Identification of CTA-Based Predictive Findings for Temporary and Permanent Neurological Dysfunction after Repair in Acute Type A Aortic Dissection

The aim of this study was to determine CT risk findings predictive of temporary neurological dysfunction (TND) and permanent neurological dysfunction (PND) after surgical repair for acute type A aortic dissection (ATAAD). A total of 255 patients (41 ± 16 years, 79% male) with ATAAD underwent aortic CT angiography (CTA) and surgical repair consecutively from January 2013 to June 2016. The CTA findings of the 255 patients for the thoracic aorta and carotid artery were analysed to identify risk factors predictive of TND and PND. Thirty-eight patients (15%) suffered TND, and 18 (7%) exhibited PND. Common carotid artery (CCA) dissection (OR = 4.63), lower enhancement of unilateral ICA (OR = 3.02) and aortic arch tears (OR = 2.83) were predictors of postoperative TND, while PND was best predicted by retrograde ascending aorta (aAO) dissection (OR = 5.62) and aortic arch tears (OR = 6.74). In CCA dissection, the extent of the entire CCA and proximal ICA (P = 0.014), a low-enhancement false lumen with re-entry (P = 0.000) and a severely narrowed true lumen without re-entry (P = 0.005) significantly increased the risk of postoperative TND. In patients with ATAAD, specific CT findings allow the individual risk of postoperative TND and PND to be identified and may guide subsequent surgical management.

as well as mortality have been evaluated in patients with ATAAD [5][6][7][8][9] , the correlation between CTA findings in ATAAD and neurological complications has rarely been discussed.
Here, we hypothesize that specific CTA findings in ATAAD are associated with postoperative neurological dysfunction. If so, CTA imaging may help individualize perioperative management and focus meticulous attention on critical procedural details. Thus, the aim of the present study was to determine the predictive value of CTA risk findings for postoperative TND and PND.

Results
Patient Characteristics. The demographics, clinical features, surgical procedures and intraoperative data of the subjects in the three groups are displayed in Table 1. The only significant difference between the groups was the occurrence of acute renal failure. There were no significant differences in terms of age, sex, body mass index, previous medical history, surgical procedures, cerebral protection method or operation time between the three groups (P > 0.05 for all comparisons). Neurological Complications. The overall incidence of postoperative neurological complications after aortic repair was 22% (56/255). In-hospital postoperative PND occurred in 18 out of 255 patients (7.1%), and TND was detected in 38 patients (14.9%). Four patients with PND (22.2%, 4/18) died, but none of the patients with TND died.  Tables 2 and 3. The CTA findings included in the analysis are as follows: diameter of the involved aAO, retrograde dissection in the aAO, entry tears in the aAO, diameter of aAO tears, aortic sinus dissection, entry tears in the aortic arch, diameter of aortic arch tears, entry tears in the descending aorta, common carotid artery (CCA) dissection, lower enhancement in the unilateral ICA, origin of CCA in the false lumen, VA dissection, and lower enhancement of unilateral VA. The number and percentage of each CTA finding between the TND, PND and no neurological dysfunction (NND) groups are also presented in Tables 2 and 3. The diameters of the involved aAO (51.2 mm ± 10.6 vs. 50.5 mm ± 8.0 or 53.1 mm ± 11.0, both P > 0.05), aAO tears (17.2 mm ± 7.4 vs. 20.5 mm ± 11.1 or 15.8 mm ± 10.0, both P > 0.05), and aortic arch tears (11.6 mm ± 8.5 vs. 8.9 mm ± 8.1 or 11.1 mm ± 6.9, both P > 0.05) were similar in the NND, TND and PND groups. A significantly higher incidence of entry tears in the aortic arch was found in both the TND (76.3%) and PND groups (83.3%) than in the NND group (50.3%, P = 0.003 and P = 0.007). In addition, lower enhancement in the unilateral ICA was significantly increased in both the TND (31.6%) and PND groups (88.9%) compared with that in the NND group (10.6%, P = 0.001 and P = 0.000). Furthermore, significantly more patients exhibited CCA dissection in the TND group (63.2%) than in the NND group (27.1%, P = 0.000), whereas the number of entry tears in the dAO was significantly lower in patients with TND (5.3%) than in the NND subjects (19.6%, P = 0.03). A significantly higher incidence of retrograde dissection in the aAO was also found in the PND group (44.4%) than in the NND cohort (13.6%, P = 0.002).
CTA Risk Findings for TND. Univariate analysis of CTA risk findings revealed that an entry tear in the aortic arch or the descending aorta, CCA dissection, and lower enhancement in the unilateral ICA were implicated in the occurrence of postoperative TND, as shown in Table 2. The following multivariate analysis, including significant preoperative acute renal failure, revealed that CCA dissection (OR = 4.63, P = 0.000), lower enhancement in the unilateral ICA (OR = 3.02, P = 0.021), and an entry tear in the aortic arch (OR = 2.83, P = 0.034) were independent CTA risk predictors for postoperative TND (Table 2, Fig. 1).

CTA Risk Findings for PND.
Univariate analysis of CTA risk findings revealed that retrograde dissection of the aAO, an entry tear in the aAO or aortic arch, and low attenuation in the unilateral ICA are implicated in the occurrence of postoperative PND, as shown in Table 3. After multivariate analysis, the significant independent CTA risk predictors of postoperative PND were retrograde dissection in the aAO (OR = 5.62, P = 0.003) and an entry tear in the aortic arch (OR = 6.74, P = 0.006) ( Table 3, Fig. 2).
Correlation between CTA Characteristics of CCA Dissection and TND. Because CCA dissection was indicated to be an independent CTA risk finding for postoperative TND, the detailed CTA characteristics of the involved CCA were further analysed to identify the correlation between these findings and TND. The location, extension and characteristics of the CCA dissections are shown in Table 4. There was no significant difference in the location of the dissection on the right side, left side or bilateral CCA between the TND and NND groups (all P > 0.05). The extension of dissections involving the proximal CCA and the entire CCA did not differ significantly between the two groups, whereas patients in the TND group exhibited a higher incidence of dissections involving the entire CCA and proximal ICA (P = 0.014) (Fig. 3). When the dissection involved the CCA, re-entry in the CCA combined with visually lower enhancement or low attenuation matter in the false lumen (Fig. 3), or no re-entry in the CCA combined with a severely narrowed true lumen compressed by a thrombosed false lumen (Fig. 3) increased the risk of postoperative TND.

Discussion
In this retrospective study of patients with ATAAD, we observed that CCA dissection, lower enhancement in the unilateral ICA, and aortic arch tears are independent CTA risk predictors for postoperative TND, while both retrograde aAO dissection and aortic arch tears are independent CTA risk predictors for PND. The extent of the entire CCA and proximal ICA involved, CCA re-entry and lower enhancement in the false lumen, and a severely narrowed true lumen in the CCA without re-entry significantly increased the risk of TND when the CCA was involved.  Table 3.  Postoperative neurological complications, including TND (with a reported incidence of 2.5-33%) and PND (with an incidence of 4.2-24%), remain challenging for patients with ATAAD 3,5,7,14-16 . In our study, the incidences of TND and PND were 14.9% and 7.1%, respectively, which are lower than those reported in most other studies. This discrepancy likely occurred because all operations were performed at our high-volume cardiovascular surgical centre with dedicated high-volume thoracic aortic surgeons. Moreover, all patients in our centre are cared for via an emergent green route, and the majority undergo emergency surgical repair within 8 hours of admission.
To increase sensitivity in the identification of risk factors, previous authors 3,5-7 have elected to combine PND and death as adverse outcomes (AO) for evaluation, but this approach may have confused the real risk predictors for PND or mortality. The most common causes of postoperative mortality involve cardiac aetiology, while neurological complications are less significant 17 . Therefore, only PND was evaluated in the present study.
As we suspected, CCA dissection was shown to be a major CTA risk predictor for TND but not for PND. A previous study 13 indicated that carotid artery dissection may well correlate with cerebral malperfusion, which presents strong potential for adverse outcomes in postoperative patients with ATAAD. In most cases requiring extensive arch repair, such as total arch replacement, unilateral antegrade selective cerebral perfusion is the cerebral protection method of choice 18 . In the setting of right CCA dissection with severe stenosis of the true lumen, however, bilateral two-vessel perfusion through the right axillary artery or the innominate artery and the left common carotid artery may be essential to avoid left cerebral artery insufficiency, which may result in cerebral infarction.
The detailed CTA characteristics of CCA dissection need to be carefully considered. Lower enhancement in the false lumen of the CCA with re-entry or a severely narrowed true lumen of the CCA without re-entry is Lower enhancement in the unilateral ICA is potentially predictive of TND. Several conditions contribute to this finding, such as the ICA originating from the false lumen, a severely narrowed true lumen of the CCA, or low-enhancement CCA originating from the false lumen of the IA. Lower enhancement indicates decreased blood flow in the carotid artery, which may be highly predictive of cerebral hypoperfusion. This finding suggests a therapeutic strategy that may resolve malperfusion 13,19 . Lower enhancement in the unilateral ICA also shows a high incidence of 88.9% in patients with PND. Therefore, the finding that this condition is not an independent predictor of PND was unexpected. This unexpected result may have occurred because this condition, which represents a manifestation of ipsilateral cerebral hypoperfusion, can be resolved via rapid establishment of cerebral flow and surgical repair of the dissection.
In our study, an aortic arch tear was found to significantly increase the risk of both TND and PND. It is reasonable to presume that emboli (thrombi in the false lumen or plaques ruptured during a surgical procedure) flow more easily into craniocervical arteries via the entry point, resulting in subsequent artery occlusion and cerebral infarction. Recent studies have demonstrated that strokes in patients with cervical artery dissection are most frequently associated with artery-to-artery embolization 20,21 . In the case of an aortic arch tear combined with a thrombus in the false lumen, CTA information could alert the surgeon to pay meticulous attention to critical procedural details to avoid dislodging the embolus from the tear. The incidence of stroke is likely to be effectively reduced by following these guidelines. Although dAO tears were excluded as a risk predictor due to their low occurrence, we still believe such tears have an effect on postoperative neurological dysfunction due to sharing similar mechanisms with aortic arch tears.
The finding that retrograde aAO dissection serves as an independent predictor of PND was unexpected. This condition is commonly associated with an aortic arch or dAO tear. Thrombogenesis in the false lumen is common due to the integrated intimal flap. Therefore, avoiding the dislodgment of a thrombus during surgical procedures may effectively reduce postoperative PND.
This study was limited by its retrospective nature. Additionally, there was some selection bias as well as a sample size and character mismatches between the three groups. The sample size of the patient groups with postoperative TND and PND was also small, which may influence the accuracy of the outcomes. Furthermore, our study focused on only in-hospital TND and PND, and long-term clinical outcomes may be more important for elucidating real CTA risk predictors. Finally, our CTA scanning procedure did not include the head, which might provide more useful information regarding the status of cerebral malperfusion.
The findings of the present study indicated that specific carotid and aortic CTA findings may help predict individual risk for postoperative TND and PND in patients with ATAAD, in addition to providing guidance for individual surgical strategies that may decrease postoperative neurological dysfunction.

Methods
Patient Characteristics. The institutional review board of Xijing Hospital approved this retrospective study and waived the need for informed consent. This study was conducted in accordance with the Declaration of Helsinki. A computerized CT database was retrospectively searched from January 2013 through June 2016; the key words "type A aortic dissection" and "surgical repair" were used. There were 443 patients in the initial database. ATAAD was defined as any nontraumatic dissection of the aorta proximal to the left subclavian artery   presenting within 14 days of symptom onset 22 . A total of 188 patients were excluded for the following reasons: non-acute type A aortic dissection (n = 6); a history of previous aortic operation (n = 32); unavailable preoperative aortic CTA data from other medical institutions (n = 114); CT image quality that could not be evaluated (n = 11); a time interval between admission and emergency surgery repair of more than 24 hours (n = 5); and postoperative mortality (n = 20). In total, 255 patients with ATAAD were finally eligible for the present study. The standard thoracic aortic operations were performed primarily by 2 dedicated principal surgeons, and the procedures were standardized. Neurological Evaluation. The in-hospital postoperative neurological symptoms and signs of patients were diagnosed after a detailed clinical assessment by neurologists and further neuroimaging. Patients were considered to have PND if they exhibited onset of focal (stroke) or global (coma) deficits or were found to have a focal infarction confirmed by CT scanning or magnetic resonance imaging of the brain [5][6][7] . TND was defined as a symptom complex involving postoperative confusion, seizures, agitation, or transient delirium with no structural abnormalities in the brain that were detectable via the usual imaging methods, and resolution of the symptoms generally occurred before hospital discharge [5][6][7] . The patients were assigned to the TND, PND, or NND group.

Image Acquisition.
CTA Findings, Definition and Parameters. CCA, ICA and VA were included the analysis, which are directly correlated with the cerebral artery, which supplies cerebral blood, and indirectly correlated with the thoracic aorta. The diameter of the involved aAO was defines as the maximum diameter of the aAO in the section orthogonal to the centreline. A retrograde aAO dissection is a primary intimal tear in the aortic arch or dAO in which the intimal flap extends to the aAO in a retrograde manner without re-entry. The diameter of the entry point was recorded as the maximum diameter in axial or oblique sagittal images. Low enhancement in the unilateral carotid artery was defined as a CT value for the unilateral artery in a slice of the C1 vertebra that was lower than that for the offside artery for more than 100 Hu.
Detailed CTA Characteristics of CCA Dissection. Due to the complexity of the impaired status, the location, extent and detailed characteristics of the CCA dissection were further analysed. The detailed characteristics of CCA dissection include (1) re-entry and lower enhancement in the false lumen; (2) re-entry and the same enhancement between the true and false lumen; (3) no re-entry but a severely narrowed true lumen; and (4) no re-entry or a severely narrowed true lumen. A severely narrowed (> 75%) true lumen was defined as a minimum diameter of the true lumen that was less than 25% of the CCA diameter in the same section 20 .
Statistical analyses. Statistical analyses were performed using SPSS software (version 17.0; SPSS, Inc, Chicago, IL, USA). Categorical variables were presented as the number and percentage, and continuous variables were described as the mean ± standard deviation. Statistical significance was determined with the χ 2 test for categorical variables and one-way analysis of variance for continuous variables. A P value < 0.05 was considered significant. Cohen's Kappa statistics were calculated for interreader agreement, for the assessment of image quality. Univariate analysis was applied to discriminate CTA risk factors for postoperative PND and TND. Significant CTA risk factors were further analysed in a multivariate logistic regression model to identify isolated CTA risk factors for TND and PND. All P values and 95% confidence intervals (CIs) were two-tailed.