Selective cannulation of the common bile duct: a prospective randomized trial comparing standard catheters with sphincterotomes

Background: Current recommendations for the use of standard catheters or sphincterotomes for the initial attempt at selective common bile duct cannulation have been made in the absence of prospective comparative data. Methods: A prospective study was carried out in which patients were randomized to undergo cannulation with a standard catheter or a sphincterotome (standard or wire-guided). Multivariate models were constructed to determine significant independent predictors of the success rates of initial and selective cannulation and the number of attempts and time needed to achieve selective cannulation. Results: Eighty-three successive patients were evaluated; 36 were excluded because they had undergone previous therapeutic endoscopic retrograde cholangiopancreatography (ERCP) or a Billroth II operation. Of the 47 patients (28 women, mean age 60.6 +/- 14.5 years), indications for ERCP included suspected bile duct stones in 41 patients, pancreatico-biliary malignancies in 4, and biliary leaks in 2. Eighteen patients were randomized to undergo selective common bile duct cannulation with standard catheter and 29 to standard/wire-guided sphincterotome. Initial common bile duct cannulation for the standard catheter and standard/wire-guided sphincterotome groups was successful in 12 (67% [95% CI: 41%, 87%]) and 28 (97% [95% CI: 82%, 100%]) patients, respectively (95% CI for the difference: -0.57 to -0.03, p = 0.009). Using intention to treat analysis, selective common bile duct cannulation was successful for standard catheter and standard/wire-guided sphincterotome patients in 17 (94% [95% CI: 73%, 99%]) and 28 (97% [95% CI: 82%, 100%]) cases, respectively (95% CI for the difference: -0.15 to +0.10, p > 0.05). The mean number of attempts required to achieve selective common bile duct cannulation were 12.4 +/- 6.0 and 2.8 +/- 3.1 (p = 0.0001). The mean time taken to achieve selective common bile duct cannulation was 13.5 +/- 6.14 and 3.1 +/- 5.1 minutes (p = 0.0001). Multivariate modeling revealed that the initial choice of catheter was the only significant independent predictor of the time taken and the number of attempts performed to achieve selective common bile duct cannulation (p = 0.0001 for each model). Conclusion: The use of standard/wire-guided sphincterotome was superior to that of standard catheter for the initial attempt at cannulation of the common bile duct. The number of attempts required may bear clinical significance with regard to the development of post-ERCP pancreatitis and warrants further study. (Gastrointest Endosc 1999;50:775-9.)