Precut Papillotomy Versus Persistence in Difficult Biliary Cannulation: A Prospective Randomized Trial

• Introduction
• Patients and Methods
• Patients
• Study Design
• Outcomes
• Statistical Analysis
• Results
• Discussion
• References

Background and Study Aims: Failed biliary cannulation occurs in up to 10 % of patients undergoing ERCP. There is some controversy as to the safety and efficacy of using precut techniques to achieve biliary cannulation in difficult cases. To date, no randomized trial has compared the success and complication rates of precut with the rates for persistence when biliary cannulation is difficult. The aim of this study was to compare the success rates and complication rates of precut with the success rates and complication rates of persistence in cases of difficult biliary cannulation.
Patients and Methods: Patients without prior sphincterotomy who required biliary cannulation were screened. A “difficult biliary cannulation” was arbitrarily defined as failed cannulation after 12 minutes. These patients were then randomized to continue treatment by needle-knife cut over the roof of the papilla or by persistence with a non-wire-guided, single-lumen papillotome. “Primary” success was defined as deep cannulation within 15 minutes of randomization. Primary and final success rates and complication rates within 30 days after ERCP were compared.
Results: Over a 38-month period a total of 642 patients were screened. Patients in whom biliary cannulation was successful within a time period of 12 minutes or less formed the reference group (n = 580). The remainder of the patients were randomly assigned to the “precut” arm (n = 32) or to the “persistence” arm (n = 30). Primary success rates and complication rates were similar in the precut and persistence arms (75 % and 4 % respectively for the precut arm vs. 73 % and 9 % for the persistence arm). The final successful cannulation rate in the entire group of 642 patients was 99.5 %.
Conclusions: In experienced hands, precut papillotomy and persistence in cannulation are equally effective in cases of difficult cannulation, with a similar complication rate.

Introduction

Endoscopic retrograde cholangiopancreatography (ERCP) has become an invaluable diagnostic and therapeutic tool for biliary and pancreatic disorders. The short-term (< 30 days) post-ERCP complication rate, with or without sphincterotomy, ranges from 5 % to 11 % [1] [2] [3] [4] [5] [6]. Major complications include pancreatitis (5 % to 7 %), post-sphincterotomy hemorrhage (0.2 % to 5 %), cholangitis (< 1 %), and perforation (< 1 %) [6] [7]. Despite the availability of specialized catheters, failed biliary cannulation occurs in up to 10 % of ERCP procedures [8] [9]. The cannulation success rate depends on patient selection, the utilization of specialized catheters, and the skill and experience of the endoscopist. In cases of difficult cannulation the endoscopist is faced with a choice between making persistent attempts at conventional biliary cannulation; the use of a needle-knife precut procedure; or other options, such as aborting the procedure, trying again later or referral to another center. The use of precut papillotomy to facilitate biliary cannulation remains controversial and the indications and timing for precut techniques vary [10].

Large prospective studies have demonstrated that difficult cannulation alone is an independent risk factor for post-ERCP complications [1] [3] [5]. Freeman et al. [3] prospectively studied 1963 consecutive patients at 11 centers in the United States. The risk of pancreatitis after a difficult cannulation (n = 710), compared with standard cannulation, increased from 4.3 % to 11.3 %, with an adjusted odds ratio of 3.41. In another study, a single-center study involving 1223 patients, the risk of pancreatitis after a difficult cannulation was 14.9 %, compared with a rate of 3.3 % for standard cannulation, with an adjusted odds ratio of 2.81 [5]. Possible reasons for the increased risk of pancreatitis may be excessive manipulation, resulting in mechanical trauma and edema of the pancreatic sphincter, or repeated contrast injection into the pancreatic duct. At what point the standard cannulation technique should be abandoned is unclear but continued attempts, indicative of problematic cannulation, are thought to increase the risk of pancreatitis. This has been postulated as a potential confounding factor in the reported high complication rates for precut papillotomy.

Dhir and Mallath [11] pointed out that more data are needed on the appropriate timing for precut papillotomy. In addition, studies are needed which distinguish between procedures in which a trainee initiates the attempts at biliary cannulation and procedures in which this is done by a senior endoscopist, as involvement of a trainee may prolong the procedure and/or result in increased papillary trauma [11]. Although precut papillotomy may improve the cannulation success rate, many prospective studies have suggested that this procedure is an independent predictor of post-ERCP complications [1] [2] [4]. The complication rate of precut sphincterotomy varies from 6 % to 30 % in different studies [1] [3] [5] [12] [13] [14], although one of these studies showed no significant increase in risk [3].

To date there has been no prospective randomized study comparing the complication and success rates of precut papillotomy with those of persistent attempts with standard techniques in biliary cannulation. We conducted a prospective randomized study comparing the cannulation and complication rates of precut papillotomy with the rates for persistent attempts at cannulation, when standard attempts by endoscopy fellows, followed by senior endoscopists failed. We hypothesized that precut papillotomy would result in a higher success rate but also a higher complication rate than persistence.

Patients and Methods

Patients

This prospective, single-center study was conducted over a 38-month period, between September 1995 and November 1998, in the Endoscopy Unit at The Wellesley Hospital, University of Toronto. From July to September each year the study was temporarily suspended in order to allow the incoming endoscopy fellows to develop their ERCP skills. The hospital’s Ethics Committee approved this study and all patients signed an informed consent form for the procedure and for enrolment into the study. All patients who had a high likelihood of having bileduct pathology which would require deep biliary cannulation, and who had no history of prior papillotomy were screened for entry into this study. Biliary cannulation was considered necessary if the patient had an abnormal ultrasound scan (dilated bile ducts or choledocholithiasis), or abnormal liver function tests. The exclusion criteria for the study were: a) age under 18 years; b) coagulopathy; c) Karnofsky performance status < 40 %; d) Billroth II gastrectomy; e) periampullary diverticulum. It was considered that these last two anatomical features could potentially make the cannulation approach either more difficult or unsafe to perform.

Study Design

The baseline patient history was recorded at the time of screening (Table [1]). All patients were prepared and sedated for ERCP as per standard medical practice. ERCP was performed in the standard way with a diagnostic Olympus duodenoscope (TJF 140F or TJF 160F; Olympus Optical Co., Tokyo, Japan) and a non-wire-guided, single-lumen papillotome (Olympus KD-21Q-1). After the procedure patients were monitored in a recovery room by nurses who were unaware of the patients’ study status. Prophylactic antibiotics were not given to the patients enrolled in the study. The patients were asked to contact their physician or the endoscopy fellow on-call at any time if they developed symptoms such as pain or fever. All randomized patients were contacted by an investigator (M.C.) by phone between day 3 and day 7 post-procedure, and again at day 30 post-procedure, in order to assess them for the development of potential side effects or complications, including pain, bleeding, cholangitis, or pancreatitis. Any additional care required, including visits by a physician or hospitalization, was recorded. The patients in the reference group were contacted and followed-up in the same way, although we only obtained follow-up data for 82.9 % of these patients.

In this study, a ”difficult biliary cannulation“ was arbitrarily defined as a failure of cannulation after a 7-minute attempt by an endoscopy fellow, followed by an additional 5-minute attempt by one of two experienced staff endoscopists (G.H., P.K.). In each instance, the start time began when the papillotome was advanced out of the instrument channel of the endoscope, in front of the papilla. Successful deep biliary cannulation was evidenced by injection of contrast dye and advancement of the papillotome into the biliary tree. If unsuccessful within the defined time periods, these patients were then randomized to the ”precut“ arm or to the ”persistence“ arm (Figure [1]). Randomization was achieved with the use of a sealed envelope with computer generated randomization codes with a 50 % probability of receiving either treatment.

Precutting was performed by a staff endoscopist using a needle-knife over the roof of the papilla, with a standard needle-knife (HPC-2; Wilson-Cook, Winston-Salem, North Carolina, USA) exposing the opening of the distal bile duct. Once biliary access had been obtained, a complete sphincterotomy with the Olympus traction papillotome was performed if indicated. We used blended cutting current with an ERBE generator (ICC 350; ERBE, Tubingen, Germany). In the persistence arm, the staff endoscopist continued to attempt cannulation with the non-wire-guided, single-lumen papillotome. In this study, as was the standard practice of these endoscopists, wire-guided cannulation was not attempted. The staff endoscopist continued to attempt cannulation for 15 minutes post-randomization (the primary study period) in patients in both arms. Institutional ethics required that the patients who had not been successfully cannulated by the end of the primary study period should be treated according to the discretion of the staff endoscopist (in a discretionary follow-on period), with a choice of continuing with the randomized procedure, crossing over to the other procedure, or aborting the procedure. Once successful biliary cannulation was achieved, patients received the therapeutic interventions they required, as indicated by the findings. In patients in the precut arm, a complete sphincterotomy was performed with the traction papillotome.

Outcomes

The primary outcomes of the study were the primary cannulation success rates and the primary complication rates for the precut and persistence arms. ”Primary cannulation success” was defined as the achievement of deep biliary cannulation within 15 minutes after randomization. If biliary cannulation still could not be achieved after 15 minutes of attempts by a staff endoscopist, the procedure was considered a primary study failure. The primary complication rates were calculated for those patients with successful primary cannulation. Post-ERCP complications were defined as any adverse clinical events or unexpected clinical outcomes related to the procedure. Specifically, this study assessed the presence of pancreatitis, bleeding, pain, cholangitis, perforation, and any unexpected adverse clinical event [15].

Clinical bleeding was defined as the presence of clinical evidence of hemorrhage with a drop of over 10 % in hemoglobin concentration, a need for endoscopic intervention, or a need for hospitalization within 7 days post-procedure. Although cases of endoscopically observed bleeding (defined as acute bleeding during ERCP that required control with local epinephrine injection but without clinical bleeding (melena or hematochezia) or a drop in hemoglobin of over 10 %) were recorded, they were not counted as formal post-ERCP complications.

Pancreatitis was defined as new-onset abdominal pain and an elevation in serum amylase levels to twice the upper limit of normal. The severity of the pancreatitis was graded as mild (requiring a hospital admission lasting 3 days or less), moderate (hospital admission of 4 - 10 days), or severe (hospital admission lasting more than 10 days or complicated by pseudocyst, abscess, or the need for surgical intervention) [15].

Abdominal pain without pancreatitis was defined as post-procedure abdominal pain that continued for over 1 hour or worsening of pain post-ERCP for over 1 hour, without evidence of pancreatitis after 24 hours. Oversedation was defined as a requirement for sedation reversal agents to be given. Suspected sphincter of Oddi dysfunction (SOD) was defined as a clinical diagnosis in which a functional or structural abnormality of the sphincter of Oddi was suspected of being the cause of recurrent abdominal pain. Manometry was not performed.

The final cannulation status after the full ERCP session and the complications in all the patients, regardless of which procedure was performed, were the secondary results. The time to deep biliary cannulation was recorded for all patients.

Statistical Analysis

In terms of the baseline characteristics of the reference and randomized groups, proportions were compared using Fisher's exact test, and age was compared using Student’ t-test. An exploratory multivariate logistic regression was fitted with difficult cannulation as the outcome, and age (categorized as < 62, the median, vs. ≥ 62), history of a prior failed ERCP attempt, outpatient procedure, and underlying pancreatic or biliary cancer as predictors.

Only two patients with missing predictors could not be used in the model. The primary complication rate was calculated for those patients with primary successful cannulation. The primary success and complication rates were compared using Fisher’s exact test. Post-randomization cannulation time was compared using Student’s t-test. The 95 % confidence intervals (CI) were determined for rates by using the binomial distribution (‘exact’ method, evaluated via the F distribution [16]); the 95 % confidence intervals for the difference between two rates were determined using the Wilson score method with continuity correction [17].

Secondary (also called ”final“) results for all patients were calculated using an intent-to-treat approach, in which randomized patients who were not successfully cannulated during the primary period were included in the arm to which they were randomized, regardless of which procedure was eventually used. Where reported, the same statistical methods were used as for the primary rates.

All P values reported are two-sided. No explicit correction was made for multiple testing. The two hypotheses regarding the primary outcomes were pre-specified. All other tests are exploratory and can at best serve to generate a new hypothesis. One interim analysis was performed for presentation purposes.

Post hoc sample size calculation: differences in primary success rate. Sample size calculations showed that 269 patients were needed in each group to achieve 80 % power to detect a difference of 10 percentage points between the null hypothesis, that the success rate is 75 %, and the alternative hypothesis, that the success rate for the precut procedure is 85 %, using a two-sided chi-squared test with continuity correction and with a significance level of 0.05.

Post hoc sample size calculation: differences in complication rates. Sample size calculations showed that 352 patients were needed in each group to achieve 80 % power to detect a difference of 7 percentage points between the null hypothesis, that the complication rate is 8 %, and the alternative hypothesis, that the complication rate for the precut procedure is 15 %, using a two-sided chi-squared test with continuity correction and with a significance level of 0.05. For a difference from 10 % to 15 %, 725 patients per group would be needed.

Results

A total of 642 patients were screened for this study (Figure [2]). The reference group (patients who were cannulated in a time of 12 minutes or less) consisted of 580 patients, in whom biliary cannulation was achieved in a mean time ± SD of 4.0 ± 3.5 minutes. The remaining 62 patients were defined as having a difficult biliary cannulation and were randomly assigned to the precut (n = 32) or persistence (n = 30) arms.

The baseline demographic and clinical features of the reference and difficult cannulation groups are shown in Table [1]. Four baseline factors were significantly associated with difficult cannulation: older age, history of prior failed ERCP attempt, outpatient procedure, and underlying pancreatic or biliary cancer. In a multivariate logistic regression model containing these variables, the significance of having an outpatient procedure was reduced (to P = 0.08), while the other three factors remained highly significant (all P < 0.01) and made independent contributions toward predicting difficult cannulation.

The baseline characteristics of the precut and persistence arms were all similar (Table [1]) although the precut arm had a higher proportion of patients who had a history of prior failed ERCP (44 % vs. 20 %).

The primary results are summarized in Table [2]. The primary cannulation success rate after randomization was similar in precut and persistence arms, being 75 % (CI 56.6 % to 88.5 %) and 73.3 % (CI 54.1 % to 87.7 %) respectively. The difference of 1.7 percentage points (CI -21.6 to 25.0) was not significant (P = 1.00). The complication rates in patients successfully cannulated within the primary period in the precut and persistence arms were 4.2 % (CI 0.1 % to 21.1 %) and 9.1 % (CI 1.1 % to 29.2 %) respectively. The difference of 4.9 percentage points (CI 15.5 to 26.8) was not significant (P = 0.60).

During the discretionary period, five of the eight patients in the precut arm in whom primary cannulation failed were eventually cannulated successfully with continued precut. In two patients cannulation was achieved using a sphincterotome, one within 15 minutes and one within 40 minutes. In one patient in the precut arm we failed to cannulate successfully during the ERCP session. This patient underwent successful laparoscopic cholecystectomy without a second ERCP. Of the eight patients, three developed post-ERCP complications, pancreatitis in two and clinical bleeding in one. One of the two patients who developed pancreatitis was crossed over to persistence with a sphincterotome.

Three of the eight patients in the persistence arm in whom primary cannulation failed were eventually successfully cannulated with continued persistence, one of these patients requiring more than 30 minutes after randomization for cannulation. Precut papillotomy was performed in three patients, who were successfully cannulated within 15 minutes. In two patients in the persistence arm we failed to cannulate successfully in the ERCP session: one went on to have a successful second ERCP 7 days later and the other needed no further ERCP due to the resolution of symptoms and indications. Of the eight patients in the persistence arm who failed primary cannulation, only one developed a post-ERCP complication (pancreatitis). Neither of the two patients who had either a second ERCP or surgery developed any complications associated with the procedure or surgery.

The secondary results are summarized in Table [3]. The results for randomized patients who were not successfully cannulated during the primary period were included in an ‘intent-to-treat’ fashion in the arm to which they were randomized, regardless of which procedure was eventually used. The final cannulation success rate during the same ERCP session was 97 % (CI 83.8 % to 99.9 %) in the precut arm and 93 % (CI 77.9 % to 99.2 %) in the persistence arm. Overall, among the 642 patients screened in this study, biliary cannulation was successful in the same session in 99.5 % of patients. The overall post-ERCP complication rate was higher in the randomized (difficult biliary cannulation) group (11.3 %, CI 4.7 % to 21.9 %) than in the reference group (4.6 %, CI 2.9 % to 6.8 %), a difference of 6.7 percentage points (CI 0.0 to 18.0). The post-ERCP pancreatitis rate was similarly higher: 6.5 % (CI 4.7 %-21.9 %) versus 0.8 % (CI 2.9 %-6.8 %), a difference of 5.7 percentage points (CI 1.0 - 15.7).

There were no reported cases of oversedation or of perforation in the entire study population.

Discussion

The long unanswered question with respect to the precut technique is whether or not the observed complication rate is a function of the technique itself or merely a reflection of the fact that the cannulation has been difficult and the prolonged attempts to achieve cannulation may result in trauma to the papilla with resultant edema accounting for the increased rate of pancreatitis [11] [13]. In this study, precutting and persistent attempts at cannulation were performed by only two experienced staff endoscopists and this minimized many confounding factors, such as the difference in experience between endoscopists and endoscopy fellows. The setting of a predefined time limit prior to randomization served to overcome the issue of papillary trauma as a factor in the complications associated with precut because patients were randomized early in the course of the ERCP. Some potential biases do exist in the study: the outcomes assessor in our study was unblinded; the secondary success and complication rates were ”contaminated” by the discretionary procedures executed by the staff endoscopist in patients whose procedures were primary study failures; and although all the randomized patients were reached for follow-up, we were only able to obtain follow-up data on 83 % of patients in the reference group.

Post hoc sample size calculations showed that very high numbers of patients would be needed to have a high probability of detecting differences of 5 to 10 percentage points in the primary success and complication rates. For a change from 10 % to 15 % in complication rates, 725 patients per group would be needed to achieve 80 % power. Using the same screening to enrollment ratio as obtained in the study, more than 15 500 patients would have had to be screened. This number is prohibitive for a prospective study, even with many participating centers. The largest prospective multicenter series to date have enrolled a maximum of 2769 patients [1] [2]. The accrual of patients for this study was slower than anticipated due to the higher than expected success rate of cannulation within 12 minutes, and the study was stopped because of the difficulty in obtaining sufficient numbers of patients. The low power is also evidenced in the wide confidence intervals presented for the primary outcomes. Although the study has low power, we feel it is important to present these results obtained in a prospective manner in a well-defined cohort of patients with difficut cannulation.

The primary cannulation success rate after randomization was very similar in the precut and persistence arms (75 % vs. 73.3 %). The complication rates in patients successfully cannulated within the primary period were very low (4.2 % in the precut arm and 9.1 % in the persistence arm), there being only one patient in the precut arm and two in the persistence arm who experienced complications. The study does not provide evidence that precut has either a higher success rate or a higher complication rate than persistence. The study suggests that the success and complication rates for these two treatment options are similar.

In this high-volume center, with experienced endoscopists, persisting with a standard cannulation technique or employing precut were equally effective and safe in cases of difficult cannulation. In some studies, precut has been used in as many as 25 % to 40 % of cases to achieve biliary access [2] [18]. In prior studies, precut papillotomy appeared to be no riskier than standard papillotomy [12] [18] [19], but it should be emphasized that these reports were non-randomized and retrospective and emanated from high-volume centers with expert endoscopists. In contrast, in prospective multicenter trials involving academic and community-based practices with endoscopists with varying levels of experience, the precut technique has been shown to be an independent risk factor for overall complications, with an adjusted odds ratio of 3.61 in one American study [1] and a relative risk of 1.87 in one Italian study [2]. In the American study, the adjusted odds ratio for pancreatitis associated with precut was 4.34. In another Italian multicenter study, the adjusted odds ratio for pancreatitis associated with precut was 2.80 [4]. Others have warned of the dangers of this procedure [6] [20].

Because our study was conducted in a high-volume center and by experienced endoscopists, the cannulation success rate was high and the complication rates were low, results which are comparable with reports from other high-volume centers. The final post-ERCP pancreatitis rate was very low: 0.8 % in the nonrandomized, reference group and 6.3 % and 6.7 % in the precut and persistence arms respectively. The relatively low rate of pancreatitis in the patients with difficult cannulation in this study is comparable to overall rates in unselected patient populations undergoing ERCP [1] [2] [3] [5]. It should be noted that during the period when the study was undertaken, prophylactic measures, such as pancreatic stenting, was not employed. In addition, known factors that may increase the complication rate, such as a history of previous Billroth II gastrectomy and periampullary diverticula, were excluded.

An interesting progression is seen in the complication rates when we pool the data from the two randomization arms: the post-ERCP complication rate was 4.6 % in the reference group, 6.5 % among the primary successes and 25.0 % in the primary failures (Figure [2], Table [3]): these rates correspond to the patients cannulated within 12 minutes of cannulation attempts, those cannulated in more than 12 minutes but less than 27 minutes, and those whose cannulation took more than 27 minutes, respectively. The pancreatitis rates followed a similar pattern: 0.8 %, 2.2 % and 18.8 % respectively for the three groups. This supports previous observations of associations between difficult cannulation and complication rates. Increased attempts at cannulation may result in mechanical trauma. The relatively high complication and pancreatitis rates in the small number of patients (n = 16) who were not successfully cannulated after 27 minutes in our study remains a point of interest.

With experienced endoscopists in a high-volume center, precut papillotomy and persistence in cannulation result in similar cannulation success and complication rates in cases of difficult biliary cannulation. It is reasonable, therefore, to assume that the endoscopist’s experience and preferences should dictate which course of action to pursue. The availability of wire-guided sphincterotomes and tapered-tip catheters provides another option in cases of difficult cannulation, although this capability has contributed to cannulation success only in patients with distorted anatomy in our experience. Other techniques that may be used to increase the success of cannulation include prior placement of a wire or stent in the pancreatic duct. It is unlikely that the use of these newer adjunctive interventions would improve the success of cannulation beyond the 99.5 % achieved in this study. On the other hand, stent protection of the pancreas has the potential to reduce pancreatitis incidence and severity, an issue not addressed by this study. Nor was the use of pancreatic stents a confounding factor in the comparison of the techniques studied.

In an American multicenter study, the rate of complications was highest when the indication for ERCP was suspected SOD (21.7 %) [1]. In our study, there were 17 patients with suspected SOD, 14 in the reference group, one in the precut arm and two in the persistence arm. Fortunately, none of them developed pancreatitis and none of them underwent prophylactic pancreatic stenting, which was not standard practice at the time of the study. The endoscopists’ experience and the small number of patients with suspected SOD both contributed to the low pancreatitis rate. In the American multicenter study, precut sphincterotomy resulted in a complication rate of 35 % in patients with suspected SOD without pancreatic stenting [1]. In another retrospective study, the post-ERCP pancreatitis rate among all 436 patients with SOD was 19.7 % [21]. Combinations of patient characteristics, including female gender, normal serum bilirubin, recurrent abdominal pain, and previous post-ERCP pancreatitis, placed patients at increasingly higher risk of developing pancreatitis [3]. At least two randomized studies have shown that in patients with pancreatic sphincter hypertension undergoing biliary sphincterotomy, pancreatic duct stenting protects significantly against post-ERCP pancreatitis, reducing this complication from 26-28 % to 5-7 % [22] [23]. This practice may therefore be warranted in high-risk patients with difficult cannulation who undergo precut papillotomy.

Clinically significant hemorrhage with melena or hematochezia which may require intervention is reported to occur in 1-2 % of cases [1] [2] [6]. Endoscopically observed bleeding is more common, occurring in up to 10-30 %, but does not represent an adverse outcome for the patient, according to many investigators [6]. In our study only one randomized patient (in the precut arm) developed delayed hemorrhage (hematemesis) 4 days after ERCP, when anticoagulation was started with heparin. Out of the 481 patients in the reference group for whom we had follow-up data, clinically significant bleeding occurred in eight patients (1.7 %) and endoscopically observed hemorrhage in another nine patients, a total of 17 patients (3.5 %). In our study, patients with coagulopathy and patients receiving anticoagulation therapy were excluded from the trial. Both these factors have been found to be a risk for delayed post-ERCP hemorrhage [1].

In conclusion, in the difficult cannulation cohort, successful cannulation rates using the precut technique were equivalent to those of persistence with the standard method. Precut papillotomy may be carried out by experienced endoscopists after standard maneuvers fail to provide biliary access and when cannulation is likely to be followed by a therapeutic intervention. The endoscopist’s experience, the anatomical configuration of the papilla, and the physician’s preference should dictate when to attempt a precut technique. The results of this study cannot be extrapolated to endoscopic practice in general but do show that in expert centers there is no difference in outcome between employing a standard technique or a precut technique in cases of difficult cannulation.

References

• 1 Freeman M L, Nelson D B, Sherman S. et al . Complications of endoscopic biliary sphincterotomy.  N Engl J Med. 1996;  335 909-918
  CrossrefPubMedSearch in Google Scholar
• 2 Loperfido S, Angelini G, Benedetti G. et al . Major early complications from diagnostic and therapeutic ERCP: a prospective multicenter study.  Gastrointest Endosc. 1998;  48 1-10
  CrossrefPubMedSearch in Google Scholar
• 3 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
  CrossrefPubMedSearch in Google Scholar
• 4 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
  CrossrefPubMedSearch in Google Scholar
• 5 Vandervoort J, Soetikno R M, Tham T C. et al . Risk factors for complications after performance of ERCP.  Gastrointest Endosc. 2002;  56 652-656
  CrossrefPubMedSearch in Google Scholar
• 6 Freeman M L. Adverse events and success of ERCP.  Gastrointest Endosc. 2002;  56 S273-S282
  CrossrefPubMedSearch in Google Scholar
• 7 Cohen S, Bacon B R, Berlin J A. et al . ERCP for diagnosis and therapy. National Institutes of Health State-of-the-Science Conference Statement; 2002 January 14-16.  Gastrointest Endosc. 2002;  56 803-809
  CrossrefPubMedSearch in Google Scholar
• 8 Cortas G A, Mehta S N, Abraham N S, Barkun A N. Selective cannulation of the common bile duct: a prospective randomized trial comparing standard catheters with sphincterotomes.  Gastrointest Endosc. 1999;  50 775-779
  CrossrefPubMedSearch in Google Scholar
• 9 Schwacha H, Allgaier H P, Deibert P. et al . A sphincterotome-based technique for selective transpapillary common bile duct cannulation.  Gastrointest Endosc. 2000;  52 387-391
  CrossrefPubMedSearch in Google Scholar
• 10 Freeman M L. Towards improving outcomes of ERCP.  Gastrointest Endosc. 1998;  48 96-102
  CrossrefPubMedSearch in Google Scholar
• 11 Dhir V, Mallath M K. Is precut papillotomy guilty as accused? [Letter].  Gastrointest Endosc. 1999;  50 143-144
  PubMedSearch in Google Scholar
• 12 Neoptolemos J P, Shaw D E, Carr-Locke D L. A multivariate analysis of preoperative risk factors in patients with common bile duct stones: implications for treatment.  Ann Surg. 1989;  209 157-161
  PubMedSearch in Google Scholar
• 13 Slot W B, Schoeman M N, Disario J A. et al . Needle-knife sphincterotomy as a precut procedure: a retrospective evaluation of efficacy and complications.  Endoscopy. 1996;  28 334-339
  PubMedSearch in Google Scholar
• 14 Rollhauser C, Johnson M, Al-Kawas F H. Needle-knife papillotomy: a helpful and safe adjunct to endoscopic retrograde cholangiopancreatography in a selected population.  Endoscopy. 1998;  30 691-696
  PubMedSearch in Google Scholar
• 15 Cotton P B, Lehman G, Vennes J. et al . Endoscopic sphincterotomy complications and their management: an attempt at consensus.  Gastrointest Endosc. 1991;  37 383-393
  CrossrefPubMedSearch in Google Scholar
• 16 Leemis L M, Trivedi K S. A comparison of approximate interval estimators for the Bernoulli parameter.  Am Stat. 1996;  50 63-68
  PubMedSearch in Google Scholar
• 17 Newcombe R G. Interval estimation for the difference between independent proportions: comparison of 11 methods.  Stat Med. 1998;  17 873-890
  CrossrefPubMedSearch in Google Scholar
• 18 Binmoeller K F, Seifert H, Gerke H. et al . Papillary roof incision using the Erlangen-type precut papillotome to achieve bile duct cannulation.  Gastrointest Endosc. 1996;  44 689-695
  PubMedSearch in Google Scholar
• 19 Huibregtse K, Katon R M, Tytgat G NJ. Precut papillotomy via fine needle-knife papillotomy: a safe and effective technique.  Gastrointest Endosc. 1986;  32 403-405
  PubMedSearch in Google Scholar
• 20 Cotton P B. Precut papillotomy: a risky technique for experts only.  Gastrointest Endosc. 1989;  35 578-579
  PubMedSearch in Google Scholar
• 21 Fogel E L, Eversman D, Jamidar P. et al . Sphincter of Oddi dysfunction: pancreaticobiliary sphincterotomy with pancreatic stent placement has a lower rate of pancreatitis than biliary sphincterotomy alone.  Endoscopy. 2002;  34 280-285
  Thieme ConnectPubMedSearch in Google Scholar
• 22 Tarnasky P R, Palesch Y 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
  PubMedSearch in Google Scholar
• 23 Fazel A, Quadri A, Catalano M F. et al . Does a pancreatic duct stent prevent post-ERCP pancreatitis? A prospective randomized study.  Gastrointest Endosc. 2003;  57 291-294
  CrossrefPubMedSearch in Google Scholar

G. B. Haber, M.D.

6th Floor, Blackhall, Lenox Hill Hospital

100 East 77th Street · New York · New York 10021 · USA

Fax: +1-212-434-2446

Email: ghaber@lenoxhill.net

References

• 1 Freeman M L, Nelson D B, Sherman S. et al . Complications of endoscopic biliary sphincterotomy.  N Engl J Med. 1996;  335 909-918
  CrossrefPubMedSearch in Google Scholar
• 2 Loperfido S, Angelini G, Benedetti G. et al . Major early complications from diagnostic and therapeutic ERCP: a prospective multicenter study.  Gastrointest Endosc. 1998;  48 1-10
  CrossrefPubMedSearch in Google Scholar
• 3 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
  CrossrefPubMedSearch in Google Scholar
• 4 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
  CrossrefPubMedSearch in Google Scholar
• 5 Vandervoort J, Soetikno R M, Tham T C. et al . Risk factors for complications after performance of ERCP.  Gastrointest Endosc. 2002;  56 652-656
  CrossrefPubMedSearch in Google Scholar
• 6 Freeman M L. Adverse events and success of ERCP.  Gastrointest Endosc. 2002;  56 S273-S282
  CrossrefPubMedSearch in Google Scholar
• 7 Cohen S, Bacon B R, Berlin J A. et al . ERCP for diagnosis and therapy. National Institutes of Health State-of-the-Science Conference Statement; 2002 January 14-16.  Gastrointest Endosc. 2002;  56 803-809
  CrossrefPubMedSearch in Google Scholar
• 8 Cortas G A, Mehta S N, Abraham N S, Barkun A N. Selective cannulation of the common bile duct: a prospective randomized trial comparing standard catheters with sphincterotomes.  Gastrointest Endosc. 1999;  50 775-779
  CrossrefPubMedSearch in Google Scholar
• 9 Schwacha H, Allgaier H P, Deibert P. et al . A sphincterotome-based technique for selective transpapillary common bile duct cannulation.  Gastrointest Endosc. 2000;  52 387-391
  CrossrefPubMedSearch in Google Scholar
• 10 Freeman M L. Towards improving outcomes of ERCP.  Gastrointest Endosc. 1998;  48 96-102
  CrossrefPubMedSearch in Google Scholar
• 11 Dhir V, Mallath M K. Is precut papillotomy guilty as accused? [Letter].  Gastrointest Endosc. 1999;  50 143-144
  PubMedSearch in Google Scholar
• 12 Neoptolemos J P, Shaw D E, Carr-Locke D L. A multivariate analysis of preoperative risk factors in patients with common bile duct stones: implications for treatment.  Ann Surg. 1989;  209 157-161
  PubMedSearch in Google Scholar
• 13 Slot W B, Schoeman M N, Disario J A. et al . Needle-knife sphincterotomy as a precut procedure: a retrospective evaluation of efficacy and complications.  Endoscopy. 1996;  28 334-339
  PubMedSearch in Google Scholar
• 14 Rollhauser C, Johnson M, Al-Kawas F H. Needle-knife papillotomy: a helpful and safe adjunct to endoscopic retrograde cholangiopancreatography in a selected population.  Endoscopy. 1998;  30 691-696
  PubMedSearch in Google Scholar
• 15 Cotton P B, Lehman G, Vennes J. et al . Endoscopic sphincterotomy complications and their management: an attempt at consensus.  Gastrointest Endosc. 1991;  37 383-393
  CrossrefPubMedSearch in Google Scholar
• 16 Leemis L M, Trivedi K S. A comparison of approximate interval estimators for the Bernoulli parameter.  Am Stat. 1996;  50 63-68
  PubMedSearch in Google Scholar
• 17 Newcombe R G. Interval estimation for the difference between independent proportions: comparison of 11 methods.  Stat Med. 1998;  17 873-890
  CrossrefPubMedSearch in Google Scholar
• 18 Binmoeller K F, Seifert H, Gerke H. et al . Papillary roof incision using the Erlangen-type precut papillotome to achieve bile duct cannulation.  Gastrointest Endosc. 1996;  44 689-695
  PubMedSearch in Google Scholar
• 19 Huibregtse K, Katon R M, Tytgat G NJ. Precut papillotomy via fine needle-knife papillotomy: a safe and effective technique.  Gastrointest Endosc. 1986;  32 403-405
  PubMedSearch in Google Scholar
• 20 Cotton P B. Precut papillotomy: a risky technique for experts only.  Gastrointest Endosc. 1989;  35 578-579
  PubMedSearch in Google Scholar
• 21 Fogel E L, Eversman D, Jamidar P. et al . Sphincter of Oddi dysfunction: pancreaticobiliary sphincterotomy with pancreatic stent placement has a lower rate of pancreatitis than biliary sphincterotomy alone.  Endoscopy. 2002;  34 280-285
  Thieme ConnectPubMedSearch in Google Scholar
• 22 Tarnasky P R, Palesch Y 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
  PubMedSearch in Google Scholar
• 23 Fazel A, Quadri A, Catalano M F. et al . Does a pancreatic duct stent prevent post-ERCP pancreatitis? A prospective randomized study.  Gastrointest Endosc. 2003;  57 291-294
  CrossrefPubMedSearch in Google Scholar

G. B. Haber, M.D.

6th Floor, Blackhall, Lenox Hill Hospital

100 East 77th Street · New York · New York 10021 · USA

Fax: +1-212-434-2446

Email: ghaber@lenoxhill.net