Biliary manometry, scintigraphy or empiric sphincterotomy for suspected sphincter of Oddi dysfunction: A decision analysis

Background. Patients with postcholecystectomy right upper quadrant/epigastric pain without objective evidence of pancreatobiliary disease (so-called Type III patients) pose a major diagnostic and therapeutic dilemma. Although these patients are usually suspected as having sphincter of Oddi dysfunction (SOD), the clinical approach to this condition is fraught with many uncertainties. Biliary manometry, considered the "gold-standard" for the diagnosis, is a difficult and relatively high-risk procedure. The accuracy of alternative techniques such as hepatobiliary scintigraphy is debated. Finally, empiric sphincterotomy (EmS) may lead to too many unnecessary procedures with a high rate of complications. Methods. We performed decision analysis by pooling published estimates and by assigning utilities of outcomes. The following assumptions were used for our base case: prevalence of SOD = 0.3, scintigraphy sensitivity/specificity = 0.75/0.95, manometry sensitivity/specificity = 1.0/1.0, response rate to EmS (if SOD was present) = 0.6, EmS complication rate = 0.10, EmS + manometry complication rate = 0.15. Utility values were assigned as follows: cure = 1.0, procedure without cure = 0.75, continued symptoms without procedure = 0.5 and complications = 0. Results. The overall values for our base case were as follows: scintigraphy = 0.73; EmS = 0.69 and manometry = 0.60. However, for SOD prevalence rates between 27 to 38%, EmS is the optimal decision unless associated with a complication rate > 10%, in which case scintigraphy is better. Conclusions. Our analysis suggests that the approach to these patients can be tailored to the prevalence of SOD in the clinical practice, and the local complication rates of sphincterotomy. If a local center has sphincterotomy complication rates greater than 10%, initial scintigraphy is optimal at any prevalence of disease. Empiric sphincterotomy may be the preferred approach given a lower complication rate and a SOD prevalence of 27 to 38%. However, the decision may be further driven towards empiric sphincterotomy if specificities of scintigraphy are lower or if lower utility values are assigned to untreated disease. Biliary manometry appears to be the least valuable procedure in this group of patients in most clinical circumstances.