OBJECTIVE Dysfunction of the sphincter of Oddi may be a cause of persistent problems after cholecystectomy. The aim was to find out whether various factors are of value in predicting abnormal manometric results and thus aid in deciding whether endoscopic manometry is indicated. METHODS 97 patients were investigated prospectively (13 men, 84 women; mean age 50.2 [29-72] years) in which endoscopic cholangiopancreatography (ERCP) had not revealed any cause of the biliary complaints. The patients were divided into three types according to four criteria: (1) history (biliary colics); (2) biochemistry (cholestasis); (3) dilated biliary tract (at ERCP); (4) contrast retention in biliary tract (at ERCP). Type I: all four criteria present; type II: positive history and one or two other criteria; type III: biliary colic only. Endoscopic manometry (EM) was performed after classification of the patients. RESULTS EM was successful in 83 of 97 patients (86%). All 15 patients of type I had sphincter of Oddi dysfunction (SOD) defined as basic sphincter pressure > 40 mm Hg. SOD was demonstrated in only 23 of 38 patients of type II (61%) and 15 of 30 patients of type III (50%) (P < 0.01). Increased rate of sphincter of Oddi contraction ("tachyoddi") was demonstrated in only four patients (4%), in two of them in combination with an increased basic sphincter pressure. Mild to moderate pancreatitis occurred within 24 hours of manometry in ten of 83 patients. CONCLUSIONS Endoscopic manometry is not necessary in patients of type I for confirming the diagnosis. But it is obligatory for patients of type II and III, because the demonstration of SOD has therapeutic consequences, and should be performed in clinical studies, if possible.