FACTORS PREDICTIVE OF UNSUCCESSFUL ERCP: A CASE-CONTROL STUDY

SL Jowett, J Romagnuolo, S Bass, M Cole, RJ Hilsden, GR May

Department of Medicine, University of Calgary, Calgary, Alberta

BACKGROUND: Failure to cannulate and/or complete therapy at ERCP is common (10-15%) and may result in complications. Pre-cut sphincterotomy can often allow successful cannulation. If failure could be predicted, it could allow referral to a centre with special expertise for other therapeutic alternatives.
AIM: To determine factors that identify a high likelihood of ERCP failure.
METHODS: In this case-control study, failure was defined as an inability to cannulate the duct of interest and complete the required therapy. Therapeutic outcomes were pre-determined. Target case enrollment is 207. Two age- and sex- matched controls were included for each case from our endoscopic database. Predictors of failure were tested using chi-square tests and logistic regression.
RESULTS: The sample to date includes 96 failures & 192 controls (mean age 63 yrs, 56% women). Indications for ERCP were the same in both groups (p=0.19): common bile duct (CBD) stones (54%), stricture (16%) and cholangitis (7%). Pancreatic duct (PD) therapy was the aim in 4.5%. Differences in predictor variables between failed and successful ERCP are shown in the table below.

The only significant variable on multivariate analysis was the duct of interest: if PD or both ducts were targeted, failure was more likely OR 3.6 [1.5-8.6]p<0.01.
CONCLUSIONS: Contrary to common opinion many variables such as acute pancreatitis, the presence of a diverticulum, or absence of a dilated duct do not significantly predict a higher rate of failed ERCP. Procedures where the PD is of interest are more likely to fail.

Funding: SL Jowett is a Cook Canada Fellow in Therapeutic Endoscopy.