Prospective study to develop an algorithm for investigation by endoscopic retrograde cholangiopancreatography or magnetic resonance cholangiopancreatography.

Background: Identifying cost-efficient and patient-friendly pathways for those who present with suspected pancreatico-biliary disease remains a challenge. Algorithms must be tailored to improve decision-making. We assessed suitable criteria from which an algorithm for selection for endoscopic retrograde cholangiopancreatography (ERCP) or magnetic resonance cholangiopancreatography could be developed.
Methods: Data on clinical, ultrasound and liver function test findings and outcome were recorded for consecutive patients undergoing ERCP and patients were stratified into different indications and a therapeutic ratio (TR) obtained for each (TR = number of therapeutic ERCP/total number of ERCP).
Results: One hundred and twenty ERCP were attempted with 112 cannulations (93.3%). Seventy-one therapeutic procedures were attempted with 64 (90.1%) successes. Forty-two (35%) investigations were normal. Seven (6%) patients suffered complications. Thirteen indications were used. The TR varied according to the indication. The TR for jaundice with biliary dilatation was 0.85. In contrast, asymptomatic patients with deranged liver function test and normal gall bladder on USS had a TR of 0.17. Gallstones with cholestatic liver function test yielded a TR of 0.54. A TR of 0.7 was considered an appropriate 'cut-off' for triage to ERCP or magnetic resonance cholangiopancreatography. An algorithm was generated based on these data. From this cohort, 50 (42%) patients would have been more appropriately investigated by magnetic resonance cholangiopancreatography, although 20 (40%) would have required therapy afterwards.
Conclusions: An algorithm that separates indications by TR (TR > 0.7) may provide a cost-efficient, patient-friendly pathway for investigation and improve the use of resources.