Radiological properties of neurological injury following acute type A aortic dissection repair

Objective The study objective was to assess the radiological properties of acute type A aortic dissection–related neurological injuries and identify predictors of neurological injury. Methods Our single-center, retrospective, observational study included all patients who underwent acute type A aortic dissection repair between January 1998 and December 2021. Multivariable analyses and Cox regression were performed to identify predictors of embolic lesions, watershed lesions, neurological injury, 30-day mortality, and late mortality. Results A total of 538 patients were included. Of these, 120 patients (22.3%) experienced postoperative neurological injury; 74 patients (13.8%) had postoperative stroke, and 36 patients (6.8%) had postoperative coma. The 30-day mortality was 22.7% in the neurological injury group versus 5.8% in the no neurological injury group (P < .001). We identified several independent predictors of neurological injury. Cerebral malperfusion (odds ratio, 2.77; 95% confidence interval, 1.53-5.00), systemic hypotensive shock (odds ratio, 1.97; 95% confidence interval, 1.13-3.43), and aortic arch replacement (odds ratio, 3.08; 95% confidence interval, 1.17-8.08) predicted embolic lesions. Diabetes mellitus (odds ratio, 5.35; 95% confidence interval, 1.85-15.42), previous cardiac surgery (odds ratio, 8.62; 95% confidence interval, 1.47-50.43), duration of hypothermic circulatory arrest (odds ratio, 1.05; 95% confidence interval, 1.01-1.08), cardiopulmonary bypass time (odds ratio, 1.01; 95% confidence interval, 1.00-1.01), ascending aortic/arch cannulation (odds ratio, 5.68; 95% confidence interval, 1.88-17.12), and left ventricular cannulation (odds ratio, 17.81; 95% confidence interval, 1.69-188.01) predicted watershed lesions. Retrograde cerebral perfusion (odds ratio, 0.28; 95% confidence interval, 0.01-0.84) had a protective effect against watershed lesions. Conclusions In this study, we demonstrated that the radiological features of neurological injury may be as important as clinical characteristics in understanding the pathophysiology and causality behind neurological injury related to acute type A aortic dissection repair.

The radiological features of neurological injuries may be as important as clinical manifestations in understanding the pathophysiology behind neurological injury related to ATAAD repair.

PERSPECTIVE
Although neurological injuries after ATAAD repair are well studied, there are limited data on the radiological properties of these injuries.By assessment of available data, this study emphasized that radiological information may provide a better understanding of the mechanisms responsible for neurological injury after ATAAD repair.
5][6][7] The pathophysiological mechanisms of preoperative neurological injury include dissection of the supra-aortic blood vessels; embolization; or systemic hypotension due to shock, hypovolemia, severe aortic regurgitation, or cardiac tamponade. 8fter ATAAD repair, 10% to 16% of patients experience postoperative stroke and 3% to 9% experience coma. 1,2,4,7,9,10Patients with postoperative neurological injury have short-and mid-term mortality rates twice those of patients with no signs of cerebral injury. 1 Previously, stroke after ATAAD repair with deep hypothermic circulatory arrest was thought to be caused by hypoperfusion during the arrest, where time was the most important factor.However, recent data suggest that postoperative neurological injury could be the result of cerebral malperfusion, hypoperfusion, or embolism occurring at any time during the perioperative period. 8urgical factors, including the arterial cannulation strategy, choice of cerebral perfusion during deep hypothermic circulatory arrest, extent of distal repair, and risk of residual intimal flaps in the downstream aorta or branch vessels are all associated with the risk of postoperative neurological injury. 1,3,6,9revious studies on neurological injury after ATAAD repair have mainly relied on the clinical manifestation of neurological injury.Only one study detailed the radiological features of cerebral injuries associated with ATAAD repair, and it was limited by a small study population. 11The aim of this study was to assess the radiological properties of ATAAD-related neurological injuries and to identify predictors of these neurological injuries based on their radiological profile rather than their clinical manifestations.

MATERIALS AND METHODS Study Design
This study was a single-center, retrospective, observational study and included all patients who underwent ATAAD surgery between January 1998 and December 2021 at our institution.Data were prospectively entered into our departmental surgical database, and additional data were c-ollected by retrospective chart review and radiological assessment.The study was approved by the Ethical Review Authority, Stockholm, Sweden (ref.2021-01185, April 23, 2021).Individual patient consent was waived.

Outcomes and Definitions
The primary outcome measures for this study were clinical neurological injury, embolic cerebral lesions, and watershed cerebral lesions.Clinical neurological injury was identified at wake up and was defined as focal neurological deficit or coma diagnosed by clinical assessment with or without radiological confirmation (computed tomography or magnetic resonance imaging [MRI]) with a symptom duration of more than 24 hours.Embolic lesions were defined as ischemia spread in the brain parenchyma, and watershed infarctions were defined as ischemic lesions occurring at the border zones between major cerebral artery territories resulting from cerebral hypoperfusion.Postoperative coma was defined as Glasgow Coma Scale motor score less than 6, 48 hours after termination of pharmacological sedation.Patients who were comatose postoperatively and regained consciousness with persisting stroke symptoms were regarded as coma patients.The radiological assessment was performed by a senior consultant neuroradiologist.Postoperative stroke was defined as focal neurological symptoms with a duration of more than 24 hours confirmed by a neurologist or with radiological findings correlating with the symptoms.If a neurological injury was presumed, a neurologist would be consulted.When classification was unclear, a neurologist reviewed the documentation of neurological symptoms.Hypotensive shock was defined as systolic blood pressure less than 90 mm Hg, clinical signs of hypotension, or the preoperative use of vasopressors or inotropes.Malperfusion was defined as clinical signs of end-organ ischemia, and cerebral malperfusion was defined as focal neurological deficit or altered state of consciousness before surgery.

Surgical Technique
The surgical technique used at our institution has been described in detail. 12In summary, repair was performed using median sternotomy, cardiopulmonary bypass, and intermittent cardioplegic arrest.Cannulation site varied at the discretion of each surgeon.In a minority of cases, aortic crossclamp was used, but in general, the distal anastomosis, resection, and inspection of the aortic arch were performed under deep or moderate hypothermic circulatory arrest, with or without the use of antegrade cerebral perfusion (ACP) or retrograde cerebral perfusion (RCP).The decision to use RCP was based on surgeon preference, and there were no patient or clinical triggers.The use of ACP was limited to arch procedures or when the anticipated circulatory arrest exceeded 30 minutes.If a limited distal repair (ascending aortic arch or hemiarch repair) was feasible, this approach was favored at our institution, but the technique used depended on the location of the intimal tear and the extent of dissection.Aortic arch procedures entailed reimplantation of any supra-aortic branch.Aortic valve replacement or total root replacement was performed when the dissection involved the coronary ostia or aortic valve, or in the presence of an aortic root aneurysm.When required, the competence of the aortic valve was restored via subcommissural plication, commissural resuspension, or valvuloplasty.Concomitant procedures (eg, coronary artery bypass) were performed when required.

Statistical Analysis
Categorical variables were presented as numbers and percentages.Continuous data were reported as median with interquartile range or mean AE standard deviation depending on the distribution of data.Chi-square test, Fisher exact test, 2-sample t tests, and Mann-Whitney U test were used for intergroup comparisons when appropriate.Patients who died intraoperatively were excluded from group analysis.Logistic regression was used for univariable and multivariable analyses to identify predictors of neurological injury and 30-day mortality.A P value of less Two variables with a variance inflation factor greater than 3.0 or correlation coefficient of greater than 0.5 were defined as being multicollinear.Multicollinearity was identified between any malperfusion and cerebral malperfusion and between clinical neurological injury and radiological neurological injury.Any malperfusion was our preferred variable when assessing 30-day mortality, whereas cerebral malperfusion was used for the analysis of neurological and radiological outcomes.In models where neurological injury was tested, neurological injury was used in the baseline model, but radiological and clinical properties were analyzed in separate models to generate the specific odds ratios (ORs) for these variables.
To reduce the number of variables in the multivariable analysis, we considered reoperation for bleeding as a marker of postoperative bleeding complications, postoperative renal replacement therapy as an indicator of renal injury, and hypotensive shock as an indicator of circulatory instability.Because the model relied on complete case analysis, lactate, postoperative myocardial infarction, fibrinogen, and recombinant factor VIIa were excluded due to numerous missing values.The analyses were performed using the Backward Wald method, and the variables remaining at the last step were reanalyzed using the Enter method to generate the final ORs.Goodness-of-fit of the multivariable model was evaluated using the Omnibus Tests of Model Coefficients and the Hosmer-Lemeshow test.Predictors of late mortality were assessed with Cox proportional hazard regression using stepwise backward elimination, including only 30-day survivors.The results of the logistic regression analyses are expressed as ORs and those of the Cox regression analysis as hazard ratios (HRs), both with 95% confidence intervals (CIs).ORs and HRs were illustrated using forest plots.Late survivals AE1 standard error were illustrated using the Kaplan-Meier method.All statistical analyses were conducted using IBM SPSS Statistics version 28.0 (IBM Corp).

Study Population
Between January 1998 and December 2021, 538 patients underwent ATAAD surgery at our institution, all of whom were included in the present study.Follow-up was conducted in January 2022 with a total of 3346 patient-years (median 5.1 [1.2-10.1],mean 6.2 AE 5.7).A total of 23 patients (4.3%) died intraoperatively.A total of 120 patients (22.3%) experienced postoperative neurological injury, of whom 74 patients (13.8%) had stroke and 36 patients (6.8%) had coma.A total of 15 patients (2.8%) were initially comatose and showed focal neurological symptoms after regaining consciousness.Ten patients (1.9%) had an unspecified neurological injury (Figure E1).A total of 78 patients (14.5%) presented with cerebral malperfusion.

Radiological Properties of Neurological Injury
Both hemispheres were affected in more than half of patients (53.4%), but unilateral embolic injuries were more common in the right hemisphere than the left in patients with stroke and coma (70.4% vs 85.7%, P ¼ .442).In patients with neurological injury, common carotid dissection was present on the left side in 6 patients (5.6%), right side in 20 patients (18.5%), and bilateral in 36 patients (33.3%).The middle cerebral artery was the most affected vascular territory in both groups (71.6% and 69.4%, respectively), but embolic lesions in the basilar artery territory were significantly more common in comatose patients (47.2% vs 20.3%, P ¼ .009)as were watershed lesions (41.7% vs 6.8%, P <.001) and mixed lesions (30.6% vs 4.1%, P <.001) (Table 3).

DISCUSSION
In this study, we showed that neurological injuries associated with ATAAD and subsequent surgical repair significantly increased postoperative mortality and morbidity.We also demonstrated that embolic injuries were more often observed in the right hemisphere, with the most common vascular territory being the middle cerebral artery.Embolic lesions in the basilar artery territory were more commonly associated with a postoperative comatose state, as were watershed lesions.As could be expected, preoperative cerebral malperfusion was associated with increased risk of postoperative neurological injury, but we could also show that RCP had a protective effect.Significant associations of embolic and watershed lesions are presented in Figure 4 and the Video Abstract.
Numerous factors have been reported to cause neurological injury after ATAAD surgery. 2 Some factors are a consequence of the pathophysiological mechanisms of ATAAD, including arch vessel branch dissection and compromised cerebral blood flow, potential thromboembolism from clot formation in the exposed false lumen, systemic hypotension due to exsanguination, cardiac tamponade, or cardiogenic shock due to valve or coronary artery involvement.[3]9 However, previous studies have relied on the clinical manifestation of neurological injury regardless of radiological appearance.We propose that assessing neurological injury based on its radiological properties rather than its clinical presentation may be more feasible for determining causality between preoperative and intraoperative factors and cerebral injury after ATAAD repair.
In our study, 22.3% of patients had postoperative neurological injury, 13.8% of whom had stroke and 6.8% were rendered comatose, regardless of preoperative state.The largest study included 9000 patients from the Society of Thoracic Surgeons Adult Cardiac Surgery Database reported a postoperative neurological complication rate of 13%.However, the proportion of patients with preoperative cerebral malperfusion was not presented, and patients who were in an unresponsive state at any time during the 24 hours preceding the procedure were excluded from the study. 9ecently, The Nordic Consortium for Acute Type A Aortic Dissection (NORCAAD) registry reported a stroke rate of 16%, and in the German Registry for Acute Aortic Dissection Type A (GERAADA) registry, 17% of patients had postoperative stroke. 1,6Most studies on ATAAD complications are self-reported by surgeons with the stroke definition including both clinical and radiological signs of neurological injury.This may lead to an underestimation of the true incidence of postoperative stroke because patients who die before radiological examinations are not included.Furthermore, a clinical examination performed by neurologists is more likely to detect minor neurological deficits.Preoperative cerebral malperfusion was present in 14.5% of patients in this study, 8.5% of patients in the NORCAAD registry, and 20.5% of patients in the GERAADA database.In the present study, 30-day mortality was 13.7%, compared with 15.7% and 16.9% for NORCAAD and GERAADA, respectively. 1,6In line with previous findings, we demonstrated that cerebral malperfusion was an independent predictor of neurological injury.However, our study showed that less than half of patients with cerebral malperfusion had persistent neurological injuries after surgery. 1,4This indicates that, in a significant proportion of patients, cerebral malperfusion may be resolved by ATAAD repair.
Diabetes mellitus is associated with an increased risk of stroke in general with several proposed mechanisms, including vascular endothelial dysfunction and increased stiffness of the cerebral arteries. 13In our study, we showed that diabetes mellitus increases the risk of watershed injuries.It is reasonable to believe that patients with diabetes mellitus are more sensitive to perioperative systemic hypotension due to vasculopathy caused by the disease.Presumably, this could render diabetic patients more sensitive to shorter periods of circulatory arrest, both because of the general metabolic consequences of reduced blood flow and hypoperfusion leads to inadequate cooling of the cerebral parenchyma.Previously, femoral cannulation has been demonstrated to increase the risk of perioperative stroke, and backwards flushing of embolic material from the dissected descending aorta has been suggested as an explanation. 9,14However, no randomized trials have been performed on cannulation strategies, and the results from the previous studies may be influenced by significant selection bias.The increased risk of stroke in patients with femoral cannulation may be related to surgical inexperience or hemodynamic instability. 14Axillary cannulation is well established and provides excellent results. 9However, the downside of this strategy is the time required to cannulate.In patients with unilateral postoperative neurological injury, we observed a higher frequency of embolic lesions in the right hemisphere compared with the left (73.6% vs 26.4%).The proposed mechanism of backward washing of embolic material using femoral cannulation most likely would lead to an increase of lesions in the left hemisphere.However, a post hoc analysis of patients with unilateral lesions, patients who had a femoral cannulation had 78.8% of lesions in the right hemisphere compared with 55.6% when other cannulation sites were used (P ¼ .21).
Two previous studies contrast our findings.Fichadiya and colleagues 11 did not identify any lateralization of unilateral strokes in patients with ATAAD, and a study of patients with cardioembolic stroke due to atrial fibrillation found that the lesions are equally distributed between the hemispheres, with a slightly higher proportion located in the left hemisphere. 15,16Embolic lesions in the right hemisphere might indicate that embolic material originates from the heart or ascending aorta.Because we only perform computed tomography or MRI on clinical suspicion of neurological injury, left-sided symptoms may be more easily recognized clinically than right-sided symptoms, which might increase the lateralization frequency in favor of the left hemisphere. 15n the absence of solid embolic material during ATAAD surgery and the higher occurrence of right-sided lesions, one may speculate that the lesions are caused by air trapped in the supra-aortic branch vessels during circulatory arrest.Although this is merely a speculation, this hypothesis may be supported by our finding that RCP has a protective effect on neurological injury.This is consistent with previous research showing that RCP during circulatory arrest reduces mortality and neurological injury when compared with straight circulatory arrest in both elective and emergency aortic surgery. 17,180][21]  Bonser and colleagues 19 showed that when using ice packs as topical cooling (which is routinely used at our institution), the addition of RCP did not provide any change in nasopharyngeal temperature, suggesting that the cooling effect of RCP is limited.Therefore, the most likely protective effect of RCP is generated by the backward flow of blood, washing out air and other potential embolic material from the supraaortic circulation.
The higher occurrence of right-sided lesions may be explained by our finding that the right common carotid artery was dissected more often than the left.This may contribute to the discrepancy between our study and the study by Fichadiya and colleagues, 11 where no side difference was observed.Because the primary entry most often is located in the ascending aorta, it is logical that the dissection more often involves the first vessel taking off from the aortic arch. 6here is some evidence that in patients with cerebral malperfusion and occlusion of the carotid artery, cannulation with extra-anatomic bypass of the carotid artery might improve neurological outcomes. 22,23With the reservation that we did not distinguish between carotid artery dissection and occlusion, our findings that carotid artery dissection does not predict neurological injury and that most patients with cerebral malperfusion do not develop postoperative neurological injury suggest that carotid artery bypass should be considered only in highly selected cases. 1

Study Limitations
This study is limited by its retrospective design and by a database missing potentially significant variables.Because of the urgency of the diagnosis, preoperative brain imaging was not performed to confirm neurological injuries present before surgery.Not all cases of neurological injury could be radiologically confirmed, probably because of the low number of MRI examinations in this study.Furthermore, radiological embolic lesions could be the result of intimal flaps or local atherosclerotic plaques causing regional hypoperfusion instead of embolic.The primary assessment was not always performed using a standardized approach (ie, Glasgow Coma Scale or National Institutes of Health Stroke Scale).However, this study is strengthened by the assessment of all radiological examinations by a neuroradiologist and by the retrospective confirmation of neurological injury by a neurologist in cases where classification was unclear.Finally, because ACP was reserved for more complex procedures, our data regarding ACP are not scientifically reliable.

Radiological properties of neurological injury following acute type A aortic dissection repair
The radiological features of neurological injuries may be as important as clinical manifestations in understanding the pathophysiology behind neurological injury related to ATAAD repair.

CONCLUSIONS
This study showed that the radiological features may be as important as clinical manifestations in understanding the pathophysiology behind neurological injury related to ATAAD repair.Although embolic lesions more often occurred in the right hemisphere causing symptoms of stroke, basilary artery obstruction and watershed lesions were associated with postoperative coma.We also showed that cerebral malperfusion and diabetes mellitus play key parts in the development of neurological injury and that RCP has a neuroprotective effect.

FIGURE 4 .
FIGURE 4. Significant predictors of embolic lesions and watershed lesions in patients undergoing surgical repair of ATAAD.Patients with coma and symptoms of focal neurological injuries (stroke) were categorized into the coma group.CPB, Cardiopulmonary bypass; Ref, reference variable; ATAAD, acute type A aortic dissection.

FIGURE E3 .
FIGURE E3.Kaplan-Meier survival curves of 30-day survivors demonstrating poorer survival in patients with neurological injury.CI, Confidence interval.
.1 in the univariable model was required for a variable to be included in the full model.Multicollinearity between continuous variables was assessed by linear regression using variance inflation factor.Spearman correlation was used for testing multicollinearity between categorical values and Pearson correlation between categorical and continuous values. than

TABLE 1 .
Baseline characteristics Values are presented as mean AE standard deviation, n (%), or median (interquartile range).COPD, Chronic obstructive pulmonary disease.*Patients who died intraoperatively were excluded from the group "No neurological injury" (n ¼ 23).

TABLE 3 .
Radiological properties of neurological injury stratified by clinical presentation Values are presented as n (%).CT, Computed tomography; MRI, magnetic resonance imaging.*Patients with coma and symptoms of focal neurological injuries (stroke) were categorized into the coma group.

Diabetes mellitus Syncope Hypotensive shock Cerebral malperfusion Carotid dissection (Ref-None) Unilateral Bilateral CPB time per minute increment Retrograde cerebral perfusion vs all other Distal technique (Ref-Ascending)
However, Forest plot illustrating the association with neurological injury.Values are expressed as OR with 95% CI.CPB, Cardiopulmonary bypass; Ref, reference variable.

Diabetes mellitus Syncope Hypotensive shock Cerebral malperfusion Carotid dissection (Ref-None) Unilateral Bilateral Retrograde cerebral perfusion vs all other Distal technique (Ref-Ascending) Hypothermic circulatory arrest technique (Ref-Cross clamp)
Forest plot illustrating the association with embolic lesions.Values are expressed as OR with 95% CI.Ref, Reference variable.

Diabetes mellitus Previous cardiac surgery Cerebral malperfusion CPB time per minute increment Hypothermic circulatory arrest technique (Ref-Straight circulatory arrest) Hypothermic circulatory arrest, time per minute increment
FIGURE 3. Forest plot illustrating the association with watershed lesions.Values are expressed as OR with 95% CI.CPB, Cardiopulmonary bypass; Ref, reference variable.Values are presented as mean AE standard deviation, n (%), or median (interquartile range).rFVIIa, Recombinant factor VIIa; PRBC, packed red blood cells; MI, myocardial infarction; CKMB, creatine phosphokinase-MB; NA, not available.*Patients who died intraoperatively were excluded from the group "No neurological injury" (n ¼ 23).

lesions Hypothermic circulatory arrest technique (Ref-Straight circulatory arrest) Hypothermic circulatory arrest time per minute increment
FIGURE E4.Forest plot illustrating independent predictors of late mortality in patients who survived to more than 30 days after surgery.Values are expressed as HR with 95% CI.COPD, Chronic obstructive pulmonary disease; Ref, reference variable.

TABLE E1 .
Univariable and multivariable logistic regression analysis for predictors of neurological injury

TABLE E2 .
Univariable and multivariable logistic regression analysis for predictors of embolic neurological injury

TABLE E3 .
Univariable and multivariable logistic regression analysis for predictors of watershed neurological injury

TABLE E4 .
Univariable and multivariable logistic regression analysis for predictors of 30-day mortality

TABLE E5 .
Univariable and multivariable Cox regression analysis for predictors of late mortality