Abstract
INTRODUCTION: The optimal treatment for biliary obstruction in pancreatic cancer remains controversial between surgical bypass and endoscopic stenting. METHODS: Retrospective analysis of unresected pancreatic cancer patients in the Healthcare Cost and Utilization Project Florida State Inpatient and Ambulatory Surgery databases (2007-2011). Propensity score matching by procedure. Primary outcome was reintervention, and secondary outcomes were readmission, overall length of stay (LOS), discharge home, death and cost. Multivariate analyses performed by logistic regression. RESULTS: In a matched cohort of 622, 20.3% (63) of endoscopic and 4.5% (14) of surgical patients underwent reintervention (p < 0.0001) and 56.0% (174) vs. 60.1% (187) were readmitted (p = 0.2909). Endoscopic patients had lower median LOS (10 vs. 19 days, p < 0.0001) and cost ($21,648 vs. $38,106, p < 0.0001) as well as increased discharge home (p = 0.0029). No difference in mortality on index admission. On multivariate analysis, initial procedure not predictive of readmission (p = 0.1406), but early surgical bypass associated with lower odds of reintervention (OR = 0.233, 95% CI 0.119, 0.434). DISCUSSION: Among propensity score-matched patients receiving bypass vs. stenting, readmission and mortality rates are similar. However, candidates for both techniques may experience fewer subsequent procedures if offered early biliary bypass with the caveats of decreased discharge home and increased cost/LOS.