Safety and limitations of line pressure-targeted cerebral perfusion strategy in aortic arch surgery

Abstract

OBJECTIVES

The optimal flow rate for selective antegrade cerebral perfusion during aortic arch surgery is unknown. While 10–15 ml/kg/min is generally recommended, our centre has adopted a line pressure-targeted, relatively low-flow antegrade cerebral perfusion strategy. We aimed to evaluate the effect of flow rate on neurological outcomes.

METHODS

Patients without preoperative neurological deficits who underwent aortic arch surgery between January 2018 and May 2023 were enrolled. Under moderate hypothermia, an initial cerebral ischaemia period of 15–20 min was permitted. Following a brief retrograde cerebral perfusion, bilateral selective antegrade cerebral perfusion was performed using balloon-tipped perfusion catheters. The flow rate was determined using a line pressure of 200 mmHg. Risk factor analysis for postoperative permanent and temporary neurological deficits was conducted.

RESULTS

A total of 262 patients were included. The median selective antegrade cerebral perfusion flow rate was 5.7 ml/kg/min. Permanent neurological deficits occurred in 2 patients (0.8%), while temporary neurological deficits occurred in 30 patients (11.5%). The low antegrade cerebral perfusion flow rate was not a risk factor for permanent or temporary neurological deficits. Prolonged cerebral ischaemia time was the only significant risk factor for temporary neurological deficits.

CONCLUSIONS

Under moderate hypothermia and with the assistance of retrograde cerebral perfusion, the line pressure-targeted, relatively low-flow selective antegrade cerebral perfusion strategy at our centre achieved favourable neurological outcomes. However, prolonged cerebral ischaemia time was a significant risk factor for temporary neurological deficits.

Open in new tabDownload slide

INTRODUCTION

Despite significant advancements in aortic arch surgery, neurological deficit remains a major complication. The incidence of permanent neurological deficits (PND) is 2.9–18.9%, whereas that of temporary neurological deficits (TND) is 0.9–10.3% [1, 2]. Therefore, cerebral protection is crucial in aortic arch surgery.

Cerebral protection methods in aortic arch surgery have evolved from the use of deep hypothermic circulatory arrest (DHCA) to the more recent widespread adoption of moderate hypothermia with selective antegrade cerebral perfusion (sACP) [2–4]. However, the optimal flow rate for sACP is unknown. While 10–15 ml/kg/min is generally recommended [2, 5], our centre has adopted a line pressure-targeted, relatively low-flow sACP strategy.

We aimed to evaluate the impact of flow rate on neurological outcomes after aortic arch surgery using a line pressure-targeted sACP strategy.

PATIENTS AND METHODS

Study population

This was a retrospective, observational study, which adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting on observational studies [6].

This study included patients who underwent aortic arch surgery after we had clearly established our sACP strategy. Consequently, it included consecutive patients who underwent surgery between January 2018 and May 2023. Patients who potentially had neurological deficits before surgery were excluded. To ensure a homogeneous population in terms of the surgical procedure, (i) patients who underwent sACP via axillary cannulation with innominate artery clamping to prevent embolic stroke due to a shaggy aorta; (ii) requiring concomitant cardiac surgery where prolonged aortic cross-clamp time was expected and necessitated performing aortic arch vessel anastomosis first; and (iii) with extensive aortic aneurysms treated through a thoracotomy approach instead of sternotomy were excluded (Fig. 1). In some patients who underwent axillary artery cannulation for reasons other than a shaggy aorta, sACP was performed using a balloon-tipped perfusion catheter inserted into the innominate artery rather than an axillary artery cannula with innominate artery clamping, not due to specific reason but rather because we were accustomed to this approach. These patients were included in this study.

Ethical statement

This study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (approval number: B-2309-853-101). Because of the retrospective design of this study, informed consent was not required.

Operative techniques

Aortic arch surgery was performed at our centre according to the following principles and sequence (Fig. 2). The detailed surgical procedure and the composition of the ACP circuit are described in the Supplementary Material.

Arterial cannulation was primarily performed through direct cannulation of the ascending aorta. The core temperature was lowered to 25°C, as measured by rectal temperature. At the same time, the nasopharyngeal temperature usually dropped below 20°C. Once the HCA period began, we allowed an initial cerebral ischaemia period of 15–20 min while trimming the arch vessels and distal anastomosis site, during which sACP was not performed. This procedure was performed to prevent blood from flowing into the surgical field and obstruct visibility.

We applied brief retrograde cerebral perfusion (RCP) for approximately 2–3 min. Bilateral sACP was performed using balloon-tipped perfusion catheters, including a 12 Fr distal perfusion catheter inserted into the innominate artery and a 9 Fr Pruitt irrigation occlusion catheter inserted into the left common carotid artery (LeMaitre Vascular Inc., Burlington, MA, USA). These catheters do not allow measurement of the tip perfusion pressure of the balloon tip. The temperature of the blood used for sACP was maintained at 25°C. The sACP flow was gradually increased in increments of 50 ml/min to achieve a line pressure of 200 mmHg measured immediately downstream from the roller pump. Right radial artery pressure was routinely measured; however, it was not used as a target to adjust the flow rate of sACP. During sACP maintenance, proximal aorta anastomosis was performed first. The left common carotid and innominate arteries were anastomosed last.

Cerebral monitoring was performed as follows: near-infrared spectroscopy (NIRS) was continuously monitored from the start of anaesthesia until the end of surgery with sensors attached to both frontal lobes. No specific target values were set for regional cerebral oxygen saturation (rSO₂) measured by NIRS, and we did not increase sACP flow beyond a line pressure of 200 mmHg alone to increase rSO₂. rSO₂ was manually recorded in the anaesthesia record every 20 min by an anaesthesiologist. Electroencephalography and transcranial Doppler imaging were not performed.

Definitions

Cerebral ischaemia time was the period from the start of HCA until the initiation of RCP. The lower-body ischaemia time was the period from the start of HCA to the initiation of lower-body perfusion. sACP time was the period between sACP initiation and discontinuation. Additional rewarming time was the time required to further increase body temperature from completing arch vessel anastomosis to terminating cardiopulmonary bypass (CPB).

The sACP flow rate (ml/min) remained constant throughout the entire procedure in most patients. In cases where flow rate adjustments were made due to line pressure changes, the average value was calculated.

rSO₂ measured using NIRS monitoring was collected at the closest values to the following time points: after intubation, CPB initiation, cerebral ischaemia initiation, sACP initiation, lower-body perfusion initiation, aortic cross-clamp release, and CPB termination.

Primary outcome

The primary outcomes were the postoperative PND and TND. PND was defined as a new focal or global cerebral dysfunction caused by a blood perfusion disturbance in the brain, which was confirmed by imaging studies and did not resolve before discharge. TND was defined as confusion, agitation, delirium, or focal neurological deficits that were completely resolved before discharge. The incidence of delirium was assessed using the confusion assessment method for the intensive care unit during the intensive care unit stay. In the general ward, it was defined as cases in which psychiatric consultation was required due to related symptoms or when antipsychotic drugs were administered. The clinical scale of the TND with symptoms related to delirium was divided into 5 grades [7].

Statistical methods

Statistical analyses were performed using R software (version 4.3.3; R Foundation for Statistical Computing, Vienna, Austria). Categorical variables are expressed as frequencies and percentages. The Shapiro–Wilk test was used to determine whether continuous variables followed a normal distribution. Continuous variables with normal distribution are presented as means and standard deviations. Continuous variables that did not follow a normal distribution are presented as medians and interquartile ranges. There was no missing data in the other variables; however, it was present in the right radial artery mean arterial pressure and rSO₂. Data imputation was not performed due to the difficulty in establishing a correlation between the missing values and other variables. Paired t-tests were performed to evaluate the differences between the rSO₂ measured at each time point and the baseline rSO₂ measured after intubation, including only cases with values available for both sides. Due to multiple comparisons, the p-values were adjusted using the Bonferroni correction method. Considering the presence of missing data, rSO₂ was not used in the risk factor analysis of PND and TND. Risk factor analysis was conducted using logistic regression. Multivariable risk factor analysis was performed using variables with a P < 0.200 in the univariable analysis, as well as variables that the researchers considered to have clinical significance (e.g. factors such as age). A stepwise backward elimination method was used for variable selection. Statistical significance was set at P < 0.05.

RESULTS

Overall, 262 patients were included. The mean age was 69.2 years, and 50% of the patients were diagnosed with acute aortic syndrome (Table 1).

Total arch replacement was performed in 75.2% of patients. The mean rectal and nasopharyngeal temperatures were 24.4°C and 19.1°C, respectively. The mean cerebral ischaemia time was 18.8 min. The median sACP flow rate was 400 ml/min, equivalent to a median of 5.7 ml/kg/min. sACP was performed with a flow rate below 10 ml/kg/min in 98.5% of patients. The average right radial artery mean arterial pressure was 28.0 mmHg during sACP without lower-body perfusion, increasing to a mean of 39.0 mmHg after initiating lower-body perfusion (Table 2).

The rSO₂ of both the right and left cerebral hemispheres showed a significant difference at all time points except after the aortic cross-clamp release, compared to the baseline (Supplementary Material, Table S1 and Fig. 3).

Among all patients, PND and TND occurred in 2 (0.8%) and 30 (11.5%) patients, respectively (Table 3). Both cases of PND were embolic ischaemic strokes and not watershed infarctions caused by cerebral hypoperfusion (Supplementary Material, Table S2). Cerebral ischaemia time [odds ratio (OR), 1.04; 95% confidence interval (CI), 0.89–1.22; P = 0.590] and sACP flow rate (ml/kg/min) (OR, 1.11; 95% CI, 0.53–2.37; P = 0.776) were not risk factors for PND. However, in the multivariable risk factor analysis, prolonged cerebral ischaemia time (OR, 1.05; 95% CI, 1.00–1.10; P = 0.035) was the only significant risk factor for TND (Table 4).

DISCUSSION

This study has 2 main findings. First, the line pressure-targeted, relatively low-flow sACP strategy (median 5.7 ml/kg/min) at our centre resulted in favourable neurological outcomes similar to those previously reported. Second, although a low sACP flow rate was not a risk factor for PND or TND, prolonged cerebral ischaemia time was a significant risk factor for TND.

sACP can be performed either unilaterally through the main arterial line via axillary artery cannulation or bilaterally using balloon-tipped perfusion catheters separately inserted into the innominate artery and the left common carotid artery. It has not yet been established whether unilateral sACP or bilateral sACP is superior [8, 9].

Similarly, the appropriate flow rate for sACP in aortic arch surgery is unknown. Currently, an acceptable flow rate is derived from several animal studies. Haldenwang et al. reported that a high flow rate of 18 ml/kg/min, compared with 8 ml/kg/min, increased intracranial pressure and caused cerebral oedema but did not enhance regional or global cerebral flow in a pig model [10]. Jonsson et al. reported that the safe minimal flow rate during sACP at a body temperature of 20°C in a pig model was 6 ml/kg/min [11]. This result was based on significant differences in mixed venous oxygen saturation and protein S-100β levels between groups, as well as differences in cerebral perfusion observed on magnetic resonance imaging. However, neurological deficits could not be clinically evaluated postoperatively. According to a recent survey conducted in European centres and meta-analyses, most centres typically use a flow rate of 10–15 ml/kg/min [2, 5]. A recently published Japanese Aortic Disease Guideline states that the ACP flow rate is generally 10 ml/kg/min [12]. However, recently published American and European guidelines do not specify appropriate ACP flow rates [13, 14]. No randomized controlled trials have investigated the flow rate of sACP in humans.

In this study, we did not determine the flow rate of the sACP as a target; instead, we determined it based on the flow at which the line pressure reached 200 mmHg. This approach primarily aimed to prevent haemolysis, which could occur in the balloon-tipped perfusion catheter (1/16 inch) and perfusion line (3/16 inch). Consequently, the median sACP flow rate was 5.7 ml/kg/min. Moreover, we performed the arch vessel anastomosis at the end of the procedure, continuing low-flow ACP with perfusion temperature at 25°C even when the body temperature had normalized. Nevertheless, the incidences of PND and TND were not higher than those previously reported.

For most patients, the flow rate was determined to be 300–400 ml/min. Since the absolute flow rate was similar across patients, the flow rate per body weight (ml/kg/min) was lower in patients with higher body weight and higher in those with lower-body weight. However, no strong evidence supports that brain weight increases linearly with body weight in humans [15, 16]. Therefore, it is difficult to infer that patients with higher body weights exhibit increased cerebral metabolism, necessitating a higher blood flow rate. Therefore, we cautiously propose that what may actually be required is not a weight-adjusted flow rate (ml/kg/min) but rather an absolute minimal flow rate (ml/min). Determining the sACP flow rate using the line pressure as the target could be a viable approach.

Recently, Friess et al. reported a method for sACP that is almost identical to ours but considers rSO₂ and line pressure [17]. They increased the sACP flow until the rSO₂ reached the awake baseline value or stopped increasing the flow if the line pressure reached 300 mmHg instead of 200 mmHg. The reaching the awake baseline rSO₂ rate was significantly higher at 8 vs 6 ml/kg/min, with no further significant improvements at 10 ml/kg/min. Despite most cases involving hemiarch replacement with a median cerebral ischaemia duration of 5 min and an sACP duration of 11 min, 14 of the 40 patients (35%) developed TND.

NIRS is widely used for aortic arch surgery. However, an established rSO₂ target remains unknown. Recent reviews suggest that evidence regarding whether cerebral oximetry reduces the incidence of PND and TND varies between studies [18, 19]. The definitions of rSO₂ desaturation also differ, ranging from 10% to 30% relative to the baseline or absolute values of 40% to 60%. In this study, rSO₂ significantly decreased during the cerebral ischaemia period compared to baseline and gradually recovered after sACP initiation. Owing to missing data, no analysis was conducted to determine whether rSO₂ desaturation is associated with PND or TND. The inability to establish this association is a major limitation of this study.

Prolonged cerebral ischaemia time was the only risk factor for TND, which is consistent with findings of an early study conducted during the DHCA era, showing that TND incidence increased with longer DHCA duration [20]. TND does not result in permanent damage; however, it may be associated with long-term cognitive dysfunction and impaired quality of life [7, 21]. When implementing our sACP strategy, the operator must determine whether it is more beneficial to prolong cerebral ischaemia time to achieve a clear surgical field and perform the operation with precision, thereby reducing lower-body ischaemia time, or to prioritize minimizing cerebral ischaemia time to lower the risk of TND.

Study limitations

This study has several other significant limitations. First, during the study population selection process, patients with a high risk of embolic stroke due to a shaggy aorta who underwent axillary cannulation and those who received sACP using the axillary cannula were excluded from the study, which may have led to selection bias. Considering that most PNDs after aortic arch surgery are caused by embolic strokes [20, 22], the low incidence of PND observed in this study may be lower than the actual incidence of PND among all patients who underwent aortic arch surgery at our centre. It is crucial to interpret the study results carefully and consider this limitation. Second, Weijs et al. recently reported that cerebral blood flow, measured using transcranial Doppler monitoring during sACP, significantly differed between patients with and without postoperative neurological deficits [23]. However, because we did not perform transcranial Doppler monitoring, direct measurement of the actual cerebral blood flow delivered through our centre’s sACP strategy was not possible. Third, this study has a retrospective design and lacks a control group for comparison with our sACP strategy, which may introduce selection bias. Finally, since this study was conducted at a single institution, its findings may not be generalizable to other centres with different patient demographics and surgical strategies.

CONCLUSION

Under moderate hypothermia and with the assistance of RCP, the line pressure-targeted, relatively low-flow sACP strategy at our centre achieved favourable neurological outcomes. However, prolonged cerebral ischaemia time was a significant risk factor for TND.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

None declared.

Conflict of interest: none declared.

ACKNOWLEDGMENTS

We would like to thank Editage (www.editage.co.kr) for English language editing. We would like to thank Division of Statistics in Medical Research Collaborating Center at Seoul National University Bundang Hospital for their invaluable expertise and guidance, actively contributing to the data analysis and consultation.

DATA AVAILABILITY

The data underlying this article will be shared upon reasonable request to the corresponding author.

Author contributions

Joon Chul Jung: Conceptualization; Data curation; Formal analysis; Writing—original draft. Hyoung Woo Chang: Methodology; Visualization. Jae Hang Lee: Resources; Supervision. Kay-Hyun Park: Resources; Writing—review and editing

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Tim Berger, Luca Di Marco and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

 
• CI

  Confidence interval

 
• CPB

  Cardiopulmonary bypass

 
• DHCA

  Deep hypothermic circulatory arrest

 
• HCA

  Hypothermic circulatory arrest

 
• NIRS

  Near-infrared spectroscopy

 
• OR

  Odds ratio

 
• PND

  Permanent neurological deficit

 
• RCP

  Retrograde cerebral perfusion

 
• rSO₂

  Regional cerebral oxygen saturation

 
• sACP

  Selective antegrade cerebral perfusion

 
• TND

  Temporary neurological deficit