The current randomized study was done to compare the results of choledochoduodenostomy (CD) and choledochojejunostomy (CJ) for choledocholithiasis, with special reference to long term results and the risk of ascendant cholangitis. From January 1978 to January 1990, 130 patients were included in the study--64 with CD (side to side, in all patients) and 66 with CJ (side to side in 25 patients and end to side in 41). No significant difference was observed between the CD and CJ groups for postoperative mortality (3.8 percent) and morbidity rates. One hundred and twenty patients (58 CD and 62 CJ) were available for long term follow-up evaluation (mean follow-up period of 29 +/- 11 months). One hundred and seven patients had no symptoms attributable to biliary disease or operation. Five patients in this group died of unrelated causes. Thirteen patients experienced biliary symptoms suggestive of cholangitis, or at least related to the bilioenteric anastomosis--six patients in the CD group and seven in the CJ group. Cholangitis was observed in the first postoperative year in eight of these 13 patients and during the second year for the five others. In the CD group, cholangitis was the result of sump syndrome (n = 3), anastomotic stricture (n = 1) and unknown causes (n = 2). In the CJ group, cholangitis was the result of anastomotic stricture (n = 3), residual intrahepatic stones (n = 1) and unknown causes (n = 3). The results of the current study confirm the good long term results of both procedures. However, it suggests that CD is preferable for choledocholithiasis for two reasons--it is technically easier and faster to perform than CJ and, unlike CJ, CD permits easy access to further endoscopic exploration or treatment if necessary.