BACKGROUND Patients referred to general surgeons for the treatment of gall-bladder stones were studied to evaluate the role of sincalide cholescintigraphy as a gall-bladder stress test in an effort to identify a group of patients whose pain was non-biliary in origin and who would not be improved by cholecystectomy. METHODS Ten asymptomatic controls and 57 patients with gallstones and abdominal symptoms were studied. All patients were interviewed by an independent assessor who identified a group of patients in whom the role of gallstones in their presentation was uncertain (clinically possibly biliary group). All patients and controls underwent sincalide cholescintigraphy. The surgeons remained blinded to the study results throughout the study period. All patients were re-evaluated 6-12 months later to establish the ultimate diagnosis based on their therapeutic response. RESULTS Several parameters of gall-bladder function were studied from analysis of the sincalide cholescintigram. Lag time, ejection period, ejection rate and ejection fraction did not differ significantly among controls, patients proven to have non-biliary disease and patients proven to have biliary disease. There were significant differences in mean gall-bladder filling fraction between proven biliary and proven non-biliary groups. However, the group of patients with clinically possibly biliary symptoms could not accurately be separated into those who benefited from cholecystectomy and those who improved without surgery on the basis of this parameter. CONCLUSIONS Significant differences in gall-bladder filling fraction between symptomatic and asymptomatic gallstone patients were identified suggesting reduced gall-bladder compliance in symptomatic patients. However, the sincalide cholescintigram failed to emerge as a useful gall-bladder stress test. Even in the 1990s, assessment by an experienced surgeon appears to be the most appropriate way to select patients for cholecystectomy.