WHipple-ABACUS, a simple, validated risk score for 30-day mortality after pancreaticoduodenectomy developed using the ACS-NSQIP database.


Conference Paper

BACKGROUND: Pancreaticoduodenectomy needs simple, validated risk models to better identify 30-day mortality. The goal of this study is to develop a simple risk score to predict 30-day mortality after pancreaticoduodenectomy. METHODS: We reviewed cases of pancreaticoduodenectomy from 2005-2012 in the American College of Surgeons-National Surgical Quality Improvement Program databases. Logistic regression was used to identify preoperative risk factors for morbidity and mortality from a development cohort. Scores were created using weighted beta coefficients, and predictive accuracy was assessed on the validation cohort using receiver operator characteristic curves and measuring area under the curve. RESULTS: The 30-day mortality rate was 2.7% for patients who underwent pancreaticoduodenectomy (n = 14,993). We identified 8 independent risk factors. The score created from weighted beta coefficients had an area under the curve of 0.71 (95% confidence interval, 0.66-0.77) on the validation cohort. Using the score WHipple-ABACUS (hypertension With medication + History of cardiac surgery + Age >62 + 2 × Bleeding disorder + Albumin <3.5 g/dL + 2 × disseminated Cancer + 2 × Use of steroids + 2 × Systemic inflammatory response syndrome), mortality rates increase with increasing score (P < .001). CONCLUSION: While other risk scores exist for 30-day mortality after pancreaticoduodenectomy, we present a simple, validated score developed using exclusively preoperative predictors surgeons could use to identify patients at risk for this procedure.

Full Text

Duke Authors

Cited Authors

  • Gleeson, EM; Shaikh, MF; Shewokis, PA; Clarke, JR; Meyers, WC; Pitt, HA; Bowne, WB

Published Date

  • November 2016

Published In

Volume / Issue

  • 160 / 5

Start / End Page

  • 1279 - 1287

PubMed ID

  • 27544541

Pubmed Central ID

  • 27544541

Electronic International Standard Serial Number (EISSN)

  • 1532-7361

Digital Object Identifier (DOI)

  • 10.1016/j.surg.2016.06.040

Conference Location

  • United States