Perfusion Technique in Total Arch Replacement to Prevent Cerebral Stroke in Patients Over 75 Years Old a Single Center Retrospective Cohort Study


 BackgroundTo evaluate the efficacy of isolated cerebral perfusion (ICP) for the prevention of cerebral infarction and the occurrence of early mortality, we retrospectively compared the outcomes between old patients with or without ICP in total arch replacement (TAR)MethodsBetween January 2003 and March 2019, 74 patients aged older than 75 years underwent elective TAR for arch aneurysm. The participants were divided into two groups according to the brain protection method used: the ICP method (ICP group, n = 46) and the selective antegrade cerebral perfusion method (non-ICP group, n = 28). ResultsThe 30-day mortality rates of both groups were zero. The in-hospital mortality rates of the ICP and non-ICP groups were 4.3% and 3.5%, respectively. The median follow-up duration was 34 months, and the 1-year survival rates were 90.1% in the ICP group and 90.7% in the non-ICP group. The incidence of postoperative cerebral infarction was significantly lower in the ICP group (n = 1, 2.1%) than in the non-ICP group (n = 5, 17.8%).ConclusionsICP could result in a lower incidence of postoperative cerebral infarction in old patients who undergo TAR.


Abstract Background
To evaluate the e cacy of isolated cerebral perfusion (ICP) for the prevention of cerebral infarction and the occurrence of early mortality, we retrospectively compared the outcomes between old patients with or without ICP in total arch replacement (TAR)

Methods
Between January 2003 and March 2019, 74 patients aged older than 75 years underwent elective TAR for arch aneurysm. The participants were divided into two groups according to the brain protection method used: the ICP method (ICP group, n = 46) and the selective antegrade cerebral perfusion method (non-ICP group, n = 28).

Results
The 30-day mortality rates of both groups were zero. The in-hospital mortality rates of the ICP and non-ICP groups were 4.3% and 3.5%, respectively. The median follow-up duration was 34 months, and the 1year survival rates were 90.1% in the ICP group and 90.7% in the non-ICP group. The incidence of postoperative cerebral infarction was signi cantly lower in the ICP group (n = 1, 2.1%) than in the non-ICP group (n = 5, 17.8%).

Conclusions
ICP could result in a lower incidence of postoperative cerebral infarction in old patients who undergo TAR.

Background
The outcomes of open surgery for aortic aneurysm have improved with the development of surgical techniques and innovation in brain protection methods [1,2 ]. Intraoperative cerebral infarction is a complication of such procedure that has not been resolved [3]. Although the incidence of hemodynamic infarction may have decreased with improvements in selective cerebral perfusion, embolic infarctions are still considered a problem, particularly in old patients [4]. To decrease the incidence of cerebral infarctions in aortic surgery, we introduced the use of isolated cerebral perfusion (ICP), a novel procedure for extracorporeal circulation [5]. To evaluate the e cacy of ICP for the prevention of cerebral infarction and the incidence of early mortality, we retrospectively compared the outcomes between old patients with or without ICP in total arch replacement (TAR).

Methods
Between January 2003 and March 2019, 74 patients aged older than 75 years underwent elective TAR for arch aneurysm at Yokohama City University Medical Center. We excluded emergency cases of type A acute aortic dissection and aortic rupture and cases of redo aortic surgery for ascending aorta and aortic arch. Patients who underwent emergency surgery for aortic dissection or rupture of the aorta were not included because these conditions might have preoperative in uence on the brain. Preoperatively, magnetic resonance imaging (MRI) or head computed tomography (CT), MR angiography (MRA), and carotid ultrasound were conducted to assess the brain, cerebral circulation, and the neck vessels, respectively.
We have been performing ICP during TAR since 2010. Until 2009, TAR was performed with conventional selective cerebral perfusion, in which the cannulae were directly inserted from the inside of the aorta after systemic perfusion cooling via various arterial cannulations.
All patients (n = 74) were divided into two groups according to the brain protection method used: the ICP method (ICP group, n = 46) and the SCP method (non-ICP group, n = 28).

Set-up during the ICP procedure
The bilateral axillary arteries were exposed before sternotomy. A 9-mm Dacron graft was anastomosed to the bilateral axillary arteries for systemic perfusion. After median sternotomy, bi-caval venous drainage was performed. The left common carotid artery (LCCA) was exposed without touching the aorta. The LCCA was proximally clamped and dissected, and a 12-Fr balloon-tipped cannula with a pressure monitor was inserted immediately. Extracorporeal circulation via the bilateral axillary artery and selective perfusion to the LCCA were simultaneously started. Separate roller pump was used to regulate the blood ow to LCCA. The mean pressure at the radial artery and LCCA were monitored and controlled between 40 and 70 mmHg. Circulatory arrest was induced at rectal temperature of 25°C. The brachiocephalic artery and left subclavian artery were clamped. ICP was immediately completed with blood ow through the bilateral axillary arteries and the LCCA ( Figure 1).
With perfusion via the ascending aorta or the right axillary artery, the selective cerebral perfusion canulae were directly inserted from the inside of the aorta after circulatory arrest. When the bilateral axially arteries were used, the brachiocephalic artery and the left subclavian artery were clamped just after circulatory arrest, and the cerebral perfusion cannula was directed into the left carotid artery.

Neurological diagnosis
The postoperative neurological symptoms based on the clinical records were retrospectively reviewed. Postoperatively, intensivists checked the neurologic state of the patients every day in the intensive care unit. For symptomatic patients, the Department of Neurology was always consulted, and neurologists completed a neurological examination via CT scan in symptomatic patients. Postoperative cerebral infarction was de ned as the persistence of permanent neurological dysfunction at discharge, which was accepted with the image and diagnosed by a neurologist. Transient neurologic dysfunction, which includes delirium, transient ischemic attack, and any neurologic dysfunction that was not judged as postoperative cerebral infarctions, was not assessed in this study.

De nitions
Cerebrovascular and carotid abnormality was de ned as a carotid artery stenosis diagnosed via carotid ultrasound, which revealed >70% of stenosis in the area, or the occluded lesion of carotid or vertebral arteries and circle of Willis diagnosed by MRA. Previous stroke was de ned as the preoperative clinical history of cerebrovascular disease that did not include transit ischemic attack.

Statistical analysis
Continuous data were expressed as median and interquartile range (25 th -75 th percentile). Discrete data were shown as counts and percentage. Continuous variables were compared with the Mann-Whitney U test, and categorical variables were compared with Pearson Χ 2 test or Fisher exact test. Kaplan-Meier analysis was performed to calculate survival.

Results
The characteristics of the patients are summarized in Table 1. There was signi cant difference in age, sex, hyperlipidemia, and chronic renal failure. Surgical outcomes are summarized in Table 2. No signi cant difference was observed in terms of operative time, pump run, aortic clamp time, lower circulatory arrest time, and amount of uid balance between the two groups. The 30-day mortality rates of both groups were zero. The in-hospital mortality rates in the ICP and non-ICP groups were 4.3% and 3.5%, respectively. The causes of death were sepsis in the ICP group and sepsis due to respiratory failure in the non-ICP group. The occurrence of postoperative cerebral infarction was signi cantly lower in the ICP group than in the non-ICP group (2.1% vs 17.8%, p = 0.026). The distribution of postoperative cerebral infarction concerning number and laterality is shown in Figure 2. Prolonged ventilation time (>72 h) and ICU stay did not signi cantly differ between the two groups ( Table 3). The median follow-up duration was 34 months, and the 1-year survival rates were 90.1% in the ICP group and 90.7% in the non-ICP group.  In the ICP group, there was one case with perioperative stroke. He was an 84-year-old man who had a plaque in his BCA on enhanced CT. He had left hemiparesis after operation; a subsequent CT scan revealed multiple small infarctions on the right side of the brain. Although he had left hemiparesis after operation, he recovered to an extent that he could walk and was therefore transferred to a rehabilitation hospital.

Discussion
Antegrade cerebral perfusion (ACP) and hypothermic circulatory arrest with or without retrograde cerebral perfusion (HCA/RCP) are often used in aortic arch surgery. Okita et al. [3] revealed the clinical outcomes concerning aortic arch surgery in Japan and showed that the incidence of stroke was 6.7% in ACP and 8.6% in HCA/RCP.
Axillary artery perfusion for arch repair was reported to be effective in preventing embolic stroke and reducing early mortality [6,7]. Kim et al. have revealed that the incidence of right-side hemisphere stroke is signi cantly lower with axillary artery perfusion in open arch repair, which indicates that axillary artery perfusion could be useful in reducing early embolic stroke and mortality [7].
However, the risk of stroke in SCP procedure with axillary perfusions is unacceptably high at about 10% in our institution until 2009 [8]. We assumed that even bilateral axillary perfusion cannot completely prevent the effect of embolisms in the aorta due to turbulent blood ow with the arti cial heart-lung machines around the arch. To reduce cerebral infarction during aortic surgery, we introduced the use of ICP, which is a novel procedure for extracorporeal circulation.
The important advantage of this procedure is that it could prevent aortic embolism during operation. It is possible for ICP to prepare the cerebral perfusion seldom touching the ascending aorta. In ICP, the systemic perfusion blood ow from the bilateral axillary perfusion went into the aorta, and the LCCA was clamped just before systemic perfusion was started to prevent turbulent ow with emboli from going to the cerebral circulation. The effectiveness of establishing cerebral perfusion before circulatory arrest for severe atheromatous aortas was reported by Shiiya et al [9]. Our method may have more advantages than their method because perfusion can be initiated without touching the arch and neck vessels, other than the LCCA.
In the non-ICP group, several methods of systemic perfusion were used based on preoperative CT ndings and surgeon's preference to prevent the occurrence of morbidity. Although in the non-ICP group, axillary artery perfusion was used in 86% (24/28) of patients and bilateral axillary artery perfusion was used in 68% (19/28) of patients, stroke occurred in 17.8% (5/28) of patients. Furthermore, cerebral infarctions of the left dominant hemisphere were recognized in three patients. (Figure 2) According to our study, using axillary artery perfusion alone cannot effectively prevent serious cerebral infarction. It was found that the incidence of stroke was lower in the ICP group than in the non-ICP group, even if the mean age was >2 years. According to our study, using axillary artery perfusion alone cannot effectively prevent cerebral infarction.
Perioperative stroke induced by aortic emboli can be prevented by isolating the blood ow to the brain completely from the atheromatous aorta by clamping the LCCA just before initiation of systemic perfusion. In the ICP group, there was one case of stroke. The existence of a plaque in his BCA, which may have been the clamp site, might have contributed to the event. In this study, we could not assess the state of arch vessels into which the selective cerebral perfusion cannula was inserted or where the clamp site could be. It could be di cult even for ICP to avoid the emboli completely induced by clamping the arch vessels where the atheroma was rich. More precise examination of the arch vessels using preoperative multidetector CT and/or intraoperative epi-aortic echography may contribute to further improvement.
A disadvantage of the ICP procedure is the long operation time due to preparation of bilateral axillary arteries and anastomosis of the grafts to them. Comparing with other studies of TAR, our operative time in both groups was longer and the amount of uid balance was higher. Moreover, a longer operative time could result in uid balance greater than 5000 mL in our study [10].
Our hospital is an educational hospital. Because some perfusion sites were selected in the non-ICP group, there could be no signi cant difference in operation time. Stroke could make subsequent quality of life worse; therefore, we believe that it is worth performing the ICP procedure even if the operation time is extended because of the possible effect of preventing stroke in aortic surgery.
The present study had some limitations. The current study was a retrospective cohort study and was conducted on patients who underwent total arch repair in a single hospital. Furthermore, because of the small number of cases, risk adjustment was not performed. Postoperative CT scan or MRI of the brain was not performed on patients who did not present with symptoms to rule out asymptomatic stroke after the operation.  The direction of blood ow in ICP is indicated by arrows. (a) Systemic perfusion during cooling (b) Cerebral perfusion during circulatory arrest