Matched comparison of 3 cerebral perfusion strategies in open zone-0 anastomosis for acute type A aortic dissection

Abstract

Open in new tabDownload slide

OBJECTIVES

The present study aims to investigate outcomes after the surgical treatment of acute type A aortic dissection in regard to three available selective cerebral perfusion strategies.

METHODS

From 2000 to 2019, patients were selected based on the employment of either retrograde cerebral perfusion (RCP), unilateral antegrade cerebral perfusion (uACP) or bilateral antegrade cerebral perfusion (bACP) during open zone-0 anastomosis. Propensity score TriMatch analysis considering several preoperative and intraoperative variables was used to identify well-balanced triplets. The primary end point of the study was a new cerebral operation-related neurologic deficit.

RESULTS

Operative times (operation time, cardiopulmonary bypass time, reperfusion time) were significantly longer in the RCP group, in which deeper hypothermia was applied (27.5 [24–28], 28 [26–28] and 16 [16–17]°C for uACP, bACP and RCP, respectively, P-value <0.001). The RCP group showed higher red blood cell concentrates and fresh frozen plasma transfusion rates. No significant difference of new cerebral operation-related neurologic deficit was observed between the 3 groups (12.9% vs 12.9% vs 11.3% for RCP, uACP and bACP, P-value = 0.86). In addition, 30-day mortality showed similar distribution independently of the cerebral perfusion strategy adopted (17.7% vs 14.5% vs 17.7% for RCP, uACP and bACP, P-value = 0.86).

CONCLUSIONS

However, based on a small sample size, the comparison showed no relevant differences in terms of neurologic outcome and 30-day mortality, confirming RCP, uACP and bACP as safe and reproducible selective cerebral perfusion strategies in surgery for acute type A aortic dissection.

INTRODUCTION

The surgical treatment of acute type A aortic dissection (ATAAD) remains challenging and is still associated with high mortality and morbidity rates, especially in the elderly [1]. The main target of surgery is the excision of the primary intimal tear and an open distal anastomosis to seal false lumen and expand true lumen, resolving malperfusion [2]. Present surgical approaches require the use of hypothermic circulatory arrest, possibly employing selective cerebral perfusion strategies. Several concepts were developed over the years: retrograde cerebral perfusion (RCP) was suggested in the 1990s as adjunct to straight deep hypothermic circulatory arrest (DHCA), whereas antegrade cerebral perfusion (ACP) became more popular in the following decades, allowing circulatory arrests at moderate hypothermia [3, 4]. Antegrade perfusion is widely used both in unilateral antegrade cerebral perfusion (uACP) and bilateral antegrade cerebral perfusion (bACP) fashion and was shown to provide favourable outcomes, especially with prolonged circulatory arrest times [5]. Although outcomes of these perfusion strategies have been extensively analyzed and bACP has currently become the most used method, a unanimous consensus on which selective cerebral perfusion should be preferred is still missing [4].

Therefore, we aimed to investigate the results of RCP, uACP and bACP after surgical treatment for ATAAD.

METHODS

Ethics statement

The study was approved by the local ethics committee (Charité Medical School, Berlin, Germany, No. EA2/096/20) and complies with the Declaration of Helsinki. Patient’s informed consent was waived due to the retrospective nature of the study.

Patient population

Between January 2000 and December 2019, we retrospectively analyzed our in-house ATAAD database for patients that underwent surgical repair with an open distal anastomosis under hypothermic circulatory arrest. Exclusion criteria were surgeries performed under DHCA alone and iatrogenic ATAAD. Stable patients with pain-to-cut time exceeding 48 h were also excluded since in some cases such dissections undergo delayed aortic repair and would bias the acute surgical setting of this analysis. To increase baseline homogeneity, patients treated with extensive aortic arch surgery were also excluded. The definition of the final study population and the partition in three groups according to the selective cerebral perfusion strategy is summarized in Fig. 1.

Surgical technique

The preferred cannulation site for arterial perfusion was the right axillary or the femoral artery for ACP and RCP, respectively. A moderate hypothermia (24–28°C) was achieved in the ACP surgeries, whereas deeper temperatures (16–17°C) were reached in the RCP group. According to the different cannulation setting and cooling protocol, in the ACP group, an arch-first approach was favoured. Based on operative reports, a clear discrimination between isolated ascending and hemiarch replacement was made, considering if the operating surgeon expressly replaced the inner arch curvature. The distal anastomosis was routinely performed using felt stripes as reinforcement. The aortic root was addressed dependently on the presence of entry tears below the sinotubular junction, on the depth of the dissection in the sinus valsalvae, on the involvement of the coronary arteries and on the aortic valve morphology. In the absence of these conditions, big effort was made to repair the aortic root, mainly through commissure resuspension, glueing of the dissected aortic walls and sealing of the new sinotubular junction with felt sutures. In case of massive disruption, the root was replaced.

Selective cerebral perfusion strategies

The setting of ACP considered a 10 ml/kg/min flow through the right axillary artery in case of uACP, with additional direct cannulation of the left common carotid artery for bACP.

In contrast, RCP was established after insertion of a standard cannula for bicaval cannulation in the superior caval vein and cardiopulmonary bypass (CPB) arrest. Target of RCP was a central venous pressure between 30 and 40 mmHg, meaning in the majority of cases a 500 ml/min flow through a separate pump. The azygos vein was not routinely clamped during RCP. Intraoperative monitoring of the selective cerebral perfusion through near infra-red spectroscopy was not available for all patients, since it was routinely performed starting from 2013. It is to mention that RCP was mainly performed in the first period object of this study (2000–2013) and was gradually replaced by ACP strategies in the following years (Fig. 2). The selection of a certain selective cerebral perfusion, when all the three strategies were available, was made according to individual preference of the operating surgeon, anatomy and intraoperative course. In case of expected long circulatory arrest times or significant perfusion differences between the 2 hemispheres (according to clinical/radiological signs of malperfusion or near infra-red spectroscopy mismatch), bACP was preferred over uACP. If ACP was considered to be disadvantageous or even harmful, i.e. in cases of massive atherosclerosis or involvement of the supra-aortic vessels, RCP was performed.

Neurologic symptomatology assessment

The study end point was a new operation-related cerebral deficit based on clinical and/or radiological diagnosis. The end point was defined as follows: the neurologic deficit had to be new (not present preoperatively), operationrelated (diagnosis was made immediately after narcosis termination at the first postoperative neurological examination) and cerebral (clinical or radiological diagnosis of cerebral accident, i.e. stroke, bleeding or oedema). Our institution is a single cardiac surgery unit, without a first aid department. This means that every ATAAD is referred to us from another hospital, where the initial diagnostic process takes place. The presence of acute preoperative neurologic dysfunction was clinically assessed at admission in our hospital. Preoperatively intubated and ventilated patients or cases requiring preoperative cardiopulmonary resuscitation, for whom no information on neurologic status could be achieved, were excluded from the primary end point. These patients did not routinely undergo brain computed tomography (CT) before surgery to save time and try to restore adequate cerebral perfusion. Postoperative neurologic deficits in these patients were classified as ‘not new’ because they could not safely be defined as operation-related.

Statistical analysis

Continuous variables were tested for normal distribution using the Shapiro–Wilk test. All variables were not normally distributed and were therefore presented as median with corresponding interquartile range (25th–75th percentile) after comparison using the Kruskal–Wallis test. Categorical data are depicted as absolute numbers with corresponding percentages and were compared using Chi square test. Propensity score matching (1:1:1) using the TriMatch-Package version 0.9.9 from R (The R Foundation for Statistical Computing), considering all variables listed in Table 1, was performed to balance the groups regarding clinically important preoperative and intraoperative variables and potential confounders. The standardized mean difference for all the mentioned variables is depicted in Fig. 3. Statistical analysis was performed using R (The R Foundation for Statistical Computing) version 4.0.2.

RESULTS

Baseline characteristics

Preoperative and intraoperative data used for propensity score matching are shown in Table 1. The primary study population was divided into 3 groups according to the selective cerebral perfusion strategy employed, obtaining 198, 95 and 531 patients for uACP, bACP and RCP, respectively. One-fourth to one-third of patients experienced acute neurologic deficit, with the presence of any sign of malperfusion for about half of patients in each subgroup. Acute shock at admission was present in 26.3%, 23.0% and 16.2% for bACP, RCP and uACP, respectively (P-value = 0.07). PENN classification allocated 10.1% (uACP), 12.6% (bACP) and 12.8% (RCP) in the Abc class (combined ischaemia). Preoperative tamponade was present in 17.7%, 23.4% and 28.4% of patients for uACP, RCP and bACP, respectively (P-value = 0.09). After propensity score matching, groups were well balanced regarding preoperative and operative variables considered for the matching.

Surgical data

Retrograde cerebral perfusion was associated with longer operative times, including CPB, X-clamping and reperfusion time. Selective cerebral perfusion time was not significantly different among the groups: 31.5 (21–41) min for uACP, 35.5 (29.5–44.3) min for bACP and 35.5 (25.8–41) min for RCP (P-value 0.09). A deeper core temperature was applied when RCP strategy was employed (16 [16–17]°C vs 27.5 [24–28]°C for uACP vs 28 [26–28]°C for bACP, P-value <0.001). The intraoperative blood product transfusion rates showed relevant differences between the groups: RCP experienced a significantly higher need for red cell concentrates and fresh frozen plasma transfusions. A more comprehensive description of operative data is summarized in Table 2.

Impact of perfusion strategies on neurologic outcome and survival

The overall rate of NPONDs was 13.3%, 19.4% and 23.3% for bACP, uACP and RCP, respectively (P-value = 0.37). Onset of NPOND was directly attributable to type A operation in 11.3% for bACP, 16.1% for uACP and 17.7% for RCP (P-value = 0.50), ending in a cerebral operation-related NPOND rate of 11.3%, 12.9% and 12.9%, respectively (P-value = 0.95). In terms of survival, no significant difference in 30-day mortality was observed, with 17.7%, 14.5% and 17.7% mortality for bACP, uACP and RCP, respectively (P-value = 0.86).

Further details regarding the perioperative outcomes are presented in Table 3.

Intraoperative and outcome data for the unadjusted cohort are available in the Supplementary Material.

DISCUSSION

The goal of the present study is to address the issues regarding the choice of optimal selective cerebral perfusion strategy in the surgical treatment of ATAAD, performing a propensity score-based head-to-head comparison. Retrograde cerebral perfusion, uACP and bACP show no statistically significant differences in terms of neurologic deficits and operative mortality, although RCP performed under lower core temperatures was associated with longer operative times and higher blood product transfusion rates. No adjustment for multiple comparison regarding these results was performed, since its impact would not have been relevant for our conclusions.

The importance of immediate surgery and the establishment of adequate cerebral perfusion to prevent brain injuries have been previously described [6, 7]. This aspect becomes clearer in presence of preoperative malperfusion and metabolic dysfunction [8, 9].

Several international registries and expert consensus documents brought valuable information regarding presentation, optimal surgical treatment and postoperative outcome of ATAAD, underlining its multifactorial nature [10, 11]. In this constellation, great effort was made to define novel risk scores and classifications, to enhance a more comprehensive dissection assessment and facilitate standardized treatment approaches [12, 13]. The wide consensus regarding fast referral for surgery is in contrast to the lack of clear indications about which perfusion strategy should be adopted during hypothermic circulatory arrest. Several retrospective analyses demonstrated the superiority of selective cerebral perfusion in adjunct to hypothermia in comparison to straight DHCA, though not showing remarkable differences between RCP and ACP [14, 15]. Moreover, DHCA was recognized as an independent risk factor for permanent neurologic injury, especially in case of prolonged circulatory arrest times [5, 16]. Based on these previously published findings, we limited our investigation to the comparison of RCP, uACP and bACP only, excluding straight DHCA. Safety and favourable results in terms of survival as well as neurologic outcomes for RCP were extensively documented for elective and for emergent aortic surgery [3, 17]. Nevertheless, experimental and metabolic investigations failed to confirm RCP’s clinical results, since additional oxygen delivery and the flushing effect of embolic material could not be supported by animal models [18]. A recently published meta-analysis showed no relevant differences in terms of mortality and postoperative stroke between RCP and ACP [19]. Although the pooled effect sizes of this meta-analysis derive from remarkable 15 365 patients, the study is limited by the lack of discriminating between unilateral and bilateral ACP as well as by the inclusion of both elective aneurysms and acute aortic dissections. The advantage of ACP over RCP is the permission of milder hypothermia, which reduces CPB time and consecutively also bleeding complications, making this technique the recommended perfusion strategy of current European guidelines [2]. Our findings are also in line with this recommendation, since the RCP group under deep hypothermia experienced longer operative times, as well as higher red blood cell, concentrates and fresh frozen plasma transfusions. The use of uACP or bACP is still a matter of debate: uACP is favoured by many surgeons due to the possibility to switch from whole body to selective cerebral perfusion just by clamping of the innominate artery, avoiding extensive preparation and manipulation of supra-aortic vessels [20]. A recent analysis investigating uACP and bACP during ATAAD surgery showed an equal incidence of left hemispheric strokes and concluded that uACP provides comparable results to bACP [21]. This trend could not be confirmed in patients with prolonged circulatory arrest times (>50 min), where bACP was demonstrated to provide higher survival rates [22]. Considering the comparison of different perfusion strategies in the setting of ATAAD, the small numbers of patients from single centre experiences led to the adoption of statistical tools like propensity score matching, to adjust for possible confounders [23]. However, the available studies couple RCP and ACP only, with no differentiation between uACP and bACP.

The decision to exclude extensive arch surgery from our analysis had several reasons: it is the source of selection bias in terms of indication (arch entries, involvement of supra-aortic branches) and surgical strategy (bACP is adopted if prolonged circulatory arrest times are scheduled) but led to a small amount of bACP surgeries. The same considerations led also to the exclusion of partial arch replacements (open zone-1 or zone-2 anastomosis), since also these approaches need more preparation of the aortic arch and the reimplantation of at least 1 supra-aortic vessel, which is not the case for a simple zone-0 open anastomosis. The definition ‘open zone-0 anastomosis’ is mandatory due to the fact that, especially for the RCP group in the first period of this analysis, a more conservative aortic resection was performed, limited to the ascending aorta only. This led us to include the hemiarch resection into the matching process, achieving a final overall 90% rate for all the groups.

Dissection involvement of the common carotid arteries was shown to strongly affect neurologic outcomes [24, 25]. The inclusion of an anatomical definition of the dissection patterns in the supra-aortic vessels would have been of interest in the assessment of our neurologic end point according to the selective cerebral perfusion strategy employed but was not available for the whole cohort, especially for patients operated in the early 2000s. Similarly, we excluded the discrimination between DeBakey I and II dissections from the matching process due to the lack of information for the older patients and the fact that the surgical strategy would not differ considering an open zone-0 anastomosis, which is the standard treatment in both scenarios. The matching process involved several preoperative and intraoperative variables which were shown to strongly affect outcomes in the available literature, with great attention to preoperative haemodynamic condition and malperfusion [12, 16, 26]. After matching, the 3 subgroups present a rough 20% rate of both preoperative neurological deficit and shock, with any sign of malperfusion in 40.3%, 46.8% and 50.0% of patients for uACP, RCP and bACP, respectively, without significant differences between the groups. Operative times were not included in the matching process, since the need for deeper temperatures in our RCP setting was associated with longer CPB and reperfusion times. The same applied for arterial cannulation site since the right axillary artery was the primary cannulation site for ACP in over 90% of patients and just 11% for RCP, where femoral cannulation was the preferred strategy (86.3%). This could be the source of confounders in terms of neurologic outcome, taken into account that axillary cannulation is associated with a lower stroke rate and retrograde femoral perfusion has been questioned due to potential expansion of the false lumen and inferior long-term survival [27]. Similarly, operative temperature could not be included in the matching process, because of the absence of overlapping between all the 3 strategies in terms of cooling protocol. Although never performed at our institution, it would be valuable to investigate RCP at moderate hypothermia for a simple hemiarch anastomosis, to finally extrapolate the effect of selective cerebral perfusion irrespective of operative temperature.

Neurologic outcome and mortality rates represented our study end-points: cerebral operation related NPOND was 11.3%, 11.3% and 12.9% (P-value = 0.95) and 30-day mortality 14.5%, 17.7% and 17.7% (P-value = 0.86) for uACP, bACP and RCP, respectively. Particular attention was paid to the definition of NPOND: we believe that the comparison of different perfusion strategies should be based on the evaluation of postoperative cerebral deficits directly connected to the type A operation only. Based on this statement, acute preoperative neurologic deficits and postoperative deficits arising from revisions/reinterventions, cardiopulmonary resuscitation, or stroke/cerebral bleeding occurring after a first normal neurologic assessment following narcosis termination led to their exclusion from our primary end-point.

On this basis, the above mentioned 11.3–12.9% rate of cerebral operation-related NPOND represents a solid confirmation regarding the comparable cerebral outcomes of uACP, bACP and RCP during open zone-0 anastomosis for ATAAD.

Limitations

This analysis is limited by its retrospective nature and the single-centre experience. Moreover, the investigation on the relatively small amount of patients after propensity score matching leads to limited statistical power, which may have influenced our results and conclusions.

Although all groups were well balanced in regard to most relevant preoperative and intraoperative variables and confounders, a residual bias cannot be precluded, also considering the different temporal distribution of the investigated perfusion strategies. The lack of information on aortic involvement in preoperative CT scan is certainly a major limitation. Furthermore, the analysis is limited by the different surgical setting of the 3 selective cerebral perfusion strategies, which does not allow to match for operative temperature, arterial cannulation site and operative times. All these variables are ultimately intrinsic components of each selective cerebral perfusion strategy. In order to state differences between the groups, it would be preferrable to adjust for multiple comparison (i.e. Bonferroni’s correction). Nevertheless, regarding the assertions on our end-point variables, the impact of P-value adjustment would have been only marginal and therefore we decided to dispense for it.

In addition, the comparison is perioperative and a longer follow-up would add more power to the present analysis, providing data on outcomes on the long term.

This analysis refers only to a sub-group of patients undergoing surgery for ATAAD, although quantitatively the most significant (open zone-0 anastomosis). Therefore, the present results cannot be extrapolated to patients undergoing more complex arch surgery or longer arrest times but clearly demonstrate the equivalence of RCP, uACP and bACP while performing an open proximal arch anastomosis for ATAAD.

CONCLUSION

The 3 investigated cerebral perfusion strategies provide similar results regarding postoperative neurologic deficits and perioperative survival. Neither strategy can be considered superior, but all of them should be part of the available armamentarium of every cardiac surgeon facing ATAAD. Retrograde cerebral perfusion combined with deep hypothermia requires a larger amount of blood products and longer operative times.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Conflict of interest: none declared.

Presented at the 35th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Barcelona, Spain, 13–16 October 2021.

Data Availability Statement

The data underlying this article are available in the article and in its online supplementary material.

Author contributions

Matteo Montagner: Conceptualization; Data curation; Formal analysis; Investigation; Writing—original draft; Writing—review & editing. Markus Kofler: Conceptualization; Data curation; Formal analysis; Methodology; Supervision; Writing—review & editing. Leonard Pitts: Data curation; Investigation; Visualization; Writing—review & editing. Roland Heck: Data curation; Investigation; Writing—review & editing. Semih Buz: Data curation; Investigation; Supervision; Writing—review & editing. Stephan Kurz: Data curation; Investigation; Supervision; Writing—review & editing. Volkmar Falk: Conceptualization; Formal analysis; Methodology; Supervision; Writing—review & editing. Jörg Kempfert: Conceptualization; Formal analysis; Methodology; Supervision; Validation; Writing—review & editing.

Reviewer information

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

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• ATAAD

  Acute type A aortic dissection

 
• ACP

  Antegrade cerebral perfusion

 
• bACP

  Bilateral antegrade cerebral perfusion

 
• CPB

  Cardiopulmonary bypass

 
• DHCA

  Deep hypothermic circulatory arrest

 
• RCP

  Retrograde cerebral perfusion

 
• uACP

  Unilateral antegrade cerebral perfusion