The improvement in post-operative mortality following pancreaticoduodenectomy between 2006 and 2016 is associated with an improvement in the ability to rescue patients after major morbidity, not in the rate of major morbidity

Background: The postoperative mortality rate of pancreaticoduodenectomy is decreasing over time. It is unknown whether this is related to reduction in incidence of major morbidity or failure to rescue. We aimed to make this determination.Methods: ACS-NSQIP was retrospectively reviewed from 2006 to 2016. Comparisons were assessed with Spearman's rank-order correlation test, chi-square test with linear-by-linear association, and multivariable hierarchical logistic regression.Results: Mortality decreased significantly from 2.9% to 1.5% (p < 0.001). This decrease was inde-pendent of preoperative variables on multivariable analysis (odds ratio [OR] 2.55, 95% confidence in-terval [CI] 1.55-;5.21, p < 0.001). In contrast, no change in incidence of major morbidity was seen on univariable (26.8% to 25.9%, p = 1.00) or multivariable analysis (OR 1.22, 95% CI 1.03-1.45, p = 0.060). Failure to rescue was observed to decrease on univariable (9.8% to 4.1%, p < 0.001) and multivariable analysis (OR 3.65, 95% CI 2.07-6.76, p < 0.001).Conclusion: There has been a sizeable reduction in the mortality rate after pancreaticoduodenectomy from 2006 to 2016. This predominantly results from a reduction in failure to rescue rate rather than a decrease in incidence of major morbidity.