1. Introduction
Acute type A aortic dissection (ATAAD) is a lethal cardiac disease involving the aorta, with catastrophic effects on end-organ perfusion. The overall mortality rate is 5.8% at 48 h, with a decrease to 4.4% at 48 h in patients with emergency surgery [
1,
2].
Despite advances in surgical techniques (central aortic therapy) and the introduction of brain protection methods like cerebral perfusion and hypothermic circulatory arrest, the incidence of neurological complications (NCs) is high and ranges between 17 and 48% [
3,
4,
5]. NCs after the emergency repair of an ATAAD are associated with higher in-hospital mortality, longer intensive care and hospital stays, and reduced long-term survival [
6,
7].
Ischemic stroke is the most clinically important NC, and we reported a 24.8% postoperative incidence in a previous study [
8]. Spinal cord ischemia (SCI), hypoxic-ischemic encephalopathy (HE), and postoperative delirium (PD) are severe complications, but there are scarce data regarding their postoperative incidence after emergency surgery for ATAAD. In a retrospective study on 278 patients with type A and B aortic dissection, there was a 3.2% preoperative incidence of HE [
9]. A systematic meta-synthesis literature review reported 67 cases of preoperative SCI caused by aortic dissection (type A and type B), while PD was present in 12–37% of patients with ATAAD [
10,
11].
Data regarding pre/postoperative NCs are heterogeneous, and there is wide variability in the categorization of NCs [
11]. For example, in the reports by the International Registry of Aortic Dissection (IRAD), 17.1% of patients presented with “any focal neurological deficits” and 15% with “coma/altered consciousness” at admission. They reported that 29 and 8.5% of those patients presented these NCs during hospitalization, without information about surgical procedures [
12]. The German Registry for Acute Aortic Dissection Type A (GERAADA) investigated hemiparesis/plegia, paraparesis/plegia, aphasia, and unconsciousness/coma as NCs [
2]. Stroke is the only NC evaluated in the reports by The Nordic Consortium for Acute Type A Aortic Dissection (NORCAAD) and The Society of Thoracic Surgeons (STS) [
6,
13].
While preoperative risk factors for NCs in ATAAD can be considered endogenous, are patient related, and cannot be changed, intraoperative risk factors depend in great measure on the surgical team and strategy.
Despite the relative association between NCs and ATTAD, risk factors for newly developed postoperative NCs, especially intraoperative factors, remain uncertain (
Table 1). For example, operative time was associated with new onset of NCs in the multi-center study GERAADA [
13]. The Nordic Consortium for Acute Type A Aortic Dissection (NORCAAD) reports only total arch replacement as a risk factor for NCs [
13], while The Society of Thoracic Surgeons (STS) reports that femoral cannulation increases the risk of stroke [
6].
The aim of this study was to determine the association between different intraoperative parameters and newly developed postoperative NCs in patients after emergency surgery for ATAAD.
2. Materials and Methods
Between January 2017 and May 2023, 240 patients were transferred to our center for ATTAD management. The diagnosis of ATAAD was based on chest computer tomography (CT) with intravenous contrast. After a cardiology consultation in the emergency department with mandatory transthoracic echocardiography, patients who were candidates for emergency surgery were transferred into the operating room. Clinical characteristics and demographic data were collected from medical records and the electronic health system.
Inclusion criteria: patients with acute type A aortic dissection according the to Stanford classification were considered for emergency surgery.
Exclusion criteria: (1) patients with ischemic or hemorrhagic stroke documented on computed tomography (CT) scans prior to surgery; (2) clinical signs of stroke, paraplegia, delirium or any neurological dysfunction prior to surgery; (3) patients with history of transient ischemic stroke, ischemic or hemorrhagic stroke or any other neurological dysfunction; (4) patients who died in the operating room or within the first 48 h after intensive care admission; (5) patients in the postoperative setting for whom the neurological status could not be evaluated; (6) patients with incomplete medical records.
Immediate newly developed NCs were evaluated in the first 72 h of ICU stay.
Ischemic stroke was confirmed by head CT when clinical suspicion was raised in the postoperative period and after initial neurological examination. The Modified Rankin Scale (mRS) was used to quantify the degree of disability at discharge based on clinical charts and neurological examination.
Anterior spinal cord ischemia was diagnosed after a physical examination for paraplegia in the postoperative period. Spine MRI was used for confirmation, with T2-weighted image hyperintensity in the anterior horns being the hallmark finding.
Hypoxic-ischemic encephalopathy was diagnosed on head CT after neurological examination. CT findings were bilateral basal ganglia hypodensities, diffuse mass effect with effacement of the cerebral sulci, loss of the normal grey-white matter differentiation due to edema, and bilaterally decreased density in a watershed distribution.
Postoperative delirium was diagnosed by the Confusion Assessment Model for the Intensive Care Unit (CAM-ICU) criteria after all other causes for neurological dysfunction were excluded by the neurologist.
2.1. Surgical Technique
Median sternotomy was the standard approach in all ATAAD cases after the induction of general anesthesia. The exposure of the axillary or femoral artery for arterial cannulation and a two-stage venous cannula in the right atrium would be considered for the aortic procedure only. Should an additional procedure such as mitral repair or replacement be required, then bicaval cannulation would be the obvious choice. Should any of the alternative arterial sites be considered unsuitable, then the direct cannulation of the dissected ascending aorta would be considered using a Seldinger technique under transesophageal echocardiographic guidance to identify the true aortic lumen. Anterograde and retrograde cold (4 °C) crystalloid cardioplegia was used for myocardial protection. Moderate hypothermia (25–28 °C) was considered when circulatory arrest was required. An open-distal anastomosis technique was preferred, using selective anterograde cerebral perfusion through the direct cannulation of the innominate and left common carotid artery. Near-infrared spectroscopy was used routinely for cerebral oxygenation monitoring. When the primary entry tear could not be addressed with a hemiarch approach, a total aortic arch replacement would be considered.
Figure 1,
Figure 2 and
Figure 3 show the three main techniques used. The bilateral selective anterograde cerebral perfusion protocol is as follows: when the target temperature was achieved and circulatory arrest was initiated, the arch was opened, and two balloon-tipped cannulas were inserted under direct vision into the innominate artery and left common carotid artery. We used 13F or 15F cannulas depending on the vessel’s diameter. Each cannula was connected to a different pump with a separate pressure monitoring line. For the left side, the perfusion pressure was 60–70 mmHg with a flow of 180–230 mL/min. The right-side perfusion pressure was 60–70 mmHg with a flow between 150 and 200 mL/min. Perfusion parameters were adapted within the mentioned intervals to reach a target value of 50–70% on NIRS. Before completing the distal anastomosis, the cannulas were extracted one at a time, and the innominate and left common carotid arteries were snared until the completion of the anastomosis. After open-distal anastomosis was performed, an arterial cannula was inserted directly into the Dacron prosthesis, CPB was restarted, and the supra-aortic vessels were unclamped after de-airing. During the rewarming of the patient, the repair of the ascending aorta and or aortic root was performed. After weaning from CPB, hemostasis was achieved, epicardial leads were placed, and sternal osteosynthesis was achieved with steel wires.
2.2. Statistical Analysis
Statistical analysis was conducted with Wizard 2 Statistical Software for Mac OS (Wizard–Statistics and Analysis®, Raipur, Chattisgarh, India).
Summary statistics are presented as absolute numbers and percentages for categorical values and as mean ± standard deviation for continuous values. Our primary outcome was calculating the incidence of newly developed neurological complications in patients after emergency surgery for ATAAD. In order to investigate the association between different intraoperative factors and the development of new NCs, multivariable analysis was performed using logistic regression and taking into account a model that included variables achieving a p-value < 0.1 in univariate analysis. A predictive modeling strategy with the backward stepwise method of entering data was then used. Variables included in the univariate analysis were the following: age, male sex, arterial hypertension, diabetes, hyperlipidemia, chorionic kidney disease, preoperative atrial fibrillation, cardiac tamponade at admission, severe aortic regurgitation, severe left ventricle dysfunction, presence of bicuspid aortic valve, a primary intimal tear in the ascending aorta, aortic arch, ascending aorta and aortic arch or no entry tear found in the ascending aorta or aortic arch (based on the preoperative chest CT), the dissection of one, two or three supra-aortic vessels, and reintervention for mediastinal bleeding. Intraoperative indexes were type of surgery (aortic arch/no aortic arch surgery, Wheat procedure, aortic root surgery, supra-coronary ascending aorta replacement, hemiarch replacement, total arch replacement, innominate artery implantation, innominate artery and left common carotid artery implantation, combined procedures), axillary or femoral artery cannulation, cardiopulmonary bypass time, aortic cross-clamp time, and cerebral perfusion time. The logistic regression results are presented as odds ratios (ORs) with confidence limits and p-values. p < 0.05 was considered statistically significant. An independent t-test or Chi-square test was used for comparing the neurologically complicated and uncomplicated groups.
4. Discussion
Acute type A aortic dissection (ATAAD) represents a catastrophic event with high mortality (5.8% at 48 h) and morbidity in the absence of emergency surgery [
1,
2,
9].
Despite advances in surgical techniques and brain protection methods, neurological complications in this group of patients rage from 17 to 48% [
3,
4,
5]. In our study, the incidence of newly developed postoperative NCs was 39.4%. The high variability in the reported incidence of NCs can be due to difficulties in recording them in critically ill patients [
4]. Data regarding pre- and postoperative NCs are heterogeneous and there is wide variability in the categorization of NCs [
2,
13,
14].
Ischemic stroke is the most clinically important newly developed perioperative NC, with a hospital mortality rate six times greater than patients without stroke [
14,
15,
16]. The incidence of ischemic stroke in this study was 23.6%, being the most frequent NC (60%), in accordance with previous reports [
3,
4,
5]. Ischemic stroke tends to be more frequent in the anterior circulation (47.91%) than the posterior one (20.83%) and is predominantly right-sided in both cases (25% right carotid artery anterior, 8.33% right vertebrobasilar artery). Supra-aortic vessel involvement was observed in 39.6% of patients with ischemic stroke, and the frequency of two-vessel involvement was significantly higher in the NC group (16.7% vs. 6.5%;
p = 0.03). One reason for the right-side dominance of lesions can be that hydraulic stress is greatest in the right lateral wall of the ascending aorta [
17]. However, 60.4% of patients with ischemic stroke did not have supra-aortic vessel involvement, suggesting there are other mechanisms present in the pathogenesis of ischemic stroke. Thromboembolism, microembolism, or hypotension may be such factors. Patients with ischemic stroke were significantly older (61.12 ± 9.27 vs. 56.52 ± 12.9;
p = 0.011), which could be explained by the aggravation of atherosclerosis with the increase in age [
18]. Sixteen patients (33.34%) with ischemic stroke died in the intensive care unit, and 43.75% had a mRs between 3 and 5, confirming the high mortality and morbidity of this postoperative complication.
Hypoxic-ischemic encephalopathy is a severe complication of global ischemia, with a spectrum of disability that ranges from complete recovery from coma to death, while data on long term outcome are rare [
19,
20,
21]. The incidence of hypoxic-ischemic encephalopathy in ATAAD is also rare. A retrospective study on 278 patients with type A and B aortic dissection found a preoperative incidence of HE of 3.2% [
9]. Blanco et al. reported 5 patients with preoperative hypoxic-ischemic encephalopathy in 24 patients with ATAAD [
4]. In our study, the incidence was 4.4% (nine patients). Systolic hypotension related to cardiac tamponade was observed in four patients (44.4%) and related to acute severe aortic regurgitation in two patients. Two patients exhibited the involvement of the supra-aortic vessels, both of them with dissection of the innominate artery and left common carotid artery, without the involvement of the subclavian arteries. To date, surgery is the best approach to protect against definitive neurological deficits when cerebral malperfusion is suspected [
22]. Axillary cannulation and bilateral anterograde cerebral perfusion should always be considered in these cases [
22]. In our study, six out of nine patients with hypoxic-ischemic encephalopathy had axillary artery cannulation and bilateral anterograde cerebral perfusion. One patient did not regain consciousness and died of septic shock.
Spinal cord ischemia in ATAAD is also a rare condition, and little data are available on its incidence [
23]. Spinal cord ischemia presents as anterior spinal cord syndrome in 87.2% of cases, while SCI in the territory of the posterior spinal artery is very rare [
24]. There are a few data regarding SCI in ATTAD. One study reports that 1% of patients presenting with ATAAD have SCI [
25] A systematic meta-synthesis literature review found 67 cases of SCI caused by aortic dissection (type A and type B). Out of 19 patients having undergone emergency surgery, full recovery in 13 patients, residual paraplegia in 4 patients, and death in 2 patients was observed. The overall mortality was 57.14% in patients presenting with SCI in the setting of ATAAD [
23]. Two patients developed SCI in our study and both of them were discharged, one without recovery and the other with a partial recovery from paraplegia.
Postoperative delirium is a severe brain disorder characterized by inattention, disturbed thinking and altered levels of consciousness [
25,
26]. A meta-analysis of 12 studies from 2016 to 2022 found that low oxygen levels, prolonged mechanical ventilation, renal dysfunction, low hemoglobin levels and prolonged intensive care stay are risk factors for PD in ATAAD [
21]. Incidence in cardiac surgery is high (32.5 and 52%) and involves between 12 and 37% of patients with ATAAD [
10,
11]. In our study, 21 (10.3%) patients developed this NC. Only one patient died in our study (mixed cardiogenic and septic shock).
Despite the relative association between NCs and ATTAD, risk factors for newly developed postoperative NCs, especially intraoperative factors, remain uncertain. A multi-center study (GERAADA) with 2137 patients found that operative time was associated with new onset of NCs [
13]. Another study enrolled 501 patients with aortic arch surgery, moderate hypothermic circulatory arrest, and selective anterograde cerebral perfusion. They found that permanent NCs were associated with operation time and temporary NCs were associated with the duration of circulatory arrest [
27]. The Nordic Consortium for Acute Type A Aortic Dissection reports only total arch replacement as a risk factor for NCs [
13], while the Society of Thoracic Surgeons reports that femoral cannulation increases the risk of stroke [
6].
While preoperative risk factors for NCs in ATAAD can be considered endogenous, are patient-related, and cannot be changed, intraoperative risk factors depend in great measure on the surgical team and strategy.
Preoperative risk factors associated in univariate analysis with newly developed postoperative NCs in this study were age (OR = 1.02; 95% CI = 0.1–1.05;
p = 0.025), aortic arch tear (OR = 2.36, 95% CI = 1.29–4.35;
p = 0.015) and ascending aorta and aortic arch entry tear (OR = 5.25; 95% CI = 1.77–15.61;
p = 0.009). Older patients have a well-documented risk of NCs, probably due to advanced atherosclerotic lesions and a higher susceptibility to brain hypoperfusion once the aortic dissection occurs [
18,
28]. Regarding primary entry tears in the aortic arch, both in the aortic arch and ascending aorta, neither GERAADA, NORCAD nor IRCAD found any association with NCs in ATAAD [
2,
12,
13]. It is possible that normal flow in the supra-aortic vessels could be more frequently reduced by an entry tear in the aortic arch causing dynamic or static obstruction by the intimal flap [
29].
Cardiopulmonary bypass time (OR = 1.011; 95% CI = 1.01–1.02;
p < 0.001), aortic cross-clamp time over three hours (OR = 2.71; 95% CI = 1.43–5.14;
p = 0.002) and cerebral perfusion times (OR = 1.02; 95% CI = 1.002–1.03;
p = 0.027) were independent risk factors associated with newly developed postoperative NCs in the present study. This is in accordance with the GERAADA study, where increased operating times were associated with NCs. They found that cardiopulmonary bypass time and circulatory arrest time were independent risk factors [
2]. Aortic cross-clamp time was not investigated, and neuroprotective strategies such as anterograde cerebral perfusion were not associated with de novo NCs [
2]. We reported in a previous study that bilateral selective anterograde cerebral perfusion over 40 min is a risk factor for new postoperative ischemic stroke [
8]. Permanent neurological dysfunction was associated with longer operation time, while increasing hypothermic circulatory arrest was associated with temporary neurologic dysfunction, in a study of 501 patients with aortic arch surgery [
27].
There were no differences between the neurologically complicated and uncomplicated groups regarding the type of surgery or site of cannulation. The most frequent procedure was ascending aorta and hemiarch replacement with moderate hypothermic circulatory arrest and bilateral anterograde selective cerebral perfusion, and the most frequent cannulation site was the axillary artery. This result is supported by data from the current literature [
2,
30,
31] and suggests that cerebral flow can be influenced by multiple factors such as flow direction, pressure, temperature, or atherosclerosis lesions.