Trends in outcomes following pancreaticoduodenectomy for pancreatic adenocarcinoma from 2005-2016: improvement has unfortunately not been uniform across hospitals

Abstract

Presenter: Christopher T Aquina MD, MPH | The Ohio State University Background: A “textbook oncologic outcome” (TOO) is a composite measure representing the “ideal” outcome for patients undergoing cancer surgery. Although operative technique for pancreaticoduodenectomy for PDAC has improved, the impact on TOO is unknown. We assessed national trends in TOO and 1-year mortality following pancreaticoduodenectomy for PDAC. Methods: Patients who underwent curative-intent pancreaticoduodenectomy for PDAC were identified within the National Cancer Database (2005-2016). TOO was defined as: adequate lymph node yield (≥12), R0 resection, non-length of stay (LOS) outlier ( < 25 days), no hospital readmission, no 90-day mortality, and receipt of chemotherapy. Trends in outcome measures (2005-2007 vs. 2014-2016), TOO, and 1-year mortality were compared using bivariate and mixed-effects analyses. Results: Among 40,342 patients, there was significant improvement in all outcome measures except rate of readmission (adequate lymph node yield: 49.1%->76.7%; R0 resection: 75.6%->78.8%; LOS outlier: 8.9%->5.2%; 90-day mortality: 7.6%->5.5%; chemotherapy receipt: 59.2%->76.4%; TOO: 16.3%->34.3%; 1-year mortality: 34.8%->28.6%) (all p<0.001). Across 418 hospitals, wide variation was present in baseline risk-adjusted rates of TOO (median = 15.7%, range = 6.9%-50.8%) and 1-year mortality (median = 32.9%, range = 17.1%-49.8%) during 2005-2007. Wide hospital-level variation was also present in the degree of improvement in adjusted TOO (OR range = 0.67-8.32) and 1-year mortality (OR range = 0.44-2.05) over time. High hospital volume (≥20 pancreatic cancer resections/year) was the only hospital factor associated with greater improvement in adjusted TOO (median OR: 3.23 vs 2.69, p<0.001) and 1-year mortality (median OR: 0.75 vs 0.83, p = 0.006). Conclusion: There have been significant improvements in outcomes following pancreaticoduodenectomy for PDAC. However, wide variation in outcomes and improvement persists across hospitals, particularly among low-volume hospitals.