Post-ERCP Pancreatitis: New Momentum

Two new papers about post-endoscopic retrograde cholangiopancreatography (ERCP) pancreatitis appear this month in Endoscopy [1] [2], and a third was published at the end of last year [3]. Studies on post-ERCP pancreatitis seem to be a never-ending story, but is there really so much news? Or does the number of publications reflect, to some extent, a sustained uncertainty in our understanding of this common complication? Probably not, since insights into post-ERCP pancreatitis have grown significantly within the last ten years. The pancreatic gland is a very sensitive organ which needs to be approached delicately, and which has unpredictable and sometimes volatile reactions, with big effects arising from small causes. It is used with increasing freedom in endoscopic therapy compared with interventions in the bile duct. The new indications for ERCP, however, seem to increase the overall risk of pancreatitis, and renewed awareness of the problem is a logical consequence. The discussion about post-ERCP pancreatitis has gained new momentum.

There is a plethora of convincing data about risk factors for post-ERCP pancreatitis. Some of these factors can be assessed before ERCP is performed; others depend on the ERCP procedure itself. The paper of Christoforidis and colleagues summarizes experience with ERCP at a single center [1]. The aim of the study was to investigate predictive symptoms and risk factors for post-ERCP pancreatitis or hyperamylasemia. Within a period of five years one endoscopist performed about 500 procedures, and the overall results are widely representative for the method. Four significant risk factors were identified, all of them in agreement with the results of prior reports [4] [5] [6] [7] [8] [9]: younger age (< 50 years), history of previous or relapsing pancreatitis, difficult cannulation, and pancreatic visualization. Additionally, female sex increased the incidence of pancreatitis and hyperamylasemia, the increase remaining just below the level of significance. The impact of pancreatic duct visualization was investigated using a grading system as suggested in previous recommendations [10]. Grade A represented visualization of the main pancreatic duct only; grade B included the visualization of the primary branches of the pancreatic duct; and in grade C, visualization of the secondary branches or acinarization of the pancreatic gland was present. The incidence of pancreatitis rose significantly with the grade of visualization, being 1.6 % in grade A, 9.3 % in grade B, and 33.3 % in grade C. Because of the low number of patients with grade C, only grade A and B subgroups were statistically compared, and the odds ratio for pancreatitis was 8.2 with grade B procedures. This is an important finding which should remind us of the initial experiences with ERCP and endoscopic sphincterotomy (EST), when endoscopists tried to strictly avoid unintended pancreatic contrast injections. Nowadays ERP is a routine procedure and we have possibly lost some of our former caution, but, as the study of Christoforidis shows, a significant part of the risk of post-ERCP pancreatitis lies in our hands - or in the hands of the endoscopy nurse who performs the contrast injections. We know that an unintended injection of only 2 ml of contrast may lead to complete visualization of the main pancreatic duct and the secondary branches [10]. Thus, awareness of the problem, mindfulness of the whole ERCP team, and continuous fluoroscopic control during all injections are the only possible means of avoiding an unintended overfilling of the pancreatic duct.

The dominant risk factor for post-ERCP pancreatitis in North America, however, is suspected sphincter of Oddi disease (SOD). This risk factor was almost absent in the study of Christoforidis et al. [1], and it has been an infrequent problem in other European reports [6] [7] [8] [9]. The paper of Fogel et al. in the current issue of Endoscopy deals with two risk factors for post-ERCP pancreatitis, i. e. SOD and pancreatic sphincterotomy [2]. The aim of the study was not to describe the risk of pancreatitis, but to analyse procedural modalities for the prevention of post-ERCP pancreatitis. The patients who were included must be regarded as high-risk candidates for pancreatitis. Indications were chronic abdominal pain, acute or recurrent pancreatitis, and chronic pancreatitis. All the patients underwent biliary and pancreatic sphincter of Oddi manometry, and patients with abnormalities revealed in the manometry studies received biliary or biliopancreatic sphincterotomy. The study shows several points of weakness, which should be considered. The results are based on a retrospective database query regarding patients who had undergone manometry studies. The methods of therapeutic measurement did not follow a fixed protocol. Around 1995 the Indianapolis group changed their treatment approach for SOD from biliary to biliopancreatic sphincterotomy, and the hypothesis of the paper is that this change in therapy decreased the incidence of post-procedural pancreatitis. Two historic patient cohorts were compared for that purpose. Unfortunately an important confounding variable was added to biliopancreatic sphincterotomy, namely pre- or post-sphincterotomy pancreatic stenting, so that it is not possible to attribute the results to a particular modality. Additionally, there was no randomization, and the endoscopists decided individually about the therapy indicated on the basis of clinical experience, rather than using the objective criteria of the underlying indication for sphincterotomy or the results of pancreatic manometry studies. Suspicion of SOD is probably the most important risk factor for post-ERCP pancreatitis, but none of the known risk factors described above were considered. Thus nothing is said about the distribution of additional confounding variables within the different treatment groups. All in all, the statistical comparison must be flawed. Nevertheless, the results are of interest for endoscopists: in patients with manometrically proven SOD, biliary sphincterotomy was associated with a 28.3 % incidence of pancreatitis, for pull-type biliopancreatic sphincterotomy followed by pancreatic stenting the incidence was 19.2 %, and for primary pancreatic stenting followed by needle-knife biliopancreatic sphincterotomy it was 10.7 %. The latter incidence was comparable to that in the group of patients with normal manometry results and without subsequent endoscopic therapy [2]. Those figures really look like an improvement of therapy in SOD-patients, and they underline the possibility of a high risk of pancreatitis after diagnostic or therapeutic ERCP with the indication of suspected SOD. But what are the conclusions based on rigorous statistical analysis?

First, the hypothesis of the study remains unproven, on the basis of the study design. A randomized investigation with stratification for prophylactic pancreatic stenting is needed to clarify whether biliopancreatic sphincterotomy in patients with SOD truly has a lower pancreatitis rate compared with biliary sphincterotomy. It is possible that biliopancreatic sphincterotomy provides a better symptomatic outcome in the treatment of SOD with pancreatic sphincter hypertension, but the authors of the study did not demonstrate such outcome data.

Second, the following questions need an answer. Who should perform ERCP and endoscopic sphincterotomy? Is SOD an indication for “experts only”? Do we need a licence for special interventions? The Indianapolis group showed their individual experience in sphincterotomy, especially in the risky group of SOD patients [2]. What would happen, if endoscopists with less experience or less ERCP practice began to treat SOD and tried to use the needle-knife for biliopancreatic sphincterotomy? Large North American multicenter studies of EST [4] and ERCP [5], an Italian multicenter study of ERCP [6], and single-center sphincterotomy studies from Germany [8] [11] have demonstrated that there is a significant variation in the risk of complications, especially pancreatitis, depending on the number of procedures performed by the endoscopist or at the endoscopy center. In the latter three studies SOD had almost no effect. This demonstrates that the “risk factor endoscopist” is a unique problem associated with ERCP and not only a challenge in SOD patients.

Third, the efficacy of procedural anticipation of post-ERCP pancreatitis is now almost proven. Pancreatic drainage seems to be helpful for decreasing the incidence of pancreatitis in high-risk patients. Fogel and co-authors point to a collection of randomized controlled trials about prophylactic pancreatic stenting [2], and several new studies could be added which show the benefit of short-term pancreatic stenting after ERCP [12] [13]. A final prospective multicenter study, which considered not only the benefit but also the practicability and economic consequences, would be helpful in the establishment of prophylactic pancreatic stenting and for adding this procedure to current recommendations.

From the results of prophylactic pancreatic stenting we can deduct that impairment of pancreatic drainage can lead to pancreatitis. Papillary edema after transpapillary procedures is the most probable cause of transient pancreatic obstruction. The observation that ERCP with biliary sphincterotomy in non-SOD patients was associated with a lower incidence of pancreatitis compared with ERCP without EST [1], or with decreased severity [14], and the fact that difficult or failed cannulation with increased mechanical papillary trauma increases the risk of pancreatitis [4] [5] underline the possible pathophysiological implication of papillary edema. The second major etiological factor for post-ERCP pancreatitis is also to some extent a mechanical one. Several studies have shown that the number of pancreatic cannulations, the amount of pancreatic contrast injection and the grade of pancreatic visualization also influence the incidence of pancreatitis [1] [4] [5]. The mechanical trauma during deep pancreatic cannulation and the injection pressure during visualization contribute to ductal epithelial or acinar injury, and this may lead to a backflow of pancreatic fluid or an influx of contrast into the interstitial space [15] [16]. Pancreatic drainage may therefore also be helpful during the healing of ERCP wounds, as it has also been successfully used previously for the therapy of traumatic pancreatic duct rupture [17] [18] [19] [20]. It would be interesting to analyse the effect of prophylactic pancreatic stenting according to grades of pancreatic contrast exposure and according to a score of cannulation trauma. It is possible that such a study might not only demonstrate that prophylactic pancreatic drainage decreases the incidence of pancreatitis, but it might also clarify the question of whether papillary edema and pancreatic contrast exposure are independent risk factors for ERCP-associated pancreatitis or whether they depend on each other.

Whenever pancreatic injury is induced during ERCP it is generally supposed to be difficult to predict the subsequent course of the patient. There is a high incidence of hyperamylasemia without further consequences. A certain group of patients has pain but no signs of pancreatitis, and a third group of patients, usually between 2 % and 10 %, develop pancreatitis. The currently employed definition of post-ERCP pancreatitis to a certain extent subserves the routine management of patients after ERCP [21] [22], i. e. discharge or admission. According to this definition, most patients have a mild or moderate course with an uneventful recovery within a few days of symptomatic therapy. Post-ERCP pancreatitis generally does not fulfil the criteria for severe pancreatitis, which are associated with local or systemic complications induced by mobilization and activation of inflammatory mediators, chemokines and pro-inflammatory cytokines. At present, routine laboratory investigations cannot provide an early prediction of severe pancreatitis, but studies of inflammatory mediators can do so within a few hours after ERCP [23]. The severity of post-ERCP pancreatitis, therefore, seems to be the result of very early inflammation events which appear during or immediately after the procedure.

The medical mitigation of the inflammatory response to ERCP-induced injury or of the systemic consequences of this reaction has been the frustrating topic of many studies. Some current reports, including a prospective, randomized, placebo-controlled trial of prednisone and allopurinol published in Endoscopy [3], and a prospective, randomized, controlled, multicenter trial investigating octreotide [24], have taken their place in a long list of studies about drugs with no effects on post-ERCP pancreatitis. Four new reports with promising results, however, have also appeared [8] [25] [26] [27], but scepticism is still recommended, given the experience with other drugs, where equivocal or contradictory results have been obtained after initial enthusiastic reports had been published.

A prospective, randomized, double-blind, placebo-controlled trial showed a significant reduction of the incidence of post-ERCP pancreatitis, from 18 % to 8 %, with the help of glyceryl trinitrate [25]. The effect of this drug is supposed to be a prolonged relaxation of the sphincter of Oddi with an improvement in ease of cannulation and consequent reduction in cannulation trauma. A 50 % incidence of systolic blood pressure below 100 mm Hg with a need for intravenous infusion of crystalloids, however, is disturbing and probably impairs the practicability of the drug in routine cases. On the other hand, there is evidence of a certain risk of ischemic heart events during ERCP [28], and nitrates could be beneficial in avoiding this. In another excellently designed study (single-center, double-blind, randomized, and placebo-controlled), interleukin-10 (IL-10) significantly reduced the incidence of post-ERCP pancreatitis from 24.4 % to 8.7 % [26]. Additionally, there was a significant dose-effect relation within the treatment group (4 µg/kg, 10.4 %; 20 µg/kg, 6.8 %) but, compared with placebo, no significant differences were obtained according to lipase, C-reactive protein (CRP), IL-6, IL-8, or tumor necrosis factor (TNF) levels up to 48 hours after ERCP. Points of weakness in this study are the relatively small number of patients providing data for analysis (n = 137), and an imbalance of confounding risk factors between the placebo and the treatment group (pancreatic sphincterotomy 28.9 % vs. 13.0 %, precut 11.1 % vs. 3.3 %; P = 0.07). Although IL-10 administration seemed to be an independent risk factor for post-ERCP pancreatitis in a logistic regression analysis of all risk factors (odds ratio 0.46, P = 0.039), and the authors stated in a letter that the P value remained significant after the Bonferrroni correction had been made [29], there were too many variables in the model for the number of patients and the number of events available, and the confirmative power of this P value appears questionable. Interestingly, another prospective, randomized, double-blind study did not demonstrate a reduction of post-ERCP pancreatitis when IL-10 8 µg/kg was used [30]. In a study by our own group we investigated the question of whether heparin can prevent post-ERCP pancreatitis [8]. This old drug has received increased attention in recent years because of its anti-inflammatory effects. The very early events of inflammation, rolling and adhesion of leukocytes are probably the target of heparin, which almost completely inhibits the activity of endothelial cells and leucocytes at the level of P-selectin and L-selectin interaction with their ligands. A total of 815 patients who underwent ERCP with sphincterotomy were studied for risk factor analysis. Among the patients, 268 received heparin for various clinical reasons. The incidence of post-ERCP pancreatitis was significantly lower in patients receiving heparin compared with others in the final multivariate model (3.4 %, 9/268 vs. 7.9 %, 43/547; P = 0.005). Within the heparin group we found evidence that there was a difference in the post-ERCP pancreatitis rates between those receiving unfractionated heparin and those having low-molecular-weight heparin (0.0 % vs. 4.3 %). Because of the study design, those data can be used only to support the hypothesis that heparin might be helpful in the prevention of post-ERCP pancreatitis. As long as the results of prospective, placebo-controlled, randomized studies are awaited - at least one trial is underway - heparin administration should be considered as a possible confounder in future studies of post-ERCP pancreatitis. Three other drugs are supposed to have beneficial effects in post-ERCP pancreatitis: native somatostatin, its synthetic analogue octreotide (both drugs are antisecretory), and the synthetic protease inhibitor gabexate mesylate. In a current meta-analysis, native somatostatin, which is a potent inhibitor of pancreatic enzyme secretion and which additionally induces a relaxation of the sphincter of Oddi, and gabexate mesylate were significantly associated with an improvement in post-ERCP pancreatitis, hyperamylasemia, and pancreatic-type pain [27]. The study was criticized, however, since only one study for gabexate mesylate was included, while eight or more randomized, controlled trials were used for the investigation of native somatostatin or octreotide [20].

Another issue in post-ERCP pancreatitis is the follow-up strategy after ERCP. The second aim of the Greek study was to analyse the prognostic value of follow-up examinations after ERCP [1]. The results of this part of the study are difficult to compare with other studies, since the amylase limit used for the definition of post-ERCP pancreatitis was four to five times the upper normal value and not, as recommended [21], a threefold elevation. Thus, several of the patients who were experiencing pain 24 hours after the procedure, with an amylase elevation below a four- to fivefold increase in amylase, would have been rated as having mild pancreatitis at other centres (19/513, 3.7 %), and, on account of the 1-day delay in discharge, in this study also they should have been labelled as having pancreatitis. This deviation from consensus might explain the low incidence of post-ERCP pancreatitis (3.3 %) in the study. Of 102 patients, 17 with any hyperamylasemia (16.8 %), or 17 of 64 (26.6 %) with any pain, at 4 hours after the procedure, had pancreatitis. The only significant prognostic parameter for the development of pancreatitis was a combination of pancreatic-type pain and elevated amylase values at 24 hours. This, however, is the common definition of post-ERCP pancreatitis, published 10 years ago, accepted as a consensus in most endoscopy centres, and used in most studies about ERCP and EST in the meantime [21]. Thus, there remain more questions than answers after one has read the paper. Should we wait 24 hours before we treat a patient who has pain and hyperamylasemia some hours after the procedure? Should we renounce checks performed on the day of ERCP? Should we determine amylase values only in symptomatic patients? The critical issues in cases of legal conflicts after complications (when informed consent has been given, and ERCP correctly executed) are the time when symptoms begin and the delay before specific diagnostic or therapeutic procedures. Thus, a limitation of blood samples on symptomatic patients at the time when symptoms begin or when they continue after ERCP might be acceptable in hospitalized patients. Symptom assessment together with routine amylase analysis should be carried out to estimate the risk of developing pancreatitis, at least when the discharge of a patient on the day of ERCP is planned. Several studies have shown that such an assessment is useful [22] [31] [32] [33] [34] [35], although there might be a problem with the timing. A time of 2 hours after ERCP might be too short because of the possibility of a delayed increase in amylase values in some patients and because of the mitigation of symptoms by sedatives and analgesics administered during ERCP. In a recent study from Australia [35], amylase values of less than 1.5 of the upper normal level had a negative predictive value of 100 % and this appears to be a reliable basis for discharge, if no other concerns are present. On the other hand, amylase elevation above 3.0 times of the upper normal level had a positive predictive value of 36.8 %, and was used as a cut-off value for definitive admission. In the grey area between 1.5 and 3.0 times of the upper normal value, it was recommended that clinical assessments of other symptoms or concerns should be made in order to decide about management. Severe pain or pain with a sustained need for analgesia after ERCP is a symptom of heterogeneous complications of ERCP including pancreatitis and perforation, and it should therefore be interpreted separately from and independent of hyperamylasemia. Those suggestions probably represent the current follow-up strategy at most endoscopy centers.

Post-ERCP pancreatitis is the result of a very complex and multifactorial reaction of the pancreatic gland to the endoscopic intervention. Most studies published in recent years describe and assess the problem by detailed differentiation into single factors involved, and thus add new information to the understanding of the causes of post-ERCP pancreatitis. A more standardized assessment of the preprocedural conditions, the techniques applied, the strategies used to avoid complications, and the final outcome according to success and complications is mandatory in studies which generate new hypotheses (see Table [1]). Additionally, studies which attempt to verify a specific hypothesis should be improved in design (multicenter, randomized, controlled) and number of cases. The organization and financing of such studies with confirmative power is the challenge of the future. Knowledge about post-ERCP pancreatitis has grown significantly, but its interpretation is still somewhat complex and sometimes application is difficult in individual situations. Perhaps we now need a more integrating approach which will summarize, in a simpler model, the multiple factors leading to post-ERCP pancreatitis or which mitigate its course.

Table 1 Overall assessment of variables related to post-endoscopic retrograde cholangiopancreatography (post-ERCP) pancreatitis Indication Symptoms, diagnostic or therapeutic strategy Risk factors Pre-ERCP (Suspicion of) sphincter of Oddi dysfunction (Suspicion of) pancreas divisum Sex Age Previous pancreatitis Prophylactic medical therapy ERCP-related Endoscopist (number of ERCP procedures per week) Papillary cannulation (easy, moderate, difficult, needle-knife) Deep pancreatic cannulation (no, yes, often) Pancreatic visualization (grade O, A, B, C) Pancreatic sphincterotomy Minor papilla procedure in pancreas divisum Prophylactic procedural therapy (pancreatic drainage, failed, not intended) Follow-up Amylase, lipase levels Pain Final technical result Success (yes, no, ongoing therapy) Pancreatitis, other (severity, outcome) Final clinical result Success (improvement of symptoms, failed, ongoing therapy)

References

• 1 Christoforidis E, Goulimaris I, Kanellos I. et al . Post-endoscopic retrograde cholangiopancreatography pancreatitis and hyperamylasaemia: patient-related and operative risk factors.  Endoscopy. 2002;  34 286-292
• 2 Fogel E L, Eversman D, Jamidar P. et al . Sphincter of Oddi dysfunction: pancreatobiliary sphincterotomy with pancreatic stent placement has a lower rate of pancreatitis than biliary sphincterotomy alone.  Endoscopy. 2002;  34 280-285
• 3 Budzynska A, Marek T, Nowak A. et al . A prospective, randomized, placebo-controlled trial of prednisone and allopurinol in the prevention of ERCP-induced pancreatitis.  Endoscopy. 2001;  33 766-772
• 4 Freeman M L, Nelson D B, Sherman S. et al . Complications of endoscopic biliary sphincterotomy.  N Engl J Med. 1996;  335 909-918
• 5 Freeman M L, DiSario J A, Nelson D B. et al . Risk factors for post-ERCP pancreatitis: a prospective, multicenter study.  Gastrointest Endosc. 2001;  54 425-434
• 6 Loperfido S, Angelini G, Beneditti G. et al . Major early complications from diagnostic and therapeutic ERCP: a prospective multicenter study.  Gastrointest Endosc. 1998;  48 1-10
• 7 Rabenstein T, Schneider H T, Bulling D. et al . Analysis of risk factors of endoscopic sphincterotomy techniques: preliminary results of a prospective study with emphasis on the reduced risk of acute pancreatitis under low-dose anticoagulation.  Endoscopy. 2000;  32 10-19
• 8 Rabenstein T, Roggenbuck S, Framke B. et al . Complications of endoscopic sphincterotomy: can heparin prevent acute pancreatitis following endoscopic retrograde cholangio-pancreatography?.  Gastrointest Endosc.. In press 2002; 
• 9 Masci E, Toti G, Mariani A. et al . Complications of diagnostic and therapeutic ERCP: a prospective multicenter study.  Am J Gastroenterol. 2001;  96 417-423
• 10 Sherman S, Lehmann G A. ERCP and endoscopic sphincterotomy induced pancreatitis.  Pancreas. 1991;  6 350-367
• 11 Rabenstein T, Schneider H T, Nicklas M. et al . The impact of skill and experience of the endoscopist on the outcome of endoscopic sphincterotomy.  Gastrointest Endosc. 1999;  50 628-636
• 12 Tarnasky P, Palesch Y, Cunningham J T. et al . Pancreatic stenting prevents pancreatitis after biliary sphincterotomy in patients with sphincter of Oddi dysfunction.  Gastroenterology. 1998;  115 1518-1524
• 13 Aizawa T, Ueno N. Stent placement in the pancreatic duct prevents pancreatitis after endoscopic sphincter dilation for removal of bile duct stones.  Gastrointest Endosc. 2001;  54 209-213
• 14 Akashi R, Kiyozumi T, Tanaka T. et al . Mechanism of pancreatitis caused by ERCP.  Gastrointest Endosc. 2002;  55 50-54
• 15 Sherman S, Ruffolo T A, Hawes R H, Lehman G A. Complications of endoscopic sphincterotomy: a prospective series with emphasis on the increased risk associated with sphincter of Oddi dysfunction and nondilated ducts.  Gastroenterology. 1991;  101 1068-1075
• 16 Bub H, Burner W, Riemann J F, Stolte M. Morphology of the pancreatic duct epithelium after traumatization of the papilla of Vater or endoscopic retrograde pancreatography with contrast media in cats.  Scand J Gastroenterol. 1983;  18 581-592
• 17 Kozarek R A. Endoscopic therapy of complete and partial pancreatic duct disruptions.  Gastrointest Endosc Clin N Am. 1998;  8 39-53
• 18 Kim H S, Lee D K, Kim I W. et al . The role of endoscopic retrograde pancreatography in the treatment of traumatic pancreatic duct injury.  Gastrointest Endosc. 2001;  54 49-55
• 19 Lau S T, Simchuk E J, Kozarek R A, Traverso L W. A pancreatic ductal leak should be sought to direct treatment in patients with acute pancreatitis.  Am J Surg. 2001;  181 411-415
• 20 Haber G B. Prevention of post-ERCP pancreatitis.  Gastrointest Endosc. 2000;  51 1-7
• 21 Cotten P B, Lehmann G A, Vennes J. et al . Endoscopic sphincterotomy complications and their management: an attempt at consensus.  Gastrointest Endosc. 1991;  37 338-393
• 22 Testoni P A, Bagnolo F, Natale C, Primignani M. Incidence of post-endoscopic retrograde-cholangiopancreatography/sphincterotomy pancreatitis depends upon definition criteria.  Dig Liver Dis. 2000;  32 419-421
• 23 Schölmerich J. Immunologic mechanisms in acute pancreatitis. In: Lankisch PG, DiMango EP (eds) Pancreatic disease. Berlin; Springer-Verlag 1999: 24-35
• 24 Testoni P A, Bagnolo F, Andriulli A. et al . Octreotide 24-h prophylaxis in patients at high risk for post-ERCP pancreatitis: results of a multicenter, randomized, controlled trial.  Aliment Pharmacol Ther. 2001;  15 965-972
• 25 Sudhindran S, Bromwich E, Edwards P R. Prospective randomized double-blind placebo-controlled trial of glyceryl trinitrate in endoscopic retrograde cholangiopancreatography-induced pancreatitis.  Br J Surg. 2001;  88 1178-1182
• 26 Deviere J, Le Moine O, van Laethem J L. et al . Interleukin 10 reduces the incidence of pancreatitis after therapeutic endoscopic retrograde cholangiopancreatography.  Gastroenterology. 2001;  120 498-505
• 27 Andriulli A, Gioacchino L, Grazia N. et al . Medical treatment can prevent pancreatic injury after ERCP: a meta-analysis.  Gastrointest Endosc. 2000;  51 1-7
• 28 Rosenberg J, Overgaard H, Andersen M. et al . Double blind randomised controlled trial of effect of metoprolol on myocardial ischaemia during endoscopic cholangiopancreatography.  BMJ. 1996;  313 258-261
• 29 de Maertelaer V, Deviere J, Le Moine O. IL-10 in post-ERCP pancreatitis (reply).  Gastroenterology. 2001;  121 233-234
• 30 Dumot J A, Conwell D L, Zuccaro G. et al . A randomized, double blind study of interleukin 10 for the prevention of ERCP-induced pancreatitis.  Am J Gastroenterol. 2001;  96 2098-2102
• 31 Gottlieb K, Sherman S, Pezzi J. et al . Early recognition of post-ERCP pancreatitis by clinical assessment and serum pancreatic enzymes.  Am J Gastroenterol. 1996;  91 1553-1557
• 32 Testoni P A, Bagnolo F, Caporuscio S, Lella F. Serum amylase measured four hours after endoscopic sphincterotomy is a reliable predictor of postprocedure pancreatitis.  Am J Gastroenterol. 1999;  94 1129-1130
• 33 Testoni P A, Caporuscio S, Bagnolo F, Lella F. Twenty-four-hour serum amylase predicting pancreatic reaction after endoscopic sphincterotomy.  Endoscopy. 1999;  31 131-136
• 34 Wojtun S, Gil M, Gil J. Recognition of ERC-induced pancreatitis in patients with choledocholithiasis by an analysis of laboratory findings.  Hepatogastroenterology. 2000;  47 550-553
• 35 Thomas P R, Sengupta S. Prediction of pancreatitis following endoscopic retrograde cholangiopancreatography by the 4-h post procedure amylase level.  J Gastroenterol Hepatol. 2001;  16 923-926