Prediction of Survival after 48 Hours of Intensive Unit Care following Repair of Ruptured Abdominal Aortic Aneurysm-Multicentric Study for External Validation of a New Prediction Score for 30-Day Mortality

Abstract

Background: Ruptured abdominal aortic aneurysm (rAAA) remains a critical life-threatening condition. We aimed to evaluate rAAA management in our center focusing on predictors of mortality at 48 hr of intensive care unit (ICU) and to develop a new mortality prediction score considering data at 48 hr postprocedure. External validation of the modified score with patient data from independent vascular surgery centers was subsequently pursued.

Methods: Clinical data of all patients admitted in our center from January 2010 to December 2017 with the diagnosis of rAAA were retrospectively reviewed for the development of the mortality prediction score. Subsequently, clinical data from patients admitted at independent centers from January 2010 to December 2017 were reviewed for external validation of the score. Statistical analysis was performed with SPSS Version 25.

Results: A total of 78 patients were included in the first part of the study: 21 endovascular aneurysm repairs (EVARs), 56 open repairs (ORs), and 1 case of conservative management. Intraoperative mortality in EVAR and OR groups was 0% vs. 24.6%, respectively (P = 0.012). Thirty-day mortality reached 50% and 33% in the OR and EVAR groups. For patients alive at 48 hr, 30-day mortality diminished to 27.6%. Several preoperative predictors of outcome were identified: smoking (P = 0.004), hemodynamic instability(P = 0.004), and elevated international normalized ratio (P < 0.0001). Dutch Aneurysm Score and Vascular Study Group of New England Score (VSGNE) were also significant predictors of outcome (area under the receiver operating characteristic curve [ROC AUC] 0.89 and 0.79, respectively; P < 0.0001). At 48 hr of ICU stay, high lactate level, high Sequential Organ Failure Assessment score, need for hemodyalitic technique, and hemodynamic instability were significant risk predictors for 30-day mortality (P < 0.05). VSGNE score was modified with the inclusion of 2 variables: hemodynamic instability and lactate level at 48 hr and a new score was attained-Postoperative Aneurysm Score (PAS). Comparing AUC for VSGNE and PAS for patients alive at 48 hr, the latter was significantly better (AUC 0.775 vs. 0.852, P = 0.039). The PAS was applied and validated in 3 independent vascular surgery centers (AUC VSGNE 0.782 vs. AUC PAS 0.820, P = 0.027).

Conclusions: Despite recent evidence on preoperative predictors of survival in an era when both EVAR and OR are available, emergent decision to withhold life-saving treatment will always be extremely difficult. Therefore, the policy in our department is to try surgical repair in all cases. It remains important, however, to identify whether late deaths can be predicted, so that unnecessary prolonged treatment can be avoided. A PAS was delineated predicting 30-day mortality significantly better in patients alive at 48 hr. The score was externally applied and validated in independent centers, corroborating the score's usefulness.