Previous studies have suggested a relationship between bile duct stones and periampullary duodenal diverticula. The aim of the present study was to examine this association in more detail, including step-wise logistic regression to identify independent predictors for the presence of diverticula. Clinical, endoscopic and radiological data were analysed from 794 consecutive subjects with bile duct stones; 44 with primary duct stones (without gall-bladder stones) and 750 with bile duct stones associated with current or previous gall-bladder stones. Frequencies of diverticula in subjects with bile duct stones were compared with those in age-matched controls who had undergone upper gastro-intestinal endoscopy using a side-viewing endoscope (n = 100) or who had undergone endoscopic retrograde cholangiopancreatography (ERCP) for pancreatic cancer (n = 100). The frequencies of diverticula in subjects with primary duct stones (70%) and bile duct stones with current or previous gall-bladder stones (25%) were significantly higher than in both control groups (7 and 8%, respectively). When subjects with bile duct stones were analysed by step-wise logistic regression, age and bile duct diameter were independent predictors of the presence of diverticula, but diverticula were unrelated to gender, mode of presentation, number of bile duct stones and outcome after endoscopic sphincterotomy. A significant association exists between duodenal diverticula and bile duct stones, perhaps because diverticula predispose to stone formation through the combined effects of bile stasis and bacterial contamination of bile.