Randomized Trial of Wire and Sphincterotome Systems for Biliary Cannulation

Overview

1. Cannulation of (placement of a small catheter into) the bile duct is critical to remove bile duct stones, divert bile leaks, and decompress biliary obstruction due to cancer. 2. Given the small size of the bile duct orifice and its close proximity to the pancreatic duct, selective biliary cannulation is the most difficult part of the endoscopic retrograde cholangiography (ERCP). 3. New small diameter sphincteromes and "short wire" systems (which allow physicians to control guidewires) offer potential, though untested advantages. 4. At most hosptial both the long and short wire systems as well as small versus standard are routinely used for clinical care. 5. Our hypothesis is that small diameter, physician controlled wires favor biliary cannulation 6. Our objective will be to assess whether small diameter sphincterotomes and "short wire" physician controlled guidewire cannulation favors successful bile duct cannulation and minimize complications.

1.0 BACKGROUND Bile duct access is the most difficult part of endoscopic retrograde cholangiography (ERCP) performed for the biliary indications of bile duct stones, leak, and stricture. In the past decade there has been an evolution in the technology and approach used to achieve selective biliary cannulation, In the classic approach a standard cannula is placed into the biliary orifice and contrast injected to confirm placement. Subsequently, a long wire is passed by the assistant deeply into the duct and the cannulatome exchange for a sphincterotome which is used to perform papillotomy. The first major change in cannulation approach has been the widespread use of the sphincterotome to obtain initial biliary access in lieu of the cannulatome. This is in part driven by economics. With the advent of endoscopic ultrasound (EUS) and magnetic resonance cholangiopancreatography (MRCP) almost all ERCP is therapeutic and sphincterotomy is typically part of this approach.1 Thus it is sensible to forgo the step of using the cannulatome which enables only diagnostic ERCP. In addition to cutting, the wire on the distal tip of the sphincterotome enables variable flexion in the direction of the bile duct, a feature lacking for the cannulatome. Additionally, high quality comparison trials suggest that the sphincterotome is superior to the cannulatome for initial access. Schwacha et al prospectively randomized patients to bile duct cannulation with the standard cannulatome versus the sphincterotome.2 After five attempts bile duct access was achieved in 62% of those in the standard catheter group compared to 84% in the sphincterotome group. Cortus et al randomized patients to \<15 attempts to access with bile duct with the cannulatome versus the sphincterotome and found that initial biliary access was obtained in 67% in the former versus 97% in the latter groups.3 In addition the mean number of cannulation attempts and time to achieve selective biliary cannulation was also less in the sphincterotome group. Given these considerations most American endoscopists use sphincterotomes for initial biliary cannulation.4 Most modern sphincterotomes have two to three lumens to enable simulatenous passage of contrast and wire into the bile duct. Traditionally, contrast was first injected to confirm that the bile duct as opposed to the pancreatic duct had been accessed. In 2004 Lella randomized 400 patients to the standard approach of contrast injection versus the passage of the guidewire to confirm bile duct cannulation.5 In the contrast group 4% of patients developed pancreatitis compared to none in the wire guided group and high pancreatic enzyme levels (5 times the upper limit of normal) developed in 20% of contrast guided versus 5% of wire guided cases. Subsequent large randomized trials have consistently demonstrated post ERCP pancreatitis rates of 2-9% in wire guided versus 10-17% in contrast guided ERCP.6-7 Wire guided cannulation has also been shown to shorten cannulation and fluoroscopy times.8 1.2 Additionally, several new technologies have been introduced with potential but unproven advantages. Traditionally the ERCP wires have been as long as the combination of the scope channel and accessory length such that an assistant could control placement during the procedure. Recently, short wire system have been developed in which wires may be stripped out of catheters with tear away lumens down to the distal most aspect (5-50cm) depending on the accessory.9 A stripped down sphincterotome accomodates a short wire which can be completely controlled by the endoscopist enabling physician controlled wire passage. A locking device on the scope allows the short wire to be fixed reliably during therapeutic maneuvers. The advantage of the short wire system is that it decreases reliance on an assistant. The disadvantage is that is may decrease the backpressure on the sphincterotome which may be achieved based on Newton's third law (pulling the wire back helps advance the tome forward). There has been little published work comparing short wire to traditional long wire systems. Available data suggests trends toward shorter fluoroscopy and procedure times.10 Additionally, the short wire system appears to enable shorter time to bile duct system placement compared to the long wire system. A recent national survey revealed that 43% of endoscopists favored the short wire system compared to 33% the long wire system.4 1.3 An additional technology frequently used to facilitate cannulation are smaller sized sphincterotomes. Theoretically, these small diameter tomes may enable access into bile duct with more diminutive orifices. One small randomized trial compared standard 5Fr tomes to 4Fr tomes. Cannulation in both groups was 84% though the study was prematurely terminated.11 In patients in whom biliary access is still not achieved often another type of sphincterotome or cannulatome will be used. If this does not work a wire or pancreatic stent will often be placed in the pancreatic duct ideally to change the orientation of the papilla in a favorable way.12 The latter strategy will also decrease the risk of post ERCP pancreatitis in difficult cases including those who have undergone multiple cannulation attempts.13 If these approaches fail an access or precut papillotomy may be performed using a needle knife. Precut sphincterotomy in the direction of the bile duct can either be performed from the orifice (gough) or above the orifice (fistulotomy).12, 14 Currently both the small and standard tomes are routinely used at LAC. The Rx System can be converted from long to short wire method using the same device. Both long and short wire approaches are routinely performed at LAC. There is no standardized approach. 2.0 OBJECTIVES AND PURPOSE To assess whether initial bile duct cannulation with a 39Fr versus 44Fr sphincterotomes and the long verus short wire systems favors successful bile duct cannulation and minimizes complications. This will help to developed an evidence based approach to initial bile duct access during endoscopic retrograde cholangiopancreatography (ERCP). 3.0 STUDY DESIGN: The study will be a prospective randomized controlled trial. All patients undergoing ERCP for standard indication including stones, leaks or jaundice will be eligible. Those who have undergone prior sphincterotomy, which ensures biliary access will be excluded. Additionally those with prior ERCP and failed bile duct cannulation will be excluded. Additionally, patients who require ERCP only for pancreatic indications such as pancreatic duct leak will be exlcluded as bile duct cannulation is not required for clinical indications. Patients will be randomized using a 2 x 2 factorial design to initial cannulation attempt with a small versus standard sphincterotome (1:1) and short versus long wire system (1:1). The Rx Biliary System (Boston Scientific; Natick, MA) will be used. In those randomized to the short wire approach, the wire will be stripped from the tear away lumen to convert from a long to short wire system. The wire will be controlled by the endoscopist "physician controlled." In those randomized to the long wire approach the long wire will be left in the lumen which will not be stripped down. The wire will be controlled by the assistant. Those randomized to the small sized tome will initially use the 39Fr tome and those randomized to the standard tome will use the 44Fr tome. In all 4 groups initial successful cannulation will be defined as the acquisition of deep bile duct access within the first 8 attempts. An attempt is defined as the placement of the cannula into the biliary orifice followed by advancement of the wire or contrast injection. At LAC+USC procedures are performed by 2-3 third year fellows who are recognized as clinically and technically excellent. Fellows at other levels; first year and second year fellows are not allowed to perform ERCP. All procedures are directly supervised by a faculty physician who takes over cannulation if the fellow is not successfully after 3-5 attempts. Faculty physicans perform the procedures alone only under rare circumstances. However, to account for this possibility randomization will be stratified by 3rd fellow+faculty and faculty alone. To account for variability of assistants we will make certain that the policy that nursing assistants rotate on monthly basis is enforced which will ensure an even distribution of technical assistance which reflects the real word scenario. If biliary cannulation is not achieved after 8 attempts any wire system or sphincterotome size may be used at the discretion of the attending endoscopist. Thus the assigned tome and wire approach may be abandoned or continued at the discretion of the faculty member performing the case. Additionally, pancreatic wire or stent placement or precut sphincterotomy will be performed at the discretion of the attending endoscopist. The primary outcome will be successful initial bile duct cannulation as defined by cannulation with the assigned technology within the first 8 attempts. The secondary outcomes will include total cannulation and fluoroscopy times, inadvertent, pancreatic duct access, and ERCP complications All patients who are inpatient will be seen in follow-up on the second procedure day and data to assess for outcomes will be collected by the study team. They will be called 7 days after the procedure. Those who are discharged on the day of the procedure will be called 1 day after the procedure and 7 days after the procedure to assess for these outcomes. Thus all patients will be assessed either with a follow up visist or phone call at 1 and 7 days post procedure. 4.0 STATISTICAL CONSIDERATONS: A 2 x 2 factorial design will be used with a 1:1 randomization to 44Fr versus 39Fr tome and 1:1 randomization to long or short wire system. Our a priori hypothesis is that there will be a 10% difference in successful initial cannulation (85% to 95%). At an alpha of 0.05 and beta of 0.10 a sample size of 498 will be needed to demonstrate a difference. A planned interim analysis will be performed after 200 patients have been enrolled. The rule for stopping are as follows; A) If significant difference in the primary outcome (cannulation within 8 attempts) is reached in one arm (wire length or cannulatome size) that arm will be stopped. An interim test of the effect of administration will be computed to determine the sample size for the remaining arm with a 2 group design moving forward. B) If both a superior wire length and cannulatome size strategy (both arms) are demonstrated after 200 cases the trial will be stopped and reported. C) If significance is reached in neither arm at 200 cases, then the 4 groups will continue until the sample size of 500 is reached. D) If significantly more complications including pancreatitis, perforation, bleeding, cholangitis are seen in one arm (wire length or cannulatome size) that arm will be stopped. Dichotomous outcomes will be compared using a Fischer's exact or chi squared test and continuous outcomes will be compared using a Wilcoxan rank sum test. Statistical analyses will be performed by biostatisticians at the Biostatistical and Bioinformatics Resource Group within the SC CTSI.