Search strategies and study selection

The search and selection strategy was the one used overall for the SEPD project, as detailed in the initial paper ⁴. Basically, two searches were performed: one for CPGs and one for original and review articles. Furthermore, references quoted in the chosen studies, as well as in the reported significant reviews, and clinical guidelines and meta-analyses, were secondarily reviewed. These references were peer reviewed, and selected when they included recommendations regarding ERCP preparation, execution and monitoring, or structural, process, and/or outcome quality indicators.

Endoscopy procedure design

Based on the literature selected and author experience, procedure-related activities were collected and sorted out. Procedures common to all endoscopic explorations and specific techniques used for specific situations were not included. The result was diagrammed in a parallel flow chart. Group proposals were reviewed and revised by a panel of SEPD-selected professionals until a definitive version was finalized.

Indicator construction

To obtain valid indicators, the quality of the knowledge available on procedure-related activities and the documents selected following the search was assessed. To this end, the GRADE evidence quality grading system was used, where quality is scored as high, moderate, low and very low ^6,7.

To ensure reliability and facilitate indicator calculations, each indicator is accompanied by a form including the following: area of application (procedures where it applies), name, formula, type, timing (pre-procedure, procedure, post-procedure), related quality dimension, justification, exclusions and clarifications, and supporting level of evidence.

Indicators common to all digestive endoscopy procedures, which are discussed in a different paper, are excluded ⁴, albeit supplementary remarks are included for some of them because of their ERCP-specific features.

ERCP procedure

The goal is to maximize procedure quality and safety in order to facilitate patient diagnosis and treatment while optimizing health outcomes (Fig. 1). It includes the following:

Fig. 1 Procedimiento de la CPRE. 

1. Patient positioning. Three positions are useful: usually the left lateral oblique-prone position is used, but the standard (left lateral) or prone position may also work.

2. Use of deep sedation or general anesthesia.

3. Duodenoscope check-up. The scope must provide adequate, high-quality viewing, and functions such as tip angulation, elevation, air and water provision, and suction must all work properly.

4. Fluoroscopy equipment preparation and focusing on the right hypochondrium.

5. Duodenoscope lubrication.

6. Duodenoscope insertion:

7. Observe and describe abnormalities in papillary area: diffuse mucosal infiltration, papillary conditions, diverticula, polyps, ampulloma. Biopsy collection is indicated in the latter case to facilitate tumor staging.

8. Biliopancreatic cannulation:

9. Act according to findings:

- In case of suspected or visualized choledocholithiasis, consider passing a balloon or using a - - Dormia basket. Should this fail, consider balloon sphincteroplasty or stent placement. - Tumor stenosis: sampling and stenting. - Non-tumor stenosis: dilation with/without stenting. - Fistula or bile leakage: sphincterotomy with/without stenting.

10. Withdraw tube. Maneuver slowly and aspirate air and gastric contents as you go.

11. End procedure. Withdraw tube completely and hand over to assistant for cleaning and preparation. Consider antibiotic prophylaxis.

Indicators

A total of 26 indicators (Table 1) have been included, of which seven refer to structure (common to all procedures), 17 refer to process, and two refer to outcome. The complete list of indicators is included in Table 2, but we shall only discuss the following because of their technique-specific status: process-pre-procedure (antibiotic prophylaxis and procedure difficulty grading [Schutz]); process-procedure (deep ductal cannulation in native papilla, common bile duct stone removal, stricture resolution, and radiation estimation). Two outcome indicators have also been developed for adverse events such as post-ERCP pancreatitis and post-ERCP bleeding, as these are highly specific and have an impact on clinical workload.

No indicator for ERCP volume per site, as recommended by some guidelines, has been included as we deem it indicative of quantity rather than quality. Furthermore, such indicator represents an attempt to measure endoscopist experience or training, an aspect already accounted for by indicators common to all endoscopic procedures ⁴.

Table 1 Quality indicators for endoscopic retrograde cholangiopancreatography (those elaborated upon in the text appear in boldface type) 

Numbering is consecutive to that established for colonoscopy markers ⁵ since this is a single project.

Indeed, not all indicators have the same weight. In fact, the American Society for Gastrointestinal Endoscopy (ASGE) de-fined as priority ERCP indicators those essential to obtain good results, including the following: appropriate indication, duct cannulation, stone removal, and post-ERCP pancreatitis ⁸.

It should be born in mind that in the GRADE system, which is used to assess the evidence supportive of indicators, "high-quality" evidence includes well-designed randomized studies, which are scarce in this setting. As a consequence, level of evidence is scored on both ends of the scale ("very high" or "low"), and "very high" levels are mainly obtained from clinical guidelines. It is because of this that many of our indicators match the ones therein. Randomized studies with moderate level of evidence are scarce, as are also other designs.

For clarification, indicator numbers (Table 1) are consecutive to those used in the prior two papers for general ⁴ and colonoscopy ⁵ indicators.

B.07. Antibiotic prophylaxis

Some clinical guidelines recommend antibiotic prophylaxis in some cases, including known or likely biliary obstruction, biliary or pancreatic fistula, pancreatic pseudocysts or necrosis, liver-transplanted immunosuppressed patients, and active bacterial cholangitis ^8,9. A Cochrane 2010 review concluded that prophylactic antibiotics reduce bacteremia and seem to prevent cholangitis, pancreatitis, and sepsis in patients undergoing elective ERCP, although their benefits may be less obvious in the subgroup of patients with uncomplicated ERCP ¹⁰. In a recent review ¹¹, the ASGE did not recommend antibiotic prophylaxis in the absence of suspected biliary obstruction or when complete biliary drainage is envisaged; they did recommend it otherwise and for liver-transplanted individuals. Appropriate antibiotics should cover biliary flora (including enterococci and Gram-negative germs).

B.08. Procedure difficulty assessment (Schutz grading)

To gain insight into the effectiveness of a procedure such as ERCP, it is key that the expected odds of success be known. In this regard, Schutz published a scale in an attempt to provide expected success data according to expected procedure difficulty (Table 2) ¹². Later, a Dutch registry analyzed the influence of such classification, recorded for one year using a standardized system (Rotterdam Assessment Form for ERCP [RAF-E]) ¹³, on procedure success, and established its relationship with difficulty levels (75.2%, 84.3%, 88.1%, respectively, for grades 3, 2, and 1), albeit within narrow ranges.

Table 2 Degrees of difficulty in diagnostic/therapeutic ERCP (Schutz) 

Endoscopists with lower numbers of procedures should likely perform grade-1 procedures exclusively, leaving those categorized as grades 2 and 3 to higher-volume institutions ^14,15. In this respect, the ASGE, based on this scale, has put forth a 4-grade scale with specific recommendations (levels 1 and 2 for basic endoscopists, levels 3 and 4 for experts) (Table 3) ¹⁶.

Table 3 ASGE degrees of complexity 

C.11. Deep cannulation of the desired duct in a native papilla

Successful cannulation of the desired duct (bile, pancreatic or both) is the foundation of the procedure's diagnostic and/or therapeutic success ^8,15, hence should be recorded in all cases. Cannulation is deemed as satisfactory when the guidewire passes through the papilla into the desired duct so that it may be advanced deeper and a contrast medium may be injected to visualize the entire duct system. All types of precut are included when the standard technique fails in order to avoid repeat endoscopic procedures or percutaneous cholangiography.

This indicator is restricted to patients with normal papillary anatomy, and also excludes:

Reference to a potential standard is based on task force agreements and the results of a meta-analysis including prospective studies, which found a percentage of 89.3% (95% CI: 0.866-0.919) for bile duct cannulation, and of 84% for pancreatic duct cannulation regardless of institution type ¹⁷. Furthermore, cannulation odds have been seen to be dependent on endoscopist experience: for ≤ 50 ERCPs/endoscopist/year, 84.2%; for > 50 ERCPs/endoscopist/year, 91% (mean, 89%) (p < 0.001) (18). Since most patients included in those studies were seen at reference sites, we decided to slightly reduce ASGE-established standards (> 90%) ⁸ as they are excessively optimistic and do not reflect the conditions extant in our setting.

A point of interest is difficult biliary cannulation, which the ESGE ¹⁹ defines as failed cannulation after five attempts, after five minutes, or after more than one pancreatic duct cannulation ²⁰. However, it is clear how this definition may apply to this deep cannulation indicator.

C.12. Choledocholithiasis extraction

Where this test is undertaken in order to remove stones, its success or failure should always be recorded, with adequate information on size and location, as well as on the potential presence of strictures or anatomical changes, including post-surgical ones ²¹.

This indicator is restricted to cases with smaller stones and normal anatomy, but this proportion may also be useful for other, more challenging stones. Within the scope of its restricted use, over 90% of stones should be successfully extracted with recourse to sphincterotomy and a balloon or basket when appropriate.

As is the case with cannulation, data from the meta-analysis show that extraction is successful in 88.3% (95% CI: 0.825-0.941) of cases ¹⁷. This is why we decided, as for cannulation, on the use of reduced, more realistic standards as compared to those established by the ASGE (> 90%) ^8,15. Meta-analysis results, however, probably encompass all types of lithiasis regardless of size ¹⁷.

Situations may be found that increase complexity and, besides stone size, also relevant are associated anatomical changes, strictures, unfeasible basket use or lithotripsy, and Mirizzi syndrome ²².

C.03. Stenosis resolution

When the procedure is aimed at stenting for biliary obstruction, the resulting success or failure, and whether it involved sphincterotomy ²³ or precut in order to access the bile duct, should always be recorded ^8,15, providing adequate information on indication, obstruction location, and presence or absence of strictures or anatomical changes, including post-surgical ones.

The indicator is restricted to the aforementioned cases but may also be useful for the rest. Placing a stent in a biliary obstruction distal to the hilum is considered as a relatively easy procedure, and an essential one for obstructive jaundice, particularly when associated with cholangitis.

As regards the cases it is limited to or indicated for, stenting should be successful in over 90% of them ^8,15, and meta-analyses have found success rates of 97.5% (95% CI: 0.967-0.984) ¹⁷. Only stenosis dilation would be inappropriate as an indicator because of its high relapse rate.

C.14. Radiation estimation

Fluoroscopy use is inherent in ERCP, this being the source of the latter's radiation exposure risk (both for patients and healthcare providers), which is similar to other radiographic studies ²⁴. Hence, measuring fluoroscopy time and including such measurement as a quality indicator has been suggested by various guidelines for more than ten years now ^8,25. The potential to optimize radiation usage without lessening effectiveness supports its inclusion ²⁶.

However, the correlation between fluoroscopy time and radiation dosage is notoriously poor ²⁷, hence other measurements should be used for this estimation whenever possible, including the dose area product (DAP) ²⁴. Various studies support the notion that received radiation dosage may be reduced without modifying fluoroscopy time ²⁴.

E.02. Adverse effects

Fewer complications have been shown to develop when endoscopists perform over 50 procedures per year ¹⁸.

E.02.1. Adverse effects: post-ERCP pancreatitis

By consensus, post-ERCP pancreatitis (PEP) ²⁸ is any newly developed or aggravated abdominal pain that is consistent with clinical pancreatitis, occurs beyond 24 hours after ERCP, is associated with an increase in serum amylase (and/or lipase) three times above normal levels, and requires hospitalization for at least two days ²⁹. This concept is highly significant since enzyme changes have been witnessed in up to 75% of patients following ERCP ³⁰. Its incidence is 3.5% (1.6% to 15.7%) ³¹. While these meta-analysis-derived figures are real, a selection of very high-risk indications (suspected sphincter of Oddi dysfunction) should be considered to establish more appropriate values. Guidelines exhibit no consensus on acceptable standards ^8,15, but we should aim at below 7%. Several factors have been shown to increase both pre-procedure and intraprocedure risk ²⁹ (Table 4).

Table 4 Post-ERCP pancreatitis risk 

Methods have been devised to reduce its incidence ³²: appropriate patient selection and indications, prophylaxis with rectal indomethacin ^33,34, use of pancreatic stents in case of initial pancreatic duct cannulation ³⁵ and guidewire cannulation.

E.02.2. Adverse effects: post-ERCP bleeding

Bleeding risk during ERCP ranges from 0.5% to 5% ³⁶, with 1% established as the most approximate figure ^37), nearly always associated with endoscopic sphincterotomy ³⁸.

It is defined as the presence of clinical complaints or blood extravasation into the gastrointestinal tract ³⁹, and one should wait for 2-3 minutes after sphincterotomy to assess it ⁴⁰. Furthermore, more severe delayed bleeding has also been described with melena, hemoglobin drops, or transfusion needs within ten days after ERCP. This type of bleeding is more easily associated with antithrombotic agents.

Balloon sphincteroplasty is safe, and its use without sphincterotomy is recommended for patients with coagulation disorders or juxtadiverticular papilla ³⁶.

Table 5 ³⁶ summarizes ERCP-related bleeding risk factors. Use of antithrombotic drugs stands out as the primary factor, although aspirin has not been shown to increase bleeding, hence sphincterotomy could be safely performed in patients on aspirin ^41,42,43.

Table 5 Bleeding risk in ERCP 

In summary, ERCP is an interventionist technique, and associated quality indicators with evidence of a high success rate in the cannulation of the desired duct (bile, pancreatic, both) include choledocholithiasis extraction and stenosis resolution. Also strong is the measurement of adverse effects, primarily post-ERCP pancreatitis and sphincterotomy-related bleeding.