Background/Aims: Endoscopic ultrasonography is a promising procedure f
or the diagnosis of extrahepatic cholestasis. Accuracy for the diagnos
is of choledocholithiasis by ultrasonography and computed tomography w
ere prospectively compared with endoscopic ultrasonography in 62 conse
cutive patients. Methods: Final diagnosis was determined by endoscopic
retrograde cholangiography with or without sphincterotomy or intraope
rative cholangiography with or without choledochoscopy. All of the pat
ients had abdominal ultrasonography, computed tomography, endoscopic u
ltrasonography, and either an endoscopic retrograde (n = 40) or intrao
perative cholangiography (n = 32) performed. Results: Choledocholithia
sis was confirmed in 22 patients. Thirteen patients had a stone with a
diameter <1 cm, and 14 had a nonenlarged common bile duct. Endoscopic
ultrasonography was more sensitive (97%) than ultrasonography (25%; P
< 0.0001) and computed tomography (75%; P < 0.02). Specificity and po
sitive predictive value were not significantly different. Negative pre
dictive value of endoscopic ultrasonography (97%) was better than that
of ultrasonography (56%; P < 0.0001) and computed tomography (78%; P
< 0.02). Results were unchanged after six patients in whom the absence
of choledocholithiasis was considered probable after follow-up were e
xcluded. Endoscopic ultrasonography results did not depend on stone di
ameter or common bile duct dilatation. Conclusions: Endoscopic ultraso
nography appears to be the best diagnostic tool for the diagnosis of c
holedocholithiasis compared with other noninvasive procedures.