Introduction

Post-ERCP pancreatitis is the most common complication after an endoscopic retrograde cholangiopancreatography (ERCP) with an incidence range from 2.9 to 14.7% in unselected patient populations resulting in pancreatitis related mortality rate of up to 3% [1,2,3].

There is substantial evidence in favor of a number of effective prophylactic measures for preventing post-ERCP pancreatitis. Rectal nonsteroidal anti-inflammatory drugs (NSAIDs), pancreatic duct (PD) stenting, and intensive intravenous (IV) hydration are commonly accepted interventions that reduce the risk of this complication [4,5,6]. It is unclear whether a combination of these measures offers additive or even synergistic risk reduction beyond the protection conferred by the individual measures [7,8,9,10].

In 2013, we assessed the Dutch clinical practice using a nationwide survey among all ERCPists. This was instigated as a preparatory means to design a nationwide post-ERCP pancreatitis trial under the auspices of the Dutch Pancreatitis Study Group [9]. There have been multiple developments in the field following this survey. International guidelines have been adapted, particularly when it comes to post-ERCP pancreatitis prophylaxis [11,12,13]. Since 2010, the guideline of the European Society of Gastrointestinal Endoscopy (ESGE) recommends rectal NSAID administration for all patients without contraindication, whereas the 2017 guideline of the American Society of Gastrointestinal Endoscopy (ASGE) recommends rectal NSAIDs only for high-risk patients [12,13,14]. Both ESGE and ASGE guidelines recommend PD stent placement in high-risk patients. The 2019 ESGE guideline state that intensive IV hydration should be considered in patients with a contraindication to NSAIDs, whereas the ASGE guideline of 2017 suggests the use of intensive IV hydration whenever possible in a moderate to high-risk populations.

These changes from 2013 onward raise questions on the potential impact of guideline changes on Dutch gastroenterologists’ clinical practice. That is why we reassessed the clinical practice of post-ERCP pancreatitis prophylaxis in The Netherlands to identify potential points of improvement to inform and educate ERCPists about current evidence-based practice.

Methods

The survey was developed following the Declaration of Helsinki and using guidelines in the conduct of survey research [15]. The Medical Ethical Review Committee of the Radboudumc in Nijmegen, the Netherlands, approved this study and waived the need for informed consent. The study was registered in the online Netherlands Trial Registry (NL8493).

Study Setting

In the 2013 and 2020 surveys, the Dutch Society of Gastroenterology (NVGE) was consulted for all affiliated gastroenterologists, respectively, 408 and 575 for these 2 years. These account for 87% (2013) and 90% (2020) of the gastroenterologists registered in the Netherlands [16]. Gastroenterologists were eligible for participation if their current employment was in a community or academic center in the Netherlands and their institutional email address could be attained. For the 2020 survey, all endoscopy departments were separately consulted on whom of their gastroenterologists performed ERCPs (n = 219). Participation was voluntary and no incentives were offered for completion of the survey. The anonymity of participants was guaranteed.

The first survey was distributed as a postal survey from August to November 2013 to all affiliated gastroenterologists of the NVGE, no reminders were sent to non-responders. At that point in time, our research group was exclusively interested in the answers of ERCPists; hence, only their responses were recorded. The online survey was distributed by email between February 6th and March 18th, 2020. Two email reminders were sent with 2 weeks interval for an optimal response rate.

Survey Design

The 2013 survey was designed based on the content of the 2010 ESGE and 2012 ASGE guidelines concerning the clinical practice of post-ERCP pancreatitis prophylaxis (Fig. 1) [14, 17]. Three domains were constructed to investigate self-reported use of guideline-recommended prophylactic measures and factors influencing the decision to utilize them. The domains were as follows: respondent characteristics, post-ERCP pancreatitis prophylaxis use, and risk factor recognition (Supplementary Appendix 1). The international guidelines recommended rectal NSAIDs and prophylactic PD stenting at that time and thorough details on application and configuration of these prophylactic measures were included. The complete questionnaire consisted of 43 questions and the majority was compulsory and closed-ended formulated and came with a predetermined list of answer choices. Depending on the content of the question, one or multiple answers could be selected. When appropriate, open-ended questions or answer choices were offered for responses not anticipated from the survey design.

Fig. 1
figure 1

Timeline of landmark trials, guideline publications, and survey administration. RCT randomized clinical trial, PD pancreatic duct, NSAID nonsteroidal anti-inflammatory drug, IV intravenous, ESGE European Society of Gastrointestinal Endoscopy, ASGE American Society of Gastrointestinal Endoscopy

Full size image

For the 2020 survey, the 2013 survey was adopted to an electronic format, containing similar domains and formatted in Google Forms (https://docs.google.com/forms). Both the ASGE 2017 and ESGE 2019 guidelines were consulted for additional recommendations relative to the 2013 survey (Fig. 1) [12, 13]. Most notably, intensive IV hydration was introduced as a prophylactic measure. Detailed questions on this prophylactic measure were therefore added to the 2020 survey. Additionally, one risk factor (previous pancreatitis) was removed due to concerns of its unambiguous interpretation caused by similarity to another risk factor. These amendments resulted in an updated survey containing 47 questions (Supplementary Appendix 2).

All gastroenterologists in the 2020 survey were required to fill in the respondent characteristics and risk factor recognition domains of each survey. ERCPists answered in-depth questions on the different prophylactic measures only when they indicated that they actually used that particular type of prophylaxis.

The draft versions were piloted by a select group of five expert ERCPists. A consensus was formulated that a risk factor was sufficiently recognized if 85% or more of the respondents considered it as such. Their feedback was further implemented to improve on readability, content validity, and unambiguity of the questions.

Statistical Analysis

Categorical data were listed as frequency and percentage. Numeric data were assessed by mean with standard deviation (SD) or median with interquartile range (IQR). The Chi-squared test was used for comparison with the 2013 survey and subgroup analysis. All statistical analyses were conducted using SPSS, version 24, with statistical significance set at a two-sided alpha level of 5%. The difference in prophylaxis use between ERCPists in 2013 and 2020 was a predefined subgroup analysis.

Results

Response

Of the 408 gastroenterologists approached by mail in the 2013 survey, 121 (30%) ERCPists responded. The total number of ERCPists among approached gastroenterologists could not be discerned for the 2013 survey. The online survey administered in 2020 reached 548 of the 566 approached gastroenterologists. Of those reached, 161 gastroenterologists responded (29%). The response rate among currently practicing ERCPists was 50% (n = 109).

Respondent Characteristics

The characteristics of the respondents of the 2013 and the 2020 survey are listed in Table 1. Baseline characteristics did not significantly differ between surveys. In the 2020 survey, the mean age of the 161 respondents was 46 year and 116 (72%) were male. Twenty-five (16%) were employed in an academic center. Of the ERCPists, the mean practice time was 13.1 years, with a median lifetime exposure of 750 ERCPs.

Table 1 Characteristics of respondents
Full size table

Pharmacological Prophylaxis

In the 2013 survey, 98% (n = 119) of the ERCPists used NSAID prophylaxis, of which 62% (n = 73) in more than 80% of the ERCPs performed (Table 2, Supplementary Appendix Table S1). In comparison, all ERCPists in 2020 used NSAIDs prophylaxis. We found no rise in the percentage of ERCPists applying NSAID prophylaxis in comparison to the 2013 survey (p = 0.18); however, there was a significant rise in the proportion of ERCPs in which prophylactic NSAIDs were used (respectively, 62% vs. 93%; p < 0.001).

Table 2 Prophylactic measures
Full size table

Diclofenac was the preferred NSAID of choice and accounting for 95% in the 2020 survey. A dosage of 100 mg was used by 95% (n = 103) of the respondents and the drug was administered rectally by 98% (n = 107) of respondents. NSAID prophylaxis was administered prior to ERCP by 78% (n = 85) of the respondents.

Prophylactic PD Stenting

Prophylactic PD stenting was used by 62% (n = 75) in the 2013 survey versus 78% (n = 85) of the ERCPists in the 2020 survey (p = 0.008, Table 2). Of the 2013 PD stent users, 85% (n = 62) used PD stents in less than 20% of the ERCPs they performed versus 80% (n = 68) in 2020 (p = 0.33).

Table 2 and Supplementary Appendix Table S2 show the responses with concerning (regularly used) stent configuration, diameter, length, shape, and flanges. In the 2020 survey, the majority (88%, n = 75) of the ERCPists checked stent dislocation through radiological imaging, 79% (n = 67) between 5 and 14 days. Given, a few (8%) respondents did indicate not checking for stent dislocation in the 2020 survey.

Intensive IV Hydration

The 2013 survey did not contain questions regarding intensive IV hydration. In the 2020 survey, intensive IV hydration was used by 33% (n = 36) (Table 2 and Supplementary Appendix Table S3). Lactated Ringer’s was the IV solution used by 31 respondents (86%). Fluid administration was most frequently in a maintenance dose (72%), and 39% reported bolus administration.

Combination of Prophylaxis

In 2020, the combination of rectal NSAID and PD stent prophylaxis was used more often compared to 2013, 69% (n = 75) versus 48% (n = 58) (p = 0.002, Table 2 and Supplementary Appendix Table S4). The percentage of procedures in which a combination of rectal NSAID and PD stent prophylaxis was used did not differ, 88% (n = 50) versus 79% (n = 59), respectively, in less than 20% of the procedures (p = 0.24). The most cited motivation for this combination was the routine use of NSAID prophylaxis and PD stent placement in the case of repeated or deep guidewire insertion/contrast injection occurred during the ERCP.

The combination of rectal NSAIDs and intensive IV hydration was used by 34% (n = 37) of the 2020 ERCPists, of which 54% (n = 20) in less than 20% of the ERCPs performed and 27% (n = 10) in more than 80% of the ERCPs performed. The most cited reasons favoring the combination were maximal risk reduction for post-ERCP pancreatitis and utilization in high-risk patients and/or procedures.

Risk Factor Recognition and Prophylaxis Use

In the 2013 survey, 11 of the 18 scientifically well-established and recognized risk factors were identified by more than 85% of the ERCPists (Fig. 2, Supplementary Appendix Table S5). In 2020, eleven of the seventeen risk factors were identified as such. There were a few differences in risk factors identified by the ERCPists in 2013 compared to 2020: balloon dilatation of the papilla was identified by more than 85% and non-dilated extrahepatic bile duct system by less than 85% of the 2020 ERCPists. Three of the risk factors (earlier post-ERCP pancreatitis, non-dilated extrahepatic bile duct system, and normal bilirubin) were categorized as risk factors less frequent by the 2020 respondents in comparison to the 2013 respondents (p = 0.048, p = 0.002, p = 0.001, respectively). Absence of chronic pancreatitis, PD cannulation, and difficult cannulation were categorized as risk factors more often by the 2020 respondents p = 0.005, p = 0.001 and p = 0.004, respectively.

Fig. 2
figure 2

Post-ERCP pancreatitis risk factor recognition by ERCPists in 2013 and 2020. Balloon dilatation papilla is without prior sphincterotomy. PEP post-ERCP pancreatitis, PD pancreatic duct, SOD sphincter of Oddi dysfunction, CBD common bile duct. *p < 0.05

Full size image

In the 2013 survey, ampullectomy was seen as the only risk factor that warranted a combination of rectal NSAID and PD stent in more than a third (n = 33) of respondents (34%, Supplementary Appendix Table S6). The ERCPists stated in 2020 that NSAID prophylaxis was the preferred prophylaxis for 14 of the risk factors (83%). In the case of ampullectomy or PD contrast injection, a similar proportion of ERCPists elected NSAID prophylaxis or combination of NSAID and PD stent prophylaxis (respectively, 38% and 39%). If the PD was cannulated, some 50% chose the combination, while 34% relied on NSAID prophylaxis.

Eighty-three percent of the 2013 ERCPists indicated the presence of a post-ERCP pancreatitis prophylaxis protocol in their center compared to 84% (n = 92) in 2020 (Supplementary Appendix Table S7). Of the respondents in 2020 who indicated that a protocol was present, NSAID prophylaxis was included in nearly 100%, PD stent prophylaxis in 31% in the 2013 survey, and 24%. Intensive IV hydration was present in 38% of the protocols in 2020.

Discussion

Rectal NSAID is the most frequently used prophylaxis to prevent post-ERCP pancreatitis in the Netherlands and is applied by all ERCPists that responded to the 2020 survey. A steep rise in NSAID prophylaxis was observed from 2013 to 2020, exemplified by an increase from 62 to 93% of the respondents in more than 80% of ERCPs. Prophylactic PD stent placement was used in selected cases by 78% of ERCPists in 2020, significantly more than in 2013. One-third of the surveyed ERCPists in 2020 acknowledged the use of intensive IV hydration. Currently, exclusive use of NSAIDs was the preferred prophylactic measure, except for ampullectomy, PD contrast injection, and PD cannulation, in which the combination of NSAID prophylaxis and PD stent was equally favored or preferred.

In 2010 and 2012, two European surveys targeting ERCPists reported 16% and 35% use of rectal NSAID, respectively [18, 19]. The results of a landmark trial published months after the 2012 survey was conducted, demonstrating a significant decrease in post-ERCP pancreatitis due to the use of rectal NSAIDs. Their use as post-ERCP pancreatitis prophylaxis became more common practice, as demonstrated by our already high 98% NSAID use rate among ERCPists in 2013. The high percentage might be attributed to ease of use and negligible costs in certain countries [20]. The favored timing (before ERCP), route (rectal), and dosage (100 mg) were answered more frequently by the 2020 respondents. The shift in answers toward these recommended administration methods in the 2020 survey could be explained by our first survey, predating more recent administration recommendations (Table 3).

Table 3 European and American Society of Gastrointestinal Endoscopy guideline overview
Full size table

It is unclear why only 78% of the ERCPists in 2020 agree with the indication of PD stent placement, as it is regarded as an evidence-based prophylactic strategy [5, 21]. Several surveys have reported considerable practice variation in prophylactic PD stent use ranging from 53 to 96% [18, 19, 22]. Given that extra manipulation of the PD increases the likelihood of post-ERCP pancreatitis, PD stent placement is not without risk [23], and the risk rises further if placement fails [24]. A 2011 survey among US gastroenterologists showed low-volume ERCPists (< 50 ERCPs/year) feel less comfortable with pancreatic duct stenting, providing a potential explanation for this discrepancy [25]. With respect to stent configuration, the best consensus in the 2020 survey was reached for diameter, with 92% using 5Fr. The flange options “internal” or “external and internal” were given by 14/89 times by 2020 respondents, despite current guidelines arguing internal flanges can impede spontaneous dislodgement [13, 26].

While each prophylactic measure confers protection on its own, their combined effect has not been confirmed by current literature. ERCPists in our 2020 survey did demonstrate consensus on applying NSAID and PD stent prophylaxis in high-risk patients with (inadvertent) PD cannulation. The additional benefit of this combination in high-risk patients is expected to follow from a multicenter randomized controlled trial [10]. Considering intensive IV hydration, a relatively small percentage of ERCPists reported using this prophylactic measure in our 2020 survey compared to 83% in a 2018 North American survey [27]. Intensive IV hydration may perhaps be the preferred prophylactic measure in North America given the elevated price of the composite rectal indomethacin in certain countries [28, 29]. Despite a network meta-analysis suggesting this combination yields the lowest post-ERCP pancreatitis incidence [30], the future results of another multicenter randomized controlled trial will provide additional clarification [9].

The risk factor recognition of surveyed endoscopists must be interpreted with caution. The most recent international guidelines are not consistent on the contribution of each individual patient- or procedure-related risk factor for the development of post-ERCP pancreatitis, such as a non-dilated extrahepatic bile duct system [12, 13]. Each risk factor’s weight has been determined by a limited number of studies with varying quality of evidence, often predating the implementation of routine NSAID prophylaxis [29, 31].

The promising results from initial pilot studies on post-ERCP pancreatitis prophylactic measures prompted multiple large-scale trials to verify their effectiveness (Fig. 1). The results have led to guidelines becoming more specific on which patients benefit most from the various prophylactic measures [12, 13]. In the Netherlands, multicenter trials of the Dutch Pancreatitis Study Group have been designed with the input of ERCPists. These trials contribute to their knowledge base and the implementation of evidence-based practice. Additionally, the implementation of a national endoscopic complication registry in 2009 has led to more awareness for post-ERCP pancreatitis and usages of potential prophylactic methods. The mandatory complication meetings, discussing individual cases of post-ERCP pancreatitis with quality improvement goals, have contributed to an increased interest in novel prophylactic measures [32, 33].

A strength of this study is its high response rate among ERCPists, which is concurrent with other surveys among gastroenterologists [18, 19, 27]. This attributed to a valuable insight into the opinions and attitudes of ERCPists with regard to the use of post-ERCP prophylaxis in the Netherlands over time. Secondly, this has been the first survey conducted since the publication of the current ESGE guideline 1 year ago [13]. The questions on the various prophylactic methods were detailed, which resulted not only in an indication of the frequency of use, but also in which manner and with what motivation these methods were applied. The process of survey design was meticulous, with questions based on the most recent guidelines. Lastly, although the prophylaxis use and risk factor recognition captured by this survey may differ for other regions of the world, we can cautiously extrapolate our results to the European clinical practice given the use of the same guidelines. Thereby, our survey can shed insight into where there are differences, and we hopefully ignite international debate on best practices for preventing post-ERCP pancreatitis.

A limitation of the current study was the lack of data on actual post-ERCP pancreatitis rates of the surveyed ERCPists. The goal of our study was to capture the nationwide clinical practice and not investigate prophylactic measures set off against risk factors, justifying our methodology. Response bias could be introduced in our survey due to the possibility that ERCPists interested in reducing post-ERCP pancreatitis rates responded more often. Recall bias is an inherent limitation of survey sampling, where responses cannot directly be considered reflective of the general practice, as demonstrated in a recent retrospective review of electronic health databases [34].

In conclusion, our most recent survey demonstrated a 100% self-reported use of rectal NSAIDs administration among ERCPists. Significantly more ERCPists place PD stents than in 2013 and intensive IV hydration has come to the stage as a novel prophylactic measure in the clinical practice. In conjunction, this survey reveals the remaining improvements to be made. Recent guideline updates and active research participation have potentially led to this increased prophylaxis implementation.