Patient-Related Factors

The patient-specific risk factors that have been associated with higher rates of PEP on multivariate analyses include younger age (<60 years), female gender, history of previous PEP, nondilated ducts, normal bilirubin level, and suspected sphincter of Oddi dysfunction (SOD). On the other hand, chronic pancreatitis, particularly chronic calcific pancreatitis and pancreatic malignancy, seem to be protective against PEP, likely because of decreased exocrine enzymatic activity and atrophy of the upstream pancreatic parenchyma from chronic obstruction. Presence of pancreas divisum (unless dorsal duct cannulation is attempted), periampullary diverticulum, Billroth II gastrectomy, allergy to contrast media, and biliary interventions in patients with a preexisting biliary sphincterotomy are not associated with an increased risk of PEP.

Patients with suspected SOD, particularly women, not only carry the highest risk for PEP, ranging from 10% to 40%, but are also more likely to develop severe pancreatitis and death. In the landmark study by Freeman and colleagues that included 2347 patients, PEP occurred in 52 of 272 (19.1%) patients with suspected SOD, and was found to be the most powerful risk factor for PEP (odds ratio [OR] 5.01, 95% confidence interval [CI] 2.73–9.22) on multivariate analysis. Although the exact reasons for such high risk remain unclear, the risk is similar with diagnostic, manometric, or therapeutic ERCP in these patients.

The risk of PEP appears to be higher among women, but this finding is confounded by the presence of SOD, which occurs much more frequently in women. In the afore mentioned study by Freeman and colleagues, female gender was significant on univariate analysis but not on multivariate analysis. However, a subsequent prospective, multicenter study by Freeman and colleagues, including 1963 ERCP procedures, found female gender to be a significant risk factor on both univariate and multivariate analyses (OR 2.51, 95% CI 1.49–4.24; P = .0001). Female gender was also shown to be a significant risk factor for PEP (relative risk [RR] 2.23, 95% CI 1.75–2.84; P <.001) along with suspected SOD (RR 4.09, 95% CI 3.37–4.96) and previous pancreatitis (RR 2.46, 95% CI 1.93–3.12) in a meta-analysis of 15 prospective clinical studies with more than 10,000 patients.

Procedure-Related Factors

Difficult cannulation, multiple pancreatic injections, pancreatic duct instrumentation, pancreatic sphincterotomy including minor papilla sphincterotomy, pancreatic sampling, and balloon dilation of an intact biliary sphincter have been identified as procedure-related factors that independently increase the risks of PEP.

Papillary trauma induced by multiple attempts at cannulation appears to be a risk factor independent of the number of contrast injections into the pancreatic duct. In a large, prospective, multicenter study including 2347 patients, Freeman and colleagues found difficultly in cannulation to be a significant risk factor for PEP (adjusted OR 2.40, 95% CI 1.07–5.36). The rates of PEP were higher with difficult (>15 attempts) or moderately difficult cannulation (6–15 attempts) in comparison with easy cannulation (≤5 attempts). The rates of PEP were 13%, 7%, and 3%, respectively. In addition, the extent of pancreatic ductal opacification has been shown to correlate with the frequency of PEP. A retrospective analysis of 14,331 ERCPs showed that patients with opacification of head only (n = 845) had significantly lower incidence of PEP (3.6% vs 8.6%; P <.001) compared with those who had opacification to the tail of the pancreas (n = 4686). However, acinarization of the pancreas may be less important than previously thought, and has not been found to be significant in large studies.

Balloon dilation of an intact biliary sphincter has been associated with a markedly increased risk of PEP. In one large randomized, controlled multicenter study of 117 patients assigned to dilation and 120 to sphincterotomy, there was a significantly increased rate of PEP in the balloon dilation group when compared with the sphincterotomy group (15.4% vs 0.8%), along with 2 deaths (1.7%) from pancreatitis following dilation, leading to termination of the study at the first interim analysis. A subsequent Cochrane systematic review including 15 randomized trials (1768 participants) confirmed a higher risk of PEP, associated with endoscopic balloon dilation, in comparison with endoscopic sphincterotomy (RR 1.96, 95% CI 1.34–2.89). However, balloon dilation after biliary sphincterotomy does not appear to increase the risk of PEP, and therefore is the preferred approach for extraction of large stones, particularly in patients with coagulopathy.

Although patients with SOD have higher rates of PEP, contrary to widely held opinion, sphincter of Oddi manometry in itself does not carry a significant additional risk. With the widespread use of aspiration instead of conventional perfusion manometry catheters, the risk of hydrostatic injury and PEP with biliary and pancreatic sphincter manometry has probably been reduced to that of cannulation with any other ERCP accessory. In an early randomized trial including 76 patients, sphincter of Oddi manometry using the aspiration catheter was associated with a significantly decreased frequency of PEP (23.5% vs 3%, P = .01), reduced mean hospital stay (5 ± 1.83 days; P = .03), and milder pancreatitis, compared with the standard perfusion system. Studies comparing the risk of PEP in patients having ERCP for suspected SOD with and without sphincter of Oddi manometry have shown no independent risk of manometry in addition to the high risk conferred from underlying SOD itself. In a large retrospective analysis of 11,497 ERCPs performed over 12 years, neither pancreatic (OR 1.43, 95% CI 0.99–2.08) nor biliary manometry (1.16, 95% CI 0.83–1.62) was found to be a statistically significant risk factor for PEP.

Data regarding the risk of PEP associated with the use of biliary self-expanding metal stents (SEMS) is conflicting, with rates ranging from 0% to 9%. Theoretically the large diameter and radial expansion of the SEMS can obstruct or distort the pancreatic orifice or common channel, leading to PEP. In a retrospective study of 544 patients with malignant biliary obstruction, the frequency of PEP was significantly higher in the SEMS group than in the plastic stent group (7.3% vs 1.3%, respectively; OR 5.7, 95% CI 1.9–17.1). The frequency of PEP was similar between covered (6.9%) and uncovered (7.5%) SEMS (OR 0.9, 95% CI 0.3–2.4). In another retrospective study, pancreatitis following SEMS insertion was observed in 22 of 370 (6%) patients, and multivariate analysis revealed that SEMS with high axial force (OR 3.69; P = .022) and nonpancreatic cancer (OR 5.52; P <.001) were significant risk factors for pancreatitis.

The osmolality of the radiologic contrast material used appears to have no association with the risk of PEP. Studies comparing low-osmolality with high-osmolality contrast media during ERCP show no significant difference in the rate of PEP. Similarly, the benefit of pure cutting current over blended current by reduction of thermal injury during endoscopic sphincterotomy and ampullectomy is controversial.

Operator-Related Factors

Operator-related risk factors, such as trainee participation and the case volume and experience of the endoscopist, have been suggested to independently contribute to the risk of PEP, but have been difficult to evaluate. Trainee involvement has been shown to increase the risk of PEP in one study (OR 1.5, 95% CI 1.029–2.057; P = .03), which presumably results from traumatic cannulation, prolonging a difficult cannulation, or delivering excess electrosurgical current during sphincterotomy.

Endoscopist case volume is thought to be inversely proportional to the risk of PEP, although most multicenter studies have failed to show this trend. In a prospective multicenter study by Freeman and colleagues, low-volume endoscopists (performing <2 ERCPs per week) had significantly lower success at bile duct cannulation (96.5% vs 91.5%; P = .0001), but lower ERCP case volume was not found to be a multivariate risk factor for PEP. Similarly, in a more recent Austrian study, although high case volume (endoscopists performing >50 vs <50 ERCPs per year) was associated with significantly higher success rates (86.9% vs 80.3%; P <.001) and lower overall complication rates (10.2% vs 13.6%; P = .007), and there was no significant difference in rates of PEP (4.9% vs 5.6%; P >.05). This result likely arises because a much higher threshold of case volume is needed to show the impact and the difference in case mix, as low-volume endoscopists tend to perform lower-risk cases.

Although it is difficult to understand the relative contribution and the interactive effect of multiple risk factors, the risk seems to be more than just additive, and a combination of factors can escalate the risk substantially. In one study, the risk of PEP increased from 5% in women with normal serum bilirubin level to 16% with addition of difficult cannulation, with further increase in risk to 42% with addition of suspected SOD as the indication for ERCP. Therefore, careful patient selection, sound endoscopic technique, and tailoring the approach of ERCP to the individual patient are the cornerstones in the prevention of PEP.

Guide-Wire Cannulation

Cannulation techniques are recognized to be important in causing PEP. The use of guide-wire–assisted cannulation instead of contrast-guided cannulation to avoid or minimize contrast injection has been shown in some studies to decrease PEP. Contrast-free guide-wire–assisted cannulation is performed either by leading with the guide wire, or inserting the catheter into the papillary orifice and advancing the guide wire without contrast injection. A recent meta-analysis of 12 randomized controlled trials (3450 patients) showed a significant reduction in PEP with a guide-wire–assisted cannulation technique compared with contrast-guided biliary cannulation, with a risk ratio of 0.51 (95% CI 0.32–0.82). In addition, the guide-wire–assisted cannulation technique was associated with greater primary cannulation success (risk ratio 1.07, 95% CI 1.00–1.15), fewer precut sphincterotomies (risk ratio 0.75, 95% CI 0.60–0.95), and no increase in other ERCP-related complications. However, some endoscopists, including the current authors, believe that cannulation without a guide wire is reasonably safe in highly skilled hands and is not associated with increased risks of PEP in comparison with guide-wire cannulation. Many advanced endoscopists use a hybrid of the 2 techniques, using minimal contrast to outline the course of the distal ducts in combination with wire probes, which may avoid dissections or passage of the guide wire outside a side branch of the pancreatic duct. In addition, use of a soft-tipped wire for cannulation may also lower the risk of PEP.

Aggressive Hydration

The use of aggressive periprocedural hydration with lactated Ringer solution also appears to reduce the incidence of PEP. In a recent pilot study of 62 patients, the incidence of PEP was significantly lower (0% vs 17%) in patients randomized to aggressive hydration with lactated Ringer solution (3 mL/kg/h during the procedure, a 20 mL/kg bolus after the procedure, and 3 mL/kg/h for 8 hours after the procedure) compared with patients who received standard hydration with the same solution (1.5 mL/kg/h during and for 8 hours after procedure).

Techniques to Minimize Ampullary Trauma

Efficient and atraumatic technique is central to minimizing the risk of PEP. The risk of PEP can also be reduced by minimizing ampullary trauma by limiting cannulation attempts and using alternative ways of accessing the duct after prolonged cannulation, such as precut or access papillotomy. There has been controversy regarding whether precut techniques increase the risk of PEP, and if the increased risk of PEP is due to the precut papillotomy itself or the prolonged cannulation attempts that often precede its use. Comparative studies of precut papillotomy with standard sphincterotomy are difficult to interpret given the difference in indications, settings, experience of the endoscopists, and the frequent combination of precut sphincterotomy with pancreatic stenting.

However, there is some evidence to suggest that early precut access, after only 5 to 10 attempts for cannulation, may decrease the risk for PEP compared with persistent attempts at cannulation, which can traumatize the papilla further. Two previous meta-analyses that included 6 randomized controlled trials comparing early precut papillotomy with persistent attempts at cannulation by a standard approach showed significant reduction in PEP (OR 0.47, 95% CI 0.24–0.91; and rate ratio 0.46, 95% CI 0.23–0.92) when early precut access was obtained. A similar trend toward a lower risk of PEP with early use of precut sphincterotomy was also seen in 2 more recent meta-analyses (including 7 randomized controlled trials), although this was not statistically significant (3.9% in the precut sphincterotomy vs 6.1% in the persistent attempts group; OR 0.58, 95% CI 0.32–1.05). In addition, early precut strategy was recently shown to substantially reduce the duration of ERCP, which may be clinically beneficial considering the relatively high percentage of elderly and advanced-grade American Society of Anesthesiologists patients.

Chemoprophylaxis

Chemoprophylaxis of PEP has also been suggested and extensively researched in an attempt to prevent or reduce the severity of PEP. Five main targets for chemoprevention include: prevention of intra-acinar trypsinogen activation (protease inhibitors such as gabexate, ulinastatin, nafamostat mesylate) ; reduction of pancreatic enzyme secretion (somatostatin and octreotide) ; relaxation of sphincter of Oddi spasm (nitroglycerin, nifedipine, phosphodiesterase-5 inhibitors) ; interruption of the inflammatory cascade (nonsteroidal anti-inflammatory drugs [NSAIDs], interleukin-10, corticosteroids, allopurinol, heparin, N -acetylcysteine) ; and prevention of infection (antibiotics).

However, numerous trials studying a variety of such pharmacologic agents have yielded disappointing or conflicting results; with the important exception of rectal administration of NSAIDs, which have been shown in several randomized controlled trials and meta-analyses to significantly reduce the incidence and severity of PEP. In the largest multicenter, randomized, placebo-controlled, double-blind clinical trial, published in the New England Journal of Medicine , 602 patients were randomized to receive a single dose of 100 mg of rectal indomethacin or placebo immediately after ERCP. PEP developed in 9.2% of the indomethacin group and in 16.9% of the placebo group ( P = .005). Moderate to severe pancreatitis developed in 13 patients (4.4%) in the indomethacin group and in 27 patients (8.8%) in the placebo group ( P = .03). A recent meta-analysis of 10 randomized controlled trials involving 2269 patients showed that NSAID use decreased the overall incidence of PEP (risk ratio 0.57, 95% CI 0.38–0.86; P = .007). Of 10 randomized controlled trials included, indomethacin was administered in 4 studies (rectal route in 3 studies and intraduodenal in 1 study), diclofenac in 5 studies (rectal in 3, oral in 1, and intramuscular in 1), and intravenous valdecoxib was used in 1 study. The pooled number needed to treat was 17. The prophylactic use of NSAIDs was also efficacious in decreasing the incidence of moderate to severe PEP (risk ratio 0.46, 95% CI 0.28–0.75; P = .002). There was no significant difference in the adverse events attributable to NSAIDs.

The mechanism of action of NSAIDs is via potent inhibition of cyclooxygenase, lipoxygenase, and phospholipase-A2 (PLA2) mediated pathways. PLA2 catalyzes the hydrolysis of cell membrane phospholipids, leading to the production of numerous inflammatory mediators such as prostanoids, leukotrienes, kinins, and platelet-activating factor, and is believed to play a critical role in the initial inflammatory cascade in acute pancreatitis. Indomethacin and diclofenac are the mostly widely studied and used NSAIDs with similar efficacy in PEP chemoprophylaxis. The rectal route in preference to other routes of administration, and administration before (in patients with high preprocedure risk) or immediately after (for patients who move into the high-risk group only during ERCP) the procedure seem to be most effective in preventing PEP. Based on these data, a good safety profile, and very low cost with one-time dosing, there has been more widespread use of NSAIDs (mainly rectal indomethacin) in patients undergoing ERCP for PEP prevention.

Combination of NSAIDs with other pharmacologic agents or ERCP techniques has also recently been studied in an attempt to achieve higher prophylactic efficacy from synergistic action. A double-blind, placebo-controlled, randomized trial (N = 300) found the combination of rectal indomethacin (100 mg) and sublingual nitrate (5 mg) given before ERCP to be significantly more likely to reduce the incidence of PEP than indomethacin suppository alone (6.7% vs15.3%, respectively; P = .016, risk ratio 0.39, 95% CI 0.18–0.86).

Although mechanistically plausible that stenting and rectal NSAIDs may complement one another by working in completely different ways, a recent network meta-analysis showed that combination of rectal NSAIDs and stents was not superior to either approach alone (NSAIDs plus stents vs NSAIDs alone: OR 1.46, 95% CI 0.79–2.69; NSAIDs plus stents vs stents alone: OR 0.70, 95% CI 0.40–1.20). Post hoc analysis by Elmunzer and colleagues also found rectal indomethacin alone to be more effective for preventing PEP than the combination of indomethacin and pancreatic stents.

Pancreatic Duct Stenting

Pancreatic duct stenting is another intervention that is increasingly used, as it has been shown to reduce the incidence and severity of PEP by mechanically facilitating pancreatic duct drainage by relieving ductal hypertension that develops as a result of transient procedure–induced stenosis of the pancreatic orifice. Several randomized controlled trials and meta-analyses have shown a 60% to 80% reduction in the incidence of pancreatitis with placement of a pancreatic duct stent in high-risk patients. A recent meta-analysis, which included 14 randomized controlled trials involving 1541 patients, indeed confirms that placement of prophylactic pancreatic stents is associated with a statistically significant reduction in PEP (RR 0.39, 95% CI 0.29–0.53; P <.001). Subgroup analysis stratified according to the severity of PEP showed that pancreatic stenting was beneficial in patients with mild to moderate PEP (RR 0.45) and in patients with severe PEP (RR 0.26). In addition, subgroup analysis performed according to patient selection demonstrated that stent placement was effective for both high-risk and mixed case groups.

Prophylactic pancreatic stenting is recommended in most patients with difficult cannulation, including double-wire cannulation (whereby a guide wire is left in the pancreatic duct to aid wire cannulation of the common bile duct) and precut sphincterotomy, pancreatic (major or minor) sphincterotomy, pancreatic endotherapy, diagnostic or therapeutic ERCP for suspected or confirmed SOD, history of PEP, balloon dilation of an intact biliary sphincter, and endoscopic ampullectomy.

However, there are a few considerations that the endoscopist must keep in mind when performing pancreatic duct stenting. First, pancreatic duct stenting can be technically challenging and potentially dangerous if attempted but unsuccessful, as it is associated with a high rate of PEP. Attempts at pancreatic duct cannulation may cause increased pancreatic orifice trauma without providing ductal decompression. In addition, pancreatic stents may cause pancreatic ductal or parenchymal injury, with a risk of permanent ductal stenosis and chronic pancreatitis changes over time. Placement of pancreatic stents made of softer materials and the use of smaller-caliber stents (3F or 4F) have been shown to lower the rates of ductal injury in comparison with conventional 5F polyethylene stents; however, 3F pancreatic stents were found to be inferior to the 5F stents for the prevention of PEP in high-risk patients in a recent network meta-analysis.

These prophylactic pancreatic duct stents are not meant to be left in indefinitely, and follow-up abdominal radiography (AXR) must be done to ensure that they have fallen out in a timely fashion (typically within 2–3 weeks). Therefore, when performing pancreatic duct stenting, consideration should also be given to the increased cost and time associated with follow-up AXR to ensure spontaneous passage of the stent, and additional upper endoscopy to remove retained stents in 5% to 10% of patients.

Despite these considerations and limited indirect evidence to suggest that NSAIDs are superior to stents for PEP prophylaxis, pancreatic stents are routinely placed in all high-risk cases, as prophylactic pancreatic duct stenting has repeatedly been shown to be effective in preventing and reducing the incidence of severe, life-threatening PEP.