Background: The performance of the Emergency Surgery Score (ESS), a validated risk calculator, in the elderly emergency general surgery (EGS) patient remains unclear. We hypothesized that ESS accurately predicts outcomes in elderly EGS patients, including octogenarians and nonagenarians. Methods: Using the 2007-2017 National Surgical Quality Improvement Program (NSQIP) database, we included all EGS patients >= 65 years old. The correlation between ESS, mortality and morbidity was assessed in the 3 patient cohorts (>65, octogenarians and nonagenarians), using the area under the curve (AUC). Results: A total of 124,335 patients were included, of which 34,215 (28%) were octogenarians and 7239 (6%) were nonagenarians. In patients >= 65 years, ESS accurately predicted mortality (AUC 0.81). For octogenarians and nonagenarians, ESS predicted mortality moderately well (AUC 0.77 and 0.69, respectively. Conclusion: ESS accurately predicts mortality and morbidity in the elderly EGS patient, but its accuracy in predicting morbidity decreases for nonagenarians. (C) 2020 Elsevier Inc. All rights reserved.