Endoscopy 2008; 40(8): 675-676
DOI: 10.1055/s-2008-1077486

© Georg Thieme Verlag KG Stuttgart · New York

Are you safe for your patients – how many ERCPs should you be doing?

N.  M.  Guda1 , 2 , M.  L.  Freeman3
  • 1Pancreatobiliary Services, St Luke’s Medical Center, Milwaukee, Wisconsin, USA
  • 2School of Medicine and Public Health, University of Wisconsin, Madison, Wisconsin, USA
  • 3Minnesota Pancreas and Liver Center, University of Minnesota and Hennepin County Medical Center, Minneapolis, Minnesota, USA
Further Information

Publication History

Publication Date:
04 August 2008 (online)

Endoscopic retrograde cholangiopancreatography (ERCP) has evolved into a highly therapeutic procedure over the past three decades. The procedure has gone through various stages of development and refinement. The procedure, and hence the supporting literature, have evolved from the initial phases of technical know-how to understanding the risks and complications and the strategies for their reduction, and finally the outcomes. The ultimate goal is to do a procedure that is therapeutically effective with the least complications.

Patient outcomes are difficult to measure, largely because of difficulties with definitions. The therapeutic intent (goal of treatment), success of the procedure (have we done what we intended to do at the outset), and complications, though initially subjective, now have objective parameters for evaluation [1] [2] [3] [4] [5].

”Quis custodiet ipsos custodes?” (Who guards the guards?). Reporting outcomes has been a challenge. It should be done because, as a healthcare provider, it helps one to better him/herself. It helps practitioners to know where they stand compared with their peers, and it improves patient outcomes by a process of continual improvement in techniques and standards. It helps the patient as a healthcare consumer to know what his/her choices are. ERCP complications are dependent on certain characteristics of the patient, the type of procedure, and also, to a degree, on the endoscopist. Whereas the former two have been well studied, there are limited data regarding the operator.

Kapral et al. have evaluated the ERCP outcomes based on a nationwide benchmarking project in Austria [6]. The study was of a voluntary reporting system of 140 registered sites covering 8 million people in Austria, where the estimated total number of ERCP procedures is about 15 000 per year. The data for benchmarking and quality assurance was abstracted using current evidence and recommendations by the endoscopy societies. The authors meticulously validated the questionnaire and ensured that the burden of time was a reasonable 6 minutes per procedure. Interestingly, only 20 % of the registered sites with a case volume of 22 % participated in this project, despite the fact that data reporting was voluntary and anonymous. There was also a dropout rate of 16.7 % concerning data completeness. This demonstrates the difficulty in performing such large, multicenter self-reporting outcome studies.

The authors need to be commended for such an endeavor. The paper helps us to understand a perspective on complications from a very divergent group of practicing endoscopists. Published data on outcomes typically tend to be from the expert centers and from dedicated endoscopists who are willing and proud to demonstrate their outcomes. Those most in need of scrutiny may be less willing to take the time and effort or risk exposure of their outcomes. This effect may be counterbalanced by the referral bias for more complex and high-risk procedures at tertiary centers, and perhaps more organized and thorough reporting of complications.

Correlating outcomes to volume of procedures performed is relatively simple and has been evaluated in other surgeries and endoscopic procedures, but the results are mixed [7] [8]. In the Austrian study by Kapral et al., endoscopists were considered high volume if they carried out more than 50 procedures a year. Their data demonstrated that high case-volume endoscopists had better diagnostic and therapeutic success (86.9 % vs. 80.3 %) with fewer complications (10.2 % vs. 13.6 %) than lower case-volume endoscopists. These data are similar to those of a previous Italian study by Loperfido et al., in which the complications were higher (7.1 % vs. 2.0 %) in centers with low volumes (< 200/year) [4]. In a multicenter study in the USA, endoscopists who carried out no more than one sphincterotomy per week had higher complication rates compared with their peers who carried out higher volumes of sphincterotomies each week [2].

These studies support the concept that a lower case-volume affects outcomes adversely. In contrast, another recent large multicenter study of assessment of risk factors for ERCP complications from the UK found no difference in overall complications between endoscopists of differing case loads or hospital type [9]. The only difference found was a decrease in the risk of post-ERCP pancreatitis when the procedure was carried out at a university hospital compared with a district hospital – interpreted to perhaps reflect the better support staff and environment available at university hospitals. Reasons for the striking difference in findings between this study and the study of Kapral et al., which is quite similar in concept and design, are difficult to postulate. Most studies of post-ERCP pancreatitis, the most common complication, have suggested that case mix is at least as important as technical factors in determining risk; any difference in technical expertise is overshadowed by the difference in patient mix, which tends to be more complex and high risk at more specialized centers.

Added together, these studies demonstrate that there is a need for each endoscopist to carry out a certain number of ERCPs and sphincterotomies in order to both minimize the risk and improve outcomes. Endsocopist experience appears to be an under-appreciated risk factor. There are no harder data on outcomes based on the volumes outside large multicenter studies where, again, bias in reporting, complexity of the case mix, and definitions of complications all play a pivotal role.

The effort by the Austrian group is a much desired effort of a nationwide collective database. However, despite their best efforts, even they were successful in collecting data from only a minority of the procedures that were carried out. However, it is probably reasonable to conclude that the sample of cases evaluated was representative of day-to-day endoscopist practice. What one should appreciate is that important outcomes include failure to successfully complete a procedure, and the risk of complications, but also the added costs of repeat procedures, treatment of complications, and patient morbidity and mortality. Whereas therapeutic success and complications are based on the complexity of the procedure, patient-related risk factors, including comorbidities, available support staff, and endoscopists’ knowledge and experience, clearly seem to play a role in the outcomes. It is thus hard to answer the question ”How many ERCPs should you be doing?” Most endoscopists might be comfortable taking care of a patient with cholangitis or routine choledocholithiasis but might not have enough experience or have maintained a high enough skill level to do more complex procedures such as altered anatomy, pancreatic endotherapy, removal of large stones, and treatment of malignant obstruction. The number of ERCPs an endoscopist needs to carry out each year in order to prevent any complications is thus not a universal number and might also vary with the level of complexity of the procedure. Based on the available data, it appears that one needs to do at least 50 procedures per year and at least one sphincterotomy per week in order to carry out standard ERCP successfully for the most common situations. For those likely to perform complex therapy or achieve a consistently high rate of success, more than 100 to 200 ERCPs annually are probably required.

There is a clear demand for physician self-reporting and benchmarking in order to improve outcomes. Just such a voluntary benchmarking project is already underway in the USA, and data from that registry will be of great interest. Before we are mandated by our patients or payors and are asked the question ”How many of these procedures have you done, what are your complications, and how well do you compare with your peers?“, let us first answer it for ourselves.

Competing interests: None


N. M. Guda, , MD 

Pancreatobiliary Services
St Luke's Medical Center

2801 W.KK River Pkwy Ste
1030 Milwaukee
Wisconsin 53215

Fax: +1-414-543-8344

Email: nguda@wisc.edu