Cross-Clamp Technique and Incidence of Stroke after Coronary Artery Bypass Grafting: Know and Respect the Aorta Principle

Giorgio Mastroiacovo (  gio.mastroiacovo@hotmail.it ) Department of Cardiovascular Surgery, IRCCS Centro Cardiologico Monzino, Milan, Italy Marco Gennari Department of Cardiovascular Surgery, IRCCS Centro Cardiologico Monzino, Milan, Italy Ilaria Giambuzzi Department of Cardiovascular Surgery, IRCCS Centro Cardiologico Monzino, Milan, Italy Sergio Pirola Department of Cardiovascular Surgery, IRCCS Centro Cardiologico Monzino, Milan, Italy Giuseppe Nanci Department of Cardiovascular Surgery, IRCCS Centro Cardiologico Monzino, Milan, Italy Alice Bonomi Department of Statistics, IRCCS Centro Cardiologico Monzino Milan, Italy Gabriele Egidy Assenza Departement of Pediatric Cardiology and Adult Congenital Heart Disease, Sant’Orsola Malpighi Hospital, Bologna, Italy Gianluca Polvani Department of Biomedical, Surgical and Dental Sciences, University of Milan, Italy Marco Agrifoglio Department of Biomedical, Surgical and Dental Sciences, University of Milan, Italy

Since the occurrence of neurological events dramatically impacts the prognosis and quality of life of the patients, any effort must be pursuit to avoid iatrogenic stroke following CABG. Atherosclerotic disease of the ascending aorta has been considered one of the most probable cause of cerebral embolization due to dislodgement of plaque debris during aortic manipulation and cross-clamping [5][6][7].
Among the strategies to decrease or eliminate aortic manipulation, off-pump CABG through an aortic "no touch" technique has been advocated as one of the most useful strategy in reducing the stroke rate [8,9], even though it might lead to worse graft patency and survival [10], especially in low-volume Centers [11,12].
However, this off-pump aortic "no touch" technique is not universally applicable and, when saphenous vein and/or free arterial aorto-coronary grafts are used, there is still risk of neurological injury due to tangential aortic clamp applied during the proximal anastomoses sewing.
To minimize aortic manipulation and trauma in standard on-pump CABG, a single aortic clamp (SAC) rather than a double aortic clamp (DAC) has also been proposed, to reduce the manipulation and stretching of the aorta [13][14][15].
The conclusions of these reports have been con icting, although in several papers SAC technique showed superiority in reducing neurologic injury following CABG causing less neuropsychological de cits and release of serum S-100 protein, a surrogate marker of cerebral injury [16][17][18].
On the other hand, SAC prolongs cardiopulmonary bypass (CPB) time, which is also per se considered an independent risk factor for cerebrovascular accidents. Indeed, other papers have reported no bene t of SAC over DAC technique in preventing the neurological lesions [19][20][21], suggesting that other factors might be accountable for stroke, such as cannulation [22](cannulation strategy, type of cannula) and even the aortic punch [23].
Given this background, we retrospectively compared the early and long-term neurological outcomes between two homogeneous CABG groups treated with either SAC or DAC technique.

Patients
From March 2006 to December 2012, 251 consecutive patients underwent myocardial revascularization on CPB through a complete median sternotomy performed by a single CABG-experienced surgeon (MA) using SAC or DAC technique. Routine pre-surgical evaluation for all patients included electrocardiogram, echocardiogram, carotid ultrasound and coronary angiography.

Ethical committee
This work is based on a retrospective chart review with follow-up information retrieved by telephone calls. All methods were carried out in accordance with relevant guidelines and regulations. Patient's consent was obtained for research purposes. The informed consent and all experimental protocols were approved by the Cardiologico Monzino's Ethics Committee (CCM 1424), December 2020.

Inclusion criteria
Inclusion criteria were isolated CABG with no other concomitant cardiac procedures, CPB support, a preoperative left ventricular ejection fraction (EF) more than 50% and two or more distal coronary anastomoses along with at least one proximal anastomosis on the ascending aorta.

Exclusion criteria
Exclusion criteria were patients with porcelain aorta or left ventricular mural thrombi identi ed by preoperative trans-thoracic/trans-oesophageal echocardiogram, angiogram and intraoperative records of atheromatous aorta (extensive calci cations, focal atheromas) and the presence of clinically relevant internal carotid artery atherosclerosis (stenosis ≥70% of luminal diameter). Given the exclusion criteria, the analysed patient population is to be considered at low to moderate risk for neurological events.

Surgical procedure
Anaesthetic protocol included thiopentone, propofol, opiates for induction, sevo urane, des urane relaxants, opiates for maintenance and propofol, relaxants and opiates for CPB time. Intraoperative monitoring included radial artery for pressure monitoring and right jugular vein for central venous catheter. Swan-Ganz catheter was not considered as needed as no patient had EF less than 50%. Urinary catheter and rectal temperature probe were also inserted. All interventions were performed on-pump, 99% of those through median sternotomy with ascending aorta and right atrial cannulation for CPB implantation, after reaching an activating clotting time more than 480 seconds under systemic heparinization. Only 3 patients (1%) underwent peripheral femoral cannulation for unstable hemodynamic conditions. In all patients a retrograde cardioplegia catheter was inserted through right atrium directly into coronary sinus. Myocardial protection was obtained using cold hematic blood cardioplegia (Buckberg) administered rstly anterograde and then retrograde every nearly 20 minutes. Patient was cooled to 32-34°C during the main time of the operation. The rewarming was achieved with tepid reperfusion and wash-out of the cardioplegia.
The choice of performing a SAC or a DAC was made by the surgeon, depending on the pre-operative imaging tests, on the visual and manipulation appearance of the aorta and on epiaortic ultrasound performed in the operating room. When it appeared to be brotic and frail based on the rst operator experience, the surgeon preferred a SAC approach.
In case of SAC, the proximal anastomosis was performed during cardiac arrest.
If the surgeon decided to perform DAC, all the grafts were linked to the arterial cannula, so that the coronaries could be perfused immediately after rewarming. After choosing the site of the proximal anastomoses, a Satinsky tangential clamp was placed to allow safe suturing. Subsequently, one by one, the proximal anastomoses were performed.
Total heparinization was antagonized with protamine sulphate after CPB weaning.

Data Collection
Pre-operative and intra-operative data were collected from clinical records. The main variable investigated was stroke, de ned as an acute focal neurological event with a positive magnetic resonance imaging, regardless of duration of clinical symptoms, or a persistent focal neurological de cit lasting longer than 24 hours. Diagnosis of stroke was determined by our cardiac surgical team (at least one cardiac surgeon and one cardiac anesthesiologist) and con rmed by a neurologist in all cases; patients with stroke underwent brain computed tomography and magnetic resonance and the presence of stroke was veri ed in all patients.
Patients were subsequently followed in our outpatient clinic (67%) or through phone calls (33%). New neurological and cardiac events (recurrence of angina, acute myocardial infarction, new revascularization procedures, whether percutaneous or surgical) as well as survival data were collected.

Data analysis
Between-group comparisons for clinical and outcome variables were performed using independent samples t-test, Wilcoxon rank sum test, chi-square analysis, or Fisher's exact test. Data are reported as mean ± standard deviation, median ( rst and third quartile), or frequency (%). All tests are two-sided. Given the exploratory nature of the study, no adjustment for multiple comparisons was performed. Data were compared using the aortic clamping technique (DAC versus SAC as the dichotomizing variable. Analyses were performed using StataV.12 (StataCorp, Texas, US).

Study Results
Patient characteristics are listed in Table 1. Surgical details are listed in Table 2. The grafts used to perform CABG were left internal mammary artery (LIMA) in situ, right internal mammary artery (RIMA) in situ, great saphenous vein (GSV) and left or right arterial artery when GVS was unavailable, always anastomosed to the ascending aorta proximally. Coronary arteries were variably grafted according to the disease as described in Table 2, with no signi cant difference between the two groups. All distal anastomosis were performed during cardiac arrest (cross-clamping). No difference between the two group was found regarding the number of distal anastomosis, performed either with an arterial (p value 0.71) or a venous graft (p value 0.071). When feasible, a sequential con guration was chosen, performing no more than two bypasses with the same graft.
In the SAC group all the proximal anastomoses were performed during the period of total aortic cross clamping (cardiac arrest). In the DAC group the proximal anastomoses were performed on a beating heart with the aid of a side clamp. We did not found any statistically signi cant differences in blood loss (median 600 cc IQR 450-750 vs 630 cc IQR 450-850) and blood product use (55 patients − 41%-vs 53 patients − 45%-) within the two groups (DAC vs SAC respectively). Surgical re-exploration for excessive post-operative bleeding was required in 5 patients for each group (4% of the entire cohort). Moreover, no difference in total hours of post-operative invasive ventilation (median of 5 hours IQR 4-7 for each group) neither in total hours of intensive care unit (ICU) stay (mean of 49 ± 27 hours) were found. Hospital stay was also comparable with a median of 7 days (IQR 7-9) of total length.
No relevant differences were found in the analysis of peri-procedural complications including the study primary outcome of stroke and transient ischemic attack (Table 3). The all-cause mortality Kaplan-Meier estimates show 81% and 73% of survival, for DAC and SAC respectively, at a median follow-up of 11 years (Fig. 1.A).
For the DAC group, 4 patients (3,4%) experienced recurrent myocardial angina and 2 patients (2%) showed acute myocardial infarction, no patient experienced stroke at follow up. Freedom from MACCE de ned as myocardial infarction, stroke and cardiac death, was 100%, 100%, 98% and 82% at 1,5, 10 and 14 years respectively (Fig. 1B). The difference between SAC and DAC is statistically signi cant (p value = 0.02). The two groups have similar incidence of MACCE in the rst 10 years ( Therefore, at the chi-squared analysis, after normalizing for presence of diabetes and peripheral arterial diseases, being part of the SAC group was a risk factor for having MACCE at follow up (p value = 0.035, HR 5,46 95% CI 1.14-26.1).

Discussion
We presented a retrospective analysis of the differences in neurological outcomes related to the clamping approach (single vs double SAC and DAC respectively) of the aorta when performing CABG in a low to moderate risk population. A single, experienced surgeon performed all the procedures. A total of 5 patients (2%) had stroke with no signi cant clinical sequalae, probably because of the limited interested cerebral anatomical area. We could not be able to nd any signi cant difference in the stroke rate between the SAC and DAC group.
Classically, many risks factors have been linked to stroke following CABG. First, the use of CPB has been advocated to be related with a certain risk of neurocognitive sequelae linked with in ammatory response and microembolism [4][5][6][7][8][9]24].
In our series, all patients underwent on-pump CABG, even if in the literature there is still an on-going debate between the usefulness of on-pump versus off-pump coronary artery bypass (OPCAB) surgery.
Prospective studies like ROOBY trial showed better results in term of myocardial infarction (lower rate in on-pump CABG) and better rate of venous graft patency at 1 year, along with better Fitz-Gibbon grade, for on-pump CABG [25]. While other studies such as the SMART trial [26] highlighted no signi cant differences between the two techniques in term of mortality, myocardial infarction, stroke and recurrent angina or readmission for cardiac or non-cardiac events. Regarding neurological outcomes, The ROOBY trial [25] showed a better scoring in the clock-drawing test in the off-pump group.
Therefore, there must be other variables linked to peri-operative stroke following CABG. Indeed, cannulation itself (including type of cannula), cross-clamping and, more widely, aortic manipulation, are showed suspected to be linked with neurocognitive impairment [12][13][14].
The debate regarding the best strategy for proximal anastomosis (SAC vs DAC) is still on-going. Chu D. et al. [27] performed a retrospective study on 1819 patients without nding any difference in terms of neurological events between SAC and DAC patients. Araque JC et al. [20] had the same results on a bigger cohort of patients and, moreover, at the univariate and multivariate analysis the degree of aortic manipulation was not associated with post-operative stroke. Their results were con rmed by a large, retrospective study on patients recruited from the Society of Thoracic Surgery database; no differences between SAC and DAC were found regarding neurological events and mortality [28].
On the other hand, several authors have showed con icting results. Remarkably, Hammon et al. [18] prospectively analyzed 237 high risk patients undergoing OPCABG vs. CABG with single vs multiple clamp technique and failed to nd an improve outcome in term of neuropsychological de cit in the OPCAB group. More frequent neuro-psychological de cits were found in the DAC group. Instead, patients undergoing CABG with single clamping had better outcome, suggesting that the mild hypothermia during CPB may allow a neuro-protective environment [29].
In our cohort none of the techniques used showed to be superior in term of stroke incidence over the other. As each technique has its own surgical advantages and disadvantages (for instance, more space to perform proximal anastomosis in SAC and direct evaluation of graft's length in DAC) the surgeon should choose the technique most appropriate to a speci c situation, given the fact that good clinical outcomes and low complication rates could be achieved with both techniques. However, according to the existing evidences, the presence of aortic atheromas, brosis or frail tissue push the operator to reduce aortic manipulation to reduce the risk of local or distal problems, and thus a SAC strategy could be preferred over DAC. OPCABG and use of double mammary graft or Y con guration could be advocated in the case of porcelain aorta to avoid excessive aortic manipulation.
On this regard an interesting aspect, as reported by Hammon et al. [18], is that neuropsychological de cits, even if absent and even not radiologically detectable early after operation, can appear over a period of six months from the index procedure, suggesting that a closer neurological follow-up should be taken into consideration in high-risk patients to better estimate the real neurological outcome .
Our study offers a very long follow-up, reaching up to 14 years. In the SAC group, 1 (0.9%) patient had stroke during follow up, while in the DAC group nobody had neurological events. Therefore, it is realistic to assume that most of the strokes after CABG happen in the early post-operative period, and that the risks related to aortic manipulation decrease in the long-term period.
Finally, it is interesting to notice that after 10 years the SAC group had a signi cant lower freedom from MACCE, often related to the progression of the vasculopathy. Indeed, patients in the SAC group had a 5times higher risk of MACCE than patients in the DAC group during long-term follow-up.
This result further underlines that an accurate analysis of the aorta at the time of surgery may reveal many pathological changes that put the patients at risk, immediately and during follow-up. Therefore, besides pre-operative imaging exams and anamnesis, a in-depth analysis of the features and consistency of the aorta at the time of surgery is crucial to decide if a DAC or a SAC is the best strategy. It is possible to assume that our patients in the SAC group had a worse vascular tissue already at time of CABG, and this would explain the higher incidence of MACCE at follow up. Hence, patients who have a worse aortic tissue should be kept in a strict long-term follow up (up to 10 years) for MACCE.

Study limitation
The main limitation of this paper is related on its retrospective design analysis on a single-operator outcome. The main bias is the analysis of pre-operatively judged low neurological risk patients, thus our conclusion could not be applicable to other subset of CABG patients. Moreover, the small sample size of patients who were able to perform a follow up is another limitation. Finally, a last bias of our study is the assessment of the single surgeon conducted on the aorta of the patients that led him to choose between SAC and DAC.

Conclusion
In conclusion, incidence of stroke seems to be independent from cross-clamping technique in a nonseverely calci ed aorta with low to moderate carotid atherosclerotic burden.
More generally, we could infer that the global rate of stroke after CABG is probably more in uenced by other factors (such as, for example, plaques extent and composition).
We do believe in the principle of knowing and respecting the aorta: a careful pre-operative and intraprocedural analysis of the aorta and of the supra-aortic vessels is the best strategy to personalize the surgical therapy to achieve optimal general and neurological outcomes.