ERCP is used selectively in patients with acute biliary pancreatitis (ABP). In patients with ABP, ERCP often is difficult and has the potential to cause further damage. In addition, the prevalence of residual choledocholithiasis in ABP is low (<30%). EUS and MRCP accurately diagnose choledocholithiasis, but the performance of MRCP may be inferior in ABP. EUS, with ERCP when a stone is seen, has been shown to be feasible. This study assessed the relative costs and outcomes of EUS and MRCP in patients with ABP compared with standard care involving selective ERCP.

A decision tree was constructed, modeling standard care for nonsevere ABP (selective ERCP) and severe ABP (ERCP with sphincterotomy and balloon sweep). The other arms included either EUS or MRCP first, with the conversion to or the addition of ERCP when a bile-duct stone was seen. Probabilities and accuracy of EUS and MRCP were taken from published data. Costs were locally quantified in Canadian dollars (CDN), including nursing/technical/professional personnel, equipment maintenance, and disposable equipment. The robustness of assumptions was tested by sensitivity analyses.

Overall, EUS in all patients with ABP was marginally dominant compared with standard care with selective ERCP ($58 CDN per patient less expensive; 0.9% fewer cases of pancreatitis [ERCP-related or recurrent]). In the severe ABP subgroup, EUS was more clearly dominant ($742 CDN per patient less expensive; 3% fewer cases of pancreatitis), and the nonsevere subgroup had an incremental cost-effectiveness ratio of $17,000 per case of pancreatitis avoided. MRCP was more expensive than EUS in both subgroups.

EUS is dominant in severe ABP. In nonsevere ABP, it is slightly more costly but is associated with fewer ERCPs and ERCP-related complications. A randomized trial would help to quantify the benefits of avoiding ERCP in these patients.