INTRODUCTION

Endoscopic retrograde cholangiopancreatography (ERCP) is a diagnostic and therapeutic technique commonly used for biliary and pancreatic conditions. The most common adverse event is post-ERCP acute pancreatitis (PEAP), which may develop in 2% to 9% of patients according to various reported series ^1,2. This may represent a serious complication with significant morbidity and also mortality ³.

Patient-related factors associated with a higher potential and severity of PEAP include age below 60 years, normal serum bilirubin and sphincter of Oddi dysfunction ^4,5. Technique-related factors have also been reported, including repeated and traumatic papillary cannulation, pancreatic sphincterotomy, biliary dilation without prior sphincterotomy, biliary sphincterotomy and precut ^6,8. Protective factors include advanced age, malignant stricture and chronic pancreatitis ⁵.

PEAP development is a multifactorial process that included damage from mechanical, thermal, chemical, hydrostatic, enzymatic and microbiological causes. Therefore, several ways of reducing PEAP risk have been studied. The most promising results were obtained with the insertion of a pancreatic stent. However, stent placement may be challenging and is not exempt from complications ^9,10. Various drugs have also been used to prevent PEAP from developing. Nonsteroidal anti-inflammatory drugs (NSAIDs) inhibit several mediators in the inflammatory cascade that play an important role in the pathogenesis of acute pancreatitis, primarily prostaglandins and phospholipase A-2. These drugs, when rectally administered, have been shown to reduce PEAP rates in a number of studies and meta-analyses. This beneficial effect is particularly seen in high-risk patients, including patients with sphincter of Oddi dysfunction ^11,12,13,14. However, conflicting results were obtained in patients with a moderate risk for PEAP development.

A cohort study was used to assess the effect of rectal diclofenac that was administered immediately before ERCP on the incidence and severity of PEAP in consecutive patients, regardless of their risk for pancreatitis.

MATERIAL AND METHODS

The study assessed a series of patients who underwent ERCP from January 1st 2009 to July 31st 2016. It was a mixed cohort study with retrospective data collected from January 1st 2009 to May 31st 2012, and prospective data collected from May 31st 2012 until July 31st 2016. A total of 1,512 ERCP procedures were carried out. The desired duct could not be cannulated but the papilla of Vater was manipulated in 93 cases. Hence, these patients were also included in the study. ERCPs were performed by two widely experienced physicians. The study was approved by the hospital research committee. All patients signed an informed consent form.

PEAP was defined according to consensus criteria including the presence of abdominal pain consistent with pancreatitis and amylase levels at least three times the upper normal limit 24 hours after the procedure ¹⁸. PEAP severity was defined as mild (less than 3 days in hospital), moderate (four to ten days in hospital) or severe (over ten days in hospital, development of necrosis or pseudocyst) ¹⁵. All patients remained in hospital for at least 24 hours after the ERCP.

Patients were compared according to whether they received diclofenac or not. An analysis of the factors associated with PEAP development was performed. These included both patient-related (age, sex, history of recurrent acute pancreatitis, prior PEAP, chronic pancreatitis, bilirubin, choledocholithiasis and malignant stricture) and technique-related parameters (biliary sphincterotomy, pancreatic sphincterotomy, precut, Wirsung cannulation, cannulation length above ten minutes, papillary dilation without sphincterotomy, biliary stent, Wirsung stent and sedation by anesthetist).

RESULTS

A total of 1,512 patients were included; 718 did not receive diclofenac and 794 received 100 mg of diclofenac that was rectally administered at the start of the procedure. Fifty-seven patients that underwent ERCP after June 1st 2012 had a contraindication to NSAIDs and were included in the non-diclofenac group. The mean age was 73 years (SD 13.86); 818 were male (54.1%) and 694 were female (45.9%).

Table 1 shows the baseline characteristics of both groups. More patients were included in the diclofenac group and significant differences were associated with choledochal stones, Wirsung cannulation, biliary sphincterotomy, male sex and sedation by an anesthetist. More biliary stents were inserted in the non-diclofenac group.

Table 1 Baseline characteristics of patients who received rectal diclofenac as compared to patients who did not 

AP: acute pancreatitis; PEAP: post-ERCP acute pancreatitis; ERCP: endoscopic retrograde cholangiopancreatography. Data are expressed as n (%).

PEAP developed in 47 patients (3.1%); 3.4% (27/47 PEAPs) of patients treated with diclofenac and 2.8% (20/47) of diclofenac-naive subjects (p = 0.554). In the diclofenac group, PEAP developed in 4.4%, 0.5% and 2.6% of subjects with intact papillae, prior sphincterotomy and extended sphincterotomy, respectively. Similar results were obtained in the non-diclofenac group: 4% with intact papillae, 0.9% with prior sphincterotomy and 2.5% with extended sphincterotomy. Only six patients had a suspected sphincter of Oddi dysfunction: four in the diclofenac group and two in the control group. There was no reported PEAP in these patients.

Plastic stents were placed in the pancreatic duct of 68 patients (4.5%), 5.3% in the diclofenac group and 3.6% in the control group (p = 0.136). PEAP was mild in 78.3% of cases (37 patients). PEAP-related mortality occurred in four subjects, two in each group. PEAP severity was similar between both groups and no significant differences were found.

Table 2 shows the results of the bivariate analysis related to the influence of the various factors on the likelihood of PEAP development. No differences were found according to diclofenac administration status. A protective effect for PEAP development was reported with regard to the presence of choledocholithiasis (OR 0.43, 95% CI: 0.23-0.80) and age above 75 years (OR 0.51, 95% CI: 0.27-0.96). Wirsung stent placement had no influence (OR 1.46, 95% CI: 0.44-4.85, p = 0.52). With regard to technique-related factors, PEAP risk was increased by pancreatic duct cannulation. The risk was 3.54 fold higher (95% CI: 1.39-8.99) for guidewire cannulation and 6.05 fold higher (95% CI: 3.04-12.03) when a guidewire and contrast agent were used simultaneously. Pancreatic sphincterotomy increased the risk by a factor of 16.25 (95% CI: 1.44-182.92), cannulation longer than ten minutes increased the risk by a factor of 2.79 (95% CI: 1.55-5.02), biliary sphincterotomy increased the risk by a factor of 1.14 (95% CI: 0.47-2.75), and biliary dilation without sphincterotomy increased the risk by a factor of 31.8 (95% CI: 1.95-517.39). However, prior sphincterotomy had a protective effect with an OR of 0.19 (95% CI: 0.04-0.78). Furthermore, extended sphincterotomy had an OR of 0.69 (95% CI: 0.13-3.49) (Table 2).

Table 2 Results of the bivariate analysis that relate the influence of various factors to the likelihood of developing post-ERCP acute pancreatitis (PEAP) 

AP: acute pancreatitis; PEAP: post-ERCP acute pancreatitis; ERCP: endoscopic retrograde cholangiopancreatography. *mean ± standard deviation.

According to the logistic regression analysis, the only variables that were significantly associated with PEAP were Wirsung cannulation with a guidewire, contrast or both; OR 2.77 (95% CI: 0.86-8.96), OR 3.60 (95% CI: 1.23-10.51) and OR 5.61 (95% CI: 2.39-13.18), respectively. As well as dilation without sphincterotomy (OR 401.52, 95% CI: 169.63-950.40). Age younger than 75 years increased the risk by a factor of 2.35 (95% CI: 1.14-4.83). Diclofenac was not associated with risk in our cohort (Table 3).

Table 3 Significant results of the logistic regression analysis with patient and ERCP related variables that potentially represent confounders for the development of post-ERCP pancreatitis 

ERCP: endoscopic retrograde cholangiopancreatography.

The cohort included a high-risk group of 394 (26.1%) patients. In this group, 183 subjects received diclofenac and 185 did not. The PEAP incidence was 8% (16 patients) in the diclofenac group and 5.1% (ten patients) in the control group. No significant differences were found between both groups (p = 0.311).

Diclofenac use did not induce more adverse effects. Eleven bleeding events were reported (1.4%) in the diclofenac group versus eight (1.1%) in the non-diclofenac group (p = 0.65). All bleeding events developed after sphincterotomy and no cases required admission to the Intensive Care Unit. Ten perforations occurred, 1% in the diclofenac group and 0.3% in the control group (p = 0.11).

DISCUSSION

NSAIDs were seen to be beneficial for PEAP prevention in randomized, controlled studies, as well as in several meta-analyses ^11,12,16,17. Such results led the European Society of Gastrointestinal Endoscopy (ESGE) to recommend the use of indomethacin or diclofenac either before or after ERCP procedures from 2010 in their guidelines ^18,19. In 2012, Elmuntzer et al. ¹² reported a multicenter, randomized, controlled study that found a significant decrease in PEAP risk among high-risk patients that received rectal indomethacin after the procedure, even though over 80% of subjects had a pancreatic stent.

In this study, a similar number of PEAPs in the group treated with diclofenac (3.4%) and in the non-diclofenac group (2.8%) was found. There was in fact a tendency towards a higher number of PEAPs in the diclofenac group (27 patients versus 20), with an OR of 1.22 (95% CI: 0.68-2.21, p = 0.49). PEAP severity was similar in both groups. Therefore, our results differ with the above report. Thiruvengadam et al. ²⁰ reported a retrospective study of a cohort of 4,017 patients, both high-risk and low-risk, who underwent ERCP. PEAP risk among the patients treated with indomethacin decreased by 65% (OR 0.35, 95% CI: 0.24-0.51, p < 0.001). Furthermore, PEAP severity was also lower for patients treated with indomethacin (a risk reduction of 83%). Finally, this study also found that PEAP risk had decreased by 64% in the group of patients with malignant strictures who received indomethacin. However, we found no such reduction in PEAP risk among patients with malignant strictures in our study (OR 1.14, 95% CI: 0.52-2.50).

Our results are in line with a prospective, randomized study of post-ERCP indomethacin in 449 patients by Levenick et al. ²¹. No drug-related efficacy with regard to the number of PEAP events was found. However, a non-significant tendency towards more PEAP episodes was found in the treated group. We cannot explain why there was a greater tendency towards pancreatitis in patients that received diclofenac. In the logistic regression analysis, we also saw that the only ERCP-related factors associated with PEAP development were pancreatic duct manipulation, whether using a guidewire, a contrast agent or both, as well as papillary dilation without sphincterotomy.

When only patients at high risk for PEAP were assessed, no differences between both groups were found. In the study by Thiruvengadam et al. ²⁰, only 10% of patients could be included in the high-risk group. The low-risk population in this study was larger than ours. Our patient distribution was more like that of the study by Levenink et al. ²¹, where around 30% of patients were high-risk. Our cohort included a high-risk group of 394 (26.1%) patients. The incidence of PEAP was 8% (16/183 patients) in the diclofenac group and 5.1% (10/185 patients) in the non-diclofenac group. No significant differences were seen between both groups (p = 0.311) and the tendency towards more PEAP events among treated patients persisted.

A strength of our study was the fact that only two endoscopists performed all ERCP procedures in a group of non-selected, consecutive patients with no relevant differences in the technique. A weakness was the mixed cohort nature with a historic retrospective cohort as some data might have been lost, which was compared to a prospective cohort.

To conclude, rectal diclofenac before ERCP was not useful to decrease the number and severity of pancreatitis events in our patient population, which included both low-risk and high-risk subjects.