Neurological Injury at Presentation Should Not Preclude Immediate Repair of Acute Type A Aortic Dissection
Peter Chiu, Torsten Rotto, Andrew B. Goldstone, Josh Whisenant, *Y. Joseph Woo, *Michael P. Fischbein
Stanford University, School of Medicine, Stanford, CA
OBJECTIVES: Neurological injury complicating the presentation of acute type A aortic dissection remains a challenge for cardiac surgeons.
METHODS: This was a retrospective review of patients undergoing open repair of acute type A aortic dissection at our institution between 1/2005 and 12/2015. Evidence of neurological injury at the time of presentation was abstracted from the medical record along with preoperative demographic and clinical data. Propensity score matching was then used to create comparable groups with and without neurological injury, and odds of in-hospital mortality was compared using logistic regression. Separately, among patients with neurological injury, fisher's exact test was used to compare persistent and transient neurological deficits. Among patients with persistent neurological deficits, a threshold for time-to-operation was evaluated using receiver operating characteristic (ROC) curves, and 95% confidence intervals were estimated with 500 bootstrap replicates.
RESULTS: There were 345 patients who underwent open repair for AcA-AoD, and 50 patients (14.5%) presented with neurological symptoms. In the matched analysis, in-hospital mortality was greater among patients who presented with neurological deficits (22.0%, 11 of 50) compared to those without (6%, 3 of 50), OR: 4.42 (95% CI: 1.15 to 16.97, p = 0.03). Among patients whose presentation included neurological symptoms, those presenting with persistent neurological deficits experienced increased mortality (35.5%, 11 of 31) compared with transient neurological deficits, (0%, 0 of 19), p = 0.003. For patients with persistent neurological deficits at presentation, complete resolution of symptoms was seen in 35.5% of patients (11 of 31). In contrast, among patients in the transient neurological deficit group, none of the preoperative deficits persisted into the postoperative period, but 21.1% (4 of 19) experienced post-operative stroke. Among patients with persistent neurological deficits at presentation, ROC curve analysis suggested that a threshold of 14.5 hours between symptoms and the start of the operation may predict (AUC 0.75, 95% CI: 0.59 to 0.92) neurological improvement. However, no patients operated on beyond 14.5 hours died. In our ROC curve analysis, there was no threshold for time-to-operation that predicted death (AUC 0.35, 95% CI: 0.19 to 0.53). Of the 50 patients with neurological symptoms at presentation, 24 underwent postoperative head imaging, and 2 patients (8.3%) demonstrated evidence of hemorrhagic conversion.
CONCLUSIONS: Neurological injury at the time of presentation with acute type A aortic dissection was associated with an increased risk of in-hospital mortality at a tertiary referral center, and patients with persistent neurological deficits fared worse than patients who presented with only transient neurological deficits. Among patients with persistent neurological deficits, a threshold of 14.5 hours, which is more liberal than currently reported in the literature, may yield prognostic information regarding neurological improvement post-operatively. However, despite the prognostic implications of time-to-operation, the lack of a threshold with respect to mortality suggests that patients beyond 14.5 hours should still proceed to the operating room for an expedient repair.
Back to 2018 Program