.c not applicable.d EIP, Endoscopic Induced Perforation.

Table 3 .
Conservative vs. operative treatment: Patient and ERCP characteristics.
a HTN, hypertension.b IHD, ischemic heart disease.c DM, diabetes mellitus.d CBD, Common bile duct.e NS, not significant.f NA, not applicable.g SD, standard deviation.

Table 4 .
Conservative versus operative treatment: Perforation type, morbidity and mortality.
a CBD, Common bile duct.b RP, retroperitoneal.c LOS, length of hospital stay.d SD, Standard deviation.e TPN, Total parenteral nutrition.f ICU, Intensive Care Unit.

Is the work clearly and accurately presented and does it cite the current literature? Yes Is the study design appropriate and is the work technically sound? Yes Are sufficient details of methods and analysis provided to allow replication by others? Yes If applicable, is the statistical analysis and its interpretation appropriate
1. Haraldsson E, Kylänpää L, Grönroos J, Saarela A, et al.: Macroscopic appearance of the major duodenal papilla influences bile duct cannulation: a prospective multicenter study by the Scandinavian Association for Digestive Endoscopy Study Group for ERCP.Gastrointest Endosc.2019; 90 (6): 957-963 PubMed Abstract | Publisher Full Text 2. Tari E, Gagyi EB, Rancz A, Veres DS, et al.: Morphology of the papilla can predict procedural safety and efficacy of ERCP-a systematic review and meta-analysis.Sci Rep. 2024; 14 (1): 7341 PubMed Abstract | Publisher Full Text ? I cannot comment.A qualified statistician is required.

Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes Competing Interests:
No competing interests were disclosed.

have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Is the work clearly and accurately presented and does it cite the current literature? Yes Is the study design appropriate and is the work technically sound? Yes Are sufficient details of methods and analysis provided to allow replication by others? Yes If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes Competing Interests:
No competing interests were disclosed.

confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. Version 1
This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

the work clearly and accurately presented and does it cite the current literature? Yes Is the study design appropriate and is the work technically sound? Yes Are sufficient details of methods and analysis provided to allow replication by others? Yes If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes Competing Interests:
No competing interests were disclosed.

confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
The criteria of endoscopist who performed the ERCP procedures are added to the text ( please see the result section , first paragraph) : board certified gastroenterologist, licensed to perform endoscopy after training in an accredited structured ERCP program.2-TheERCP-inducedperforation cases were occurred in the first time , this has also been clarified in the result section first paragraph.3-Unfortunately, the duration of the procedure was not documented in the ERCP reports , neither stored on the patients electronic files.For standardizing the impact of comorbidities on the outcomes, it is recommended to use a standardized and validated comorbidity index such Charleson Comorbidity Index (CCI) (Ref 1, Ref Reviewer Report 05 February 2024 https://doi.org/10.5256/f1000research.142334.r238886© 2024 Wadhwa V et al.This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Vaibhav Wadhwa University of Texas Health Science Center, Texas, USA Prateek Harne University of texas health science center, Texas, USA Thank you for submitting your detailed work to this journal.Authors present data on ERCPinduced perforation which serves as a helpful guide for gastroenterologists.There are some major revisions that are suggested: Patient demographic table.