The heart and the aortic arch are exposed through a median sternotomy. Cardiopulmonary bypass is established from the right atrium to the right axillary artery. At 26 °C of body temperature, the supra-aortic vessels are clamped, the ascending aorta and the aortic arch are incised, and a cuffed endotracheal cannula, connected to an arterial line geared by a separate roller pump, is inserted into the descending thoracic aorta. Perfusion is started in the distal body, while the brain is perfused through the right axillary artery. Once the aortic arch has been replaced with a Dacron graft and the supra-aortic vessels have been reimplanted on the graft, the arterial line in the descending thoracic aorta is clamped and removed. The supra-aortic vessel clamps are removed, the proximal part of the Dacron graft is clamped, and systemic cardiopulmonary bypass is resumed via the right axillary artery.

From January 2002 through December 2005, this technique was used in 12 consecutive patients on an emergency basis, due to acute aortic dissection that required total arch replacement. Within the first 30 postoperative days, 1 patient (8.3%) died, and no patient had permanent neurologic deficits.

This simple technique ensures a full-flow antegrade total-body perfusion during all phases of the surgical procedure, thereby eliminating ischemia–reperfusion syndrome and yielding excellent clinical results.

In aortic arch surgery, especially in the setting of acute aortic dissection, several perfusion and cannulation techniques developed in recent years have focused attention on the organ most sensitive to ischemia: the brain. However, a high incidence of complications involving other organs and tissues, and the considerable morbidity and mortality rates that ensue, have been reported in the worldwide literature regardless of the surgical strategies adopted.

While conscious of the importance of affording the best protection to the central nervous system, we turned our attention also to the protection of the rest of the body, in order to ensure enough time to perform safe and unhurried surgery of the aortic arch.

To date, hypothermia, either alone or in association with different methods of selective brain perfusion, has been used to ensure periods of circulatory arrest. Yet circulatory arrest, quite aside from the deleterious side effects of profound hypothermia, exposes all organs and tissues to ischemia–reperfusion injuries.

The surgical technique that we propose ensures continuous, total-body antegrade perfusion, central cannulation, and full-flow cardiopulmonary bypass (CPB) under moderate systemic hypothermia.

On arrival in our emergency room, all patients underwent transesophageal echocardiography to locate the entry point and to determine the extent of the aortic dissection. Because most of the patients were operated upon at night, on an emergency basis, Doppler ultrasonography was unavailable for study of the extracranial vessels; and in no patient was preoperative cardiac catheterization or coronary arteriography performed, due to their unstable clinical condition. In 11 patients, we observed moderate-to-severe aortic valve insufficiency due to aortic valve cusp prolapse. Ten patients had a history of arterial hypertension, 3 had diabetes mellitus, and 8 were current tobacco smokers. The interval from initial symptoms to surgery varied from 6 to 21 hr (mean, 9 hr). On admission to the hospital, 3 patients were taken immediately to the operating room because of cardiogenic shock due to cardiac tamponade. For the other patients, the elapsed time from hospital admission to surgery varied from 1 to 3 hr.

After the intravenous administration of heparin, an 8-mm polytetrafluoroethylene graft (W.L. Gore & Associates, Inc.; Flagstaff, Ariz) was anastomosed end-to-side to the right axillary artery and cannulated with a 21F femoral artery cannula. The right atrium was cannulated with a standard 36F 2-stage cannula for venous drainage.

Two arterial lines (lines 1 and 2), geared by separate single-head roller pumps (Stöckert Instrumente GmbH; Munich, Germany), were set. A Y connector was inserted on arterial line 1. The 1st line (line 1a) was used for systemic perfusion through the right axillary artery; the 2nd line (line 1b) was clamped. Arterial line 2 was connected to the side branch of the right axillary artery cannula and clamped. In case of a need to perfuse the left carotid artery (1 instance, in our series), a Y-connector cardioplegia set (Edwards Lifesciences; Irvine, Calif) was inserted on arterial line 2. One branch was connected to the right axillary arterial perfusion cannula, and the other branch was connected to a 12F retrograde cardioplegia catheter (Edwards Lifesciences) for left carotid artery perfusion. The arterial lines were all made from PVC Medy 6H (Sorin Biomedica Cardio S.p.A.; Via Crescentino, Italy) and were the following sizes: lines 1, 1a, and 1b, 3/8 in × 3/32 in; and line 2, 1/4 in × 3/32 in (Fig. 1).

Fig. 1 Two arterial lines (lines 1 and 2), geared by separate roller pumps (P1 and P2), are set. A Y connector is inserted in arterial line 1: line 1a is used for systemic perfusion through the right axillary artery, while line 1b is clamped. Arterial (more ...)

Right atrium-to-right axillary CPB was instituted, a left ventricular vent was placed through the right superior pulmonary vein, and the patient was cooled. At a nasopharyngeal temperature of 26°C, the supra-aortic vessels were clamped. No ice bags were placed around the patient's head, and no other forms of cerebral protection were used. Arterial line 1a was clamped and axillary artery perfusion was started through arterial line 2. Myocardial protection was achieved by retrograde administration of cold crystalloid cardioplegic solution (St. Thomas's Hospital No. 1), delivered every 30 min in the coronary sinus: the 1st dose was 10 mL/kg of body weight; the subsequent doses were 5 mL/kg of body weight, with an infusion pressure not above 40 mmHg. Antegrade administration of the cardioplegic solution was avoided to prevent any additional lesions of the coronary ostia, which were often involved in the dissecting process. Pericardial cooling was achieved with ice-cold saline solution. After removal of the ascending aorta and the aortic arch, we chose a collagen-coated InterGard™ tubular graft (InterVascular; Montvale, NJ) of adequate size to replace the aortic arch. In this series of patients, no branched grafts were used to reimplant the supra-aortic vessels. In no instance was an elephant trunk procedure performed. The true lumen of the descending thoracic aorta was cannulated with an endotracheal cannula of 8 mm internal diameter, which was passed through the InterGard tubular graft and connected with arterial line 1b. The endotracheal cannula was tightly cuffed with saline solution to the descending thoracic aorta, and antegrade thoracic perfusion was started (Fig. 2). The opening of the aortic arch and the insertion of the endotracheal cannula into the descending thoracic aorta required a period of circulatory arrest in the distal body that ranged from 3 to 5 min, while the brain was perfused through the right axillary artery without interruption. Once the distal anastomosis of the vascular graft with the proximal descending thoracic aorta was complete and the supra-aortic arterial vessels were reimplanted on the graft, arterial line 1b was clamped and removed. The supra-aortic arterial clamps were removed, arterial line 2 was clamped, and total-body antegrade perfusion was resumed through arterial line 1a (Fig. 3).

Fig. 2 At 26°C of body temperature, the supra-aortic vessels are clamped. Arterial line 1a is clamped, and axillary artery perfusion is started through arterial line 2. Cold crystalloid cardioplegic solution (St. Thomas's No. 1) is used to protect (more ...)

Fig. 3 The aortic arch is replaced with a Dacron, collagen-impregnated tubular graft. Once the distal anastomosis is complete and the supra-aortic arterial vessels are implanted on the graft, arterial line 1b is clamped and removed. The supra-aortic artery (more ...)

During the rewarming phase, the proximal anastomosis of the vascular prosthesis to the ascending aorta was completed. In all cases, the aortic valve showed no disease. In 11 patients, the valve was resuspended to the aortic wall because of aortic valve cusp prolapse.

During CPB, the nasopharyngeal temperature was kept at 26°C, and the perfusion flow rate was maintained at a constant 50 mL · min⁻¹ · kg⁻¹. When the supra-aortic vessels were clamped, the flow rate into the axillary artery varied from 10 to 15 mL · min⁻¹ · kg⁻¹, in order to obtain a right radial artery pressure of 50 to 60 mmHg; and the remainder of the perfusate was administered into the descending thoracic aorta at a flow rate that varied from 35 to 40 mL · min⁻¹ · kg⁻¹, in order to obtain a total antegrade body perfusion at full flow during the entire surgical procedure.

Postoperatively, daily neurologic examinations were performed by a neurologist. A set of 3 neurocognitive tests was conducted by the same psychologist the day before the patients' discharge from the hospital: Raven's standard progressive matrices test, the Stroop task test, and the Rey auditory verbal learning test, to evaluate, respectively, the function of the right cerebral hemisphere and the parietal lobe, the dominant frontal lobe, and the limbic system.

During the first 6 months after discharge from the hospital, all of the surviving patients underwent follow-up transesophageal echocardiography or computed tomographic scanning.

The mean total CPB time was 188 ± 46 min (range, 110–235 min); the mean myocardial ischemia time was 113 ± 29 min (range, 63–138 min); and the mean supra-aortic vessel cross-clamp time was 60 ± 21 min (range, 33–97 min).

The only death in 30 days was that of 1 patient (8.3%), who died on the 2nd postoperative day, of low-cardiac-output syndrome. The mean hospital stay of the survivors was 12 ± 3 days. No permanent neurologic deficits were observed. One patient experienced a temporary neurologic deficit. Cognitive functions were preserved in all cases. Postoperatively, no cases of pulmonary insufficiency, infection, or ARDS were observed. The Lung Injury Score ranged from 0 to 2 (mean, 1.1), and the median intubation time was 13 hr (range, 6–9 hr). Hourly urine output was ≥1 mL/kg in all patients during their intensive care unit stay. The preoperative mean ± SD BUN and serum creatinine levels were 38.7 ± 9.5 and 0.85 ± 0.17 mg/dL, respectively. Postoperative levels of those same values were 56.7 ± 10.9 and 1.11 ± 0.35 mg/dL; differences were statistically significant only for the BUN values (Table I). Postoperative AST, ALT, and LDH values all increased without reaching statistical significance (Table I). The total mean blood loss from drainage was 900 ± 175 mL.

TABLE I. Pre- and Postoperative Measurements of Renal and Hepatic Function

During the first 6 months of follow-up, we observed no additional intimal tears at the level of the insertion of the distal perfusion cannula into the proximal descending aorta.

In aortic arch surgery, selection of the arterial cannulation site for CPB is of crucial importance. Our analysis of the literature showed that permanent neurologic damage results most often from strokes due to emboli or malperfusion and is not dependent upon the method of cerebral protection used during surgery.^2,3,6–9 The femoral artery has long been the preferred site of arterial cannulation. It is easy and fast to access, and it ensures a safe backup while the chest is opened in hemodynamically unstable patients. Major complications of femoral artery cannulation result from reversing the blood flow, which can cause embolic showers to the brain, the kidneys, and other organs. At the beginning of perfusion, when the body is warm, the effects of emboli are particularly dangerous.² Moreover, acute aortic dissection can redirect the blood flow into the false lumen through a distal re-entry site,³ thereby causing malperfusion of various organs. Antegrade perfusion during CPB offers the opportunity to reduce the incidence of these possible complications.

At present, the right axillary artery seems to be the preferred site for arterial cannulation in aortic arch surgery.^4,5 This artery is rarely involved in the dissection,^4,5 is rarely affected by atherosclerotic plaques, and is of adequate size to carry full-flow CPB in patients who have a large body-surface area. The brain is perfused with blood flow that has not passed through the often-atherosclerotic ascending aorta or arch. Cannulation of the right axillary artery ensures continuous antegrade cerebral perfusion, avoids manipulation of the supra-aortic vessels (which are often atherosclerotic or dissected), leaves a free operative field, and eliminates the need to change the cannulation site when the arch is replaced. The right axillary artery can be cannulated directly or by interposing a vascular graft. We prefer interposition because it affords the possibility of monitoring right radial artery pressure.

Unilateral selective cerebral perfusion has been criticized due to the risk of cerebral hyperperfusion or of inadequate perfusion of the contralateral hemisphere. On the basis of experimental data, Tanaka and colleagues¹¹ recommended a minimum mean carotid arterial pressure of 39.8 ± 6.2 mmHg at a flow rate of 50% of the physiologic level for safe, unilateral, selective cerebral perfusion at moderate hypothermia (25°C). These conclusions have been confirmed by several clinical reports.^7,12,13 Using Tanaka's studies as a guide, we conducted selective unilateral cerebral perfusion at a flow rate of 10 to 15 mL · min⁻¹ · kg⁻¹ in order to obtain a right radial artery mean pressure of 50 to 60 mmHg. We think that these higher values of mean arterial pressure in the right radial artery have to be reached because of the distal hyperperfusion provoked by axillary artery cannulation via the graft interposition technique. However, in no instance did the distal hyperperfusion provoke an evident clinical sequel in the right arms of our patients. In our series, cerebral hyperperfusion was avoided; we controlled hypertension by lowering the flow rate or by unclamping the origin of the left subclavian artery if the returning blood did not disturb the operative field. Postoperatively, no high-resistance complications such as hemorrhagic strokes were observed in our patients. The efficacy of contralateral hemispheric perfusion through the circle of Willis in this series of patients was evaluated intraoperatively, at the initiation of unilateral cerebral perfusion via the right axillary artery, by observing the amount of blood that returned from the left carotid artery before clamping it. The amount of blood that flowed down from the left carotid artery was judged satisfactory in 11 patients; in 1 patient, we selectively cannulated this artery to counter the risk of left-hemisphere hypoperfusion. Preoperative evaluation of the circle of Willis or more sophisticated monitoring of contralateral hemispheric perfusion during surgery could be used in elective surgical cases.

To ensure a bloodless operative field during surgery on the aortic arch, deep hypothermic circulatory arrest (DHCA) has been used alone or in association with various methods of retrograde or antegrade cerebral perfusion. Although DHCA is the simplest surgical technique, several disadvantages are related to it: suboptimal cerebral protection is reported after only 25 min of DHCA,⁶ with decreased survival after 60 min.⁷ Other authors have found that DHCA is an independent risk factor for cerebral injuries.⁹

Better results have been reported with antegrade cerebral perfusion conducted in association with periods of hypothermic circulatory arrest^2,3,9 at systemic temperatures of up to 25°C.^10,14,16 The rationale for raising the body temperature during periods of circulatory arrest is to reduce the risks of coagulopathy and pulmonary insufficiency and to shorten the time of CPB. But maintaining circulatory arrest at moderate systemic hypothermia raises 2 questions: Can the spinal cord tolerate the prolonged periods of ischemia? What happens to the rest of the body when CPB is resumed? It is well known that normothermic CPB in itself can trigger a systemic inflammatory response syndrome that ends in diffuse endothelial cell injury and parenchymal dysfunction. Ischemia–reperfusion is a potent stimulator of endothelial-cell apoptosis and can induce multisystem organ dysfunction. Cooper and associates¹⁵ showed experimentally that endothelial cells malfunction in large-caliber pulmonary veins and renal arteries in pigs undergoing DHCA, which causes renal insufficiency and ARDS. These data are confirmed in the clinical setting by the high incidence of pulmonary, renal, and mesenteric insufficiency reported in the literature in patients who undergo periods of circulatory arrest.^17–19 In 2003, Allen and coworkers²⁰ proposed avoiding the “all-body reperfusion syndrome” by modifying the composition of the reperfusate.

The goal of our cannulation-and-perfusion technique is to reduce at minimum the period of circulatory arrest, while providing a dry operative field at the level of the aortic arch. Cannulation of the right axillary artery and the descending thoracic aorta provides total antegrade body perfusion at full flow in mild hypothermia during the entire surgical procedure. Perfusing the brain at a temperature well above 20°C at a pressure of 50 to 70 mmHg preserves the autoregulation mechanism of cerebral blood flow.^21,22 In the present series of patients, the mean time of unilateral cerebral perfusion was always longer than 30 min; of particular importance, 1 patient reached 97 min without postoperative signs of cerebral hyperperfusion (edema) or hypoperfusion (stroke). Should right axillary artery cannulation not be sufficient to ensure satisfactory contralateral cerebral perfusion, as shown by scanty intraoperative blood flow from the unclamped left common carotid artery, the latter artery can easily be cannulated from inside the opened aorta. In such cases, we think that mild systemic hypothermia can be safely used.

Once the true lumen of the descending thoracic aorta has been cannulated with a cuffed cannula, full-flow antegrade perfusion of the thoracoabdominal organs, as happens in the usual CPB, is guaranteed. In our series of patients, pulmonary function was preserved in all cases, as shown by the short postoperative intubation times, the absence of episodes of pulmonary insufficiency or infection, and the good Lung Injury Scores upon serial chest radiography. Probably, these results are due to continuous perfusion of the lungs through the bronchial collateral circulation. The postoperative renal and hepatic functional tests showed values higher than the preoperative ones, but only the increase in BUN values reached statistical significance. Moreover, increases in these values are normally observed in patients who undergo CPB for routine cardiac surgical procedures.^23,24

The aim of this technique is to avoid any period of circulatory arrest during aortic arch surgery, especially in the setting of acute aortic dissection. Ensuring antegrade perfusion at full flow for the entire body minimizes the systemic inflammatory syndrome and eliminates the ischemia–reperfusion syndrome. The presence of a cuffed perfusion cannula in the proximal descending aorta does not affect the execution of the distal anastomosis. Because the cuff of the endotracheal cannula is under low pressure and has a large area, inflation of the cuff to eliminate blood backflow does not provoke further trauma to the freshly dissected aorta and ensures a completely dry operative field.

The major limitations of this study are the small series of patients and the absence of comparative studies of other perfusion techniques, but the excellent clinical results so far obtained appear to confirm the validity of our technique. We believe that use of this perfusion technique will enable safe increases in the systemic temperature to 28 to 30°C during CPB.