External Validation of a Rapid Ruptured Abdominal Aortic Aneurysm Score

Background: The Rapid Ruptured Abdominal Aortic Aneurysm Score (RrAAAS) was developed from Vascular Study Group of New England (VSGNE) data (649 ruptured abdominal aortic aneurysm (rAAA) patients, repaired both open and endovascularly), using preoperative age, creatinine, and blood pressure. This study validates that model using the larger National Vascular Quality Initiative (VQI) data set and compares its performance to previous models. Methods: The VQI registry was queried for patients undergoing rAAA repair from 2006 to 2016. The performance of our original model, RrAAAS, was tested on this data set excluding VSGNE patients (VQI minus VSGNE), and its performance was then compared to the performance of the Glasgow Aneurysm Score (GAS) and Edinburgh Ruptured Aneurysm Score (ERAS). Results: VQI contained 2,704 eligible patients, of which 715 had been contributed by VSGNE. The discrimination of RrAAAS was similar to GAS or ERAS (area under a receiver operator characteristic curve = 0.66). Neither GAS nor ERAS provides a direct prediction of mortality; observed mortality in the VQI minus VSGNE cohort tended to be somewhat lower than predictions of the original RrAAAS. A recalibrated equation predicting the percent mortality was Mortality (%) = 16 + 12*(age > 76) + 8*(creatinine > 1.5) + 20*(systolic blood pressure < 70). Conclusions: The previously described RrAAAS has similar discrimination as the GAS and ERAS, is easier to obtain in an emergency setting, and has been recalibrated to reflect the experience of a large national sample. The RrAAAS could be useful for clinicians caring for these patients and could be used for risk adjustment when comparing regional differences in mortality associated with rAAA repair.