Background: Current stratification systems for patients presenting with acute type A aortic dissection rely on signs of malperfusion to predict mortality. The authors sought to develop an algorithm to readily risk stratify these patients using admission characteristics. Methods: Two hundred sixty-nine consecutive patients who underwent type A repair between 2002 and 2015 were reviewed for easily obtainable preoperative demographics and laboratory values deemed a priori as potential predictors of operative mortality. Multiple logistic regression analysis was performed to determine independent significance, and linear regression was performed to generate the concomitant regression expression of the variables significant on bivariate analysis. Results: Operative mortality was 16% (43/269) and was 29% (34/119) among patients who presented with malperfusion. Upon multivariate analysis, creatinine (p = 0.008), liver malperfusion (p = 0.006), and lactic acid level (p = 0.0007) remained independent significant predictors. Regression coefficients allowed the generation of a risk score as 5.5 × (lactic acid [mmol/L]) + 8 × (creatinine [mg/dL]) ± 8 (+ if liver malperfusion presents, – if no liver malperfusion). Upon receiver-operating characteristic curve analysis this model generated a c-statistic of 0.75. Operative mortality among patients within the lowest tertile (risk score < 7) was 4%, whereas patients in the middle (7 to 20) and highest (≥20) tertiles had mortality rates of 14% 37%, respectively. Conclusions: Although still requiring external validation, the innovative risk score presented necessitates knowledge of lactic acid, serum creatinine, and liver function tests. The algorithm predicts operative mortality with high accuracy and offers clinicians a novel tool to improve preoperative guidance and prognosis.
Bibliographical notePublisher Copyright:
© 2018 The Society of Thoracic Surgeons