Technical principles of direct innominate artery revascularization: A comparison of endarterectomy and bypass grafts

Transcription

1 Technical principles of direct innominate artery revascularization: A comparison of endarterectomy and bypass grafts Kenneth J. Cherry, Jr., MD, James L. McCuUough, MD, John W. Hallett, Jr., MD, Peter C. Pairolero, MD, and Peter Gloviczki, MD, Rochester, Minn. Occlusive lesions of the innominate artery that require operation occur infrequently. Direct repair has been performed with low morbidity and mortality. There is debate over the best method of direct reconstruction. Twenty-six patients undergoing transsternal innominate artery repair over a 12-year period were reviewed to determine if either grafting or endarterectomy was superior and what technical factors might be responsible for success or failure. Most of the patients were women. Twenty-four of the patients had atherosclerotic lesions, whereas the other two had Takayasu's arteritis. Either neurologic or right upper extremity symptoms were present in 24 patients. Sixteen patients had grafting, and 10 underwent endarterectomy. There was one death. There were no strokes or transient ischemic attacks. Three patients experienced recurrence of their symptoms; all had failures of reconstruction. The use of a bifurcated graft in one patient was probably responsible for one recurrence of symptoms. Single limb grafts with added side arms are probably preferable to bifurcated grafts. Innominate artery grafting and innominate endarterectomy are equally effective, although grafting is applicable to more patients. Direct transsternal repair is the procedure of choice to correct innominate occlusive disease in patients who are good candidates for correction. (J VAsc SuRG 1989;9: ) Occlusive disease of the innominate artery that requires operation occurs infrequently. The optimal operative treatment for innominate stenoses has been the subject of some debate; several authors favor extraanatomic bypasses as the procedure of choice, even in patients who are good candidates. 1'2 However, others have performed direct innominate reconstruction with low stroke and death rates. 3-8 Carlson et al/felt that endarterectomy was the preferred direct reconstruction, whereas Crawford et al. 9 reported that their results were improved when they abandoned endarterectomy for graft reconstruction. The purpose of this article was to review all patients undergoing innominate endarterectomy or aortic-origin grafting for innominate artery occlusive disease to determine if one of these methods was superior and to determine what technical factors were involved in operative success or failure. From the Section of Vascular Surgery, Department of Surgery, Mayo Clinic and Mayo Foundation. Presented at the Twelfth Annual Meeting of the Midwestem Vascular Surgical Society, Rochester, Minn., Sept , Reprint requests: Kenneth J. Cherry, Jr., MD, Mayo Clinic, 200 First St. S.W., Rochester, MN MATERIAL AND METHODS The records of all patients undergoing direct transsternal revascularization of the innominate artery for occlusive disease at the Mayo Clinic from January 1976 through January 1988 were reviewed. Patients with aneurysmal disease of the innominate artery and those whose stenotic lesions of the great vessels did not include an innominate artery lesions were excluded from the study. Age, sex, cardiovascular risk factors, etiology, initial symptoms, arteriographic findings, type of revascularization, operative techniques, and results were examined. Follow-up information was obtained from review of charts and telephone contact with patients, families, and physicians. There were 26 patients identified. One patient was operated on in 1976; the other patients were operated on in 1980 or later. Twentyfour of the patients had atherosclerotic lesions. The other two patients, one an Asian man and the other a white woman, had Takayasu's arteritis. This latter diagnosis was established by the clinical manifestations and the pattern of involvement was confirmed on arteriography. One diagnosis was confirmed by biopsy. Fifteen patients (57.7%) were women. Ages of 718

2 Volume 9 Number 5 May 1989 Comparison of endarterectomy and bypass grafts 719 Table I. Symptoms of 26 patients with innominate artery occlusive disease Neurologic ~ Upper extremity ~ Asymptomatic No. of No. of No. of Symptoms patients Symptoms patients patients Stroke 0 TIA 4 Amaurosis fugax 6 Vertebrobasilar 8 Global 2 Total 20 (76.9%) *Ten patients had both neurologic and upper extremity symptoms. Microembolization 5 2 Claudication (53.8%) 2 (7.7%) Table II. Early results Early relief Patients Stroke / TIA Death of symptoms Endarterectomy Graft all patients ranged from 20 to 71 years, with a median of 56.5 years. The median age for men was 58 years and for women 49. Smoking was the most commonly associated risk factor and occurred in 92.3% of patients, followed by hypertension in 50.0%, coronary artery disease in 26.9%, diabetes in 23.1%, and hyperlipidemia in 19.2%. Initial signs and symptoms. Twenty patients (76.9%) had neurologic symptoms as their initial complaints. There were no strokes. Ten of the patients had symptoms referable to the anterior cerebral circulation, whereas eight had vertebrobasilar symptoms. Two patients had global ischemia. Six of the patients had amaurosis fugax of the right eye, and four had right hemispheric transient ischemic attacks (TIAs). Fourteen patients had symptoms of the right upper extremity. Ten of these also had neurologic symptoms. All four patients who had upper extremity complaints as their only manifestation ofinnominate arterial occlusive disease had evidence of microembolization to their hands, rather than claudication. None had gangrene, but one was severely ischemic with multiple ara:erial occlusions in the hand and fingers. Of the 10 patients who had both neurologic and extremity complaints, nine had upper extremity claudication rather than microembolization, reflecting their tightly stenotic rather than ulcerative lesions (Table I). Two of the patients showed no symptoms. Lesions of the innominate artery were suspected clini- cauy because of a diminished or absent pulse. One of these patients was operated on primarily for coronary artery disease and had concomitant aortoinnominate bypass grafting. The other patient showing no symptoms had a juxtarenal aortic occlusion and a 50% stenosis of her left internal carotid artery. She underwent aortoinnominate grafting before delayed infrarenal aortic reconstruction. Arteriography. All patients underwent percutaneous transfemoral arteriography. All arteriograms were reviewed by the authors. Seven of the 26 patients had disease of the innominate artery with no significant disease of either the left common carotid or left subclavian artery. However, three of those seven had concomitant carotid bifurcation or vertebral artery disease. In addition to their innominate artery lesions, the remaining 19 patients had significant disease of the other great vessels. The roentgenographic findings for atherosclerotic disease were typical of the classic descriptions offered by Carlson et al.3 Lesions most often involved the proximal third of the artery and involved the.lateral wall. Extension was usually into the subclavian artery rather than the common carotid artery. The two patients with Takayasu's arteritis had widespread disease. Operation. There were 10 innominate endarterectomies (38.5%) and 16 aortic-origin bypass graftings (61.5%). One patient undergoing irmominate endarterectomy also had aortic-origin grafting to the left common carotid and subclavian arteries. Seven

3 720 Cherry et al. Journal of VASCULAR SURGERY artery disease, experienced a fatal arrhythmia at the conclusion of her grafting operation. There were four major postoperative complications. One patient had Klebsiella pneumonia, one had Homer's syndrome that required blepharoplasty, one had a noninfected sternal dehiscence that required repeat operation, and one had a hyperperfusion syndrome with no neurologic sequelae. All patients who survived had relief of symptoms in the postoperative period or remained without symptoms (96.2%) (Table II). Fig. 1. Arteriogram of a 21-year-old white woman with Takayasu's arteritis reconstructed 16 months previously with a bifurcated graft. Note stenoses at the origins of the graft limbs. endarterectomies were closed primarily and three had Dacron patch angioplasties. Nine of the 16 grafts were single limb, three were bifurcated, and four, to multiple arteries, were single limb grafts with side arms added. The left innominate vein was divided and oversewn in four of the 26 patients. There was transient left upper extremity swelling in one of these four, which lasted less than 24 hours, There was one concomitant innominate grafting with coronary artery bypass graft, two concomitant vertebral reconstructions, and three simultaneous carotid bifurcation reconstructions. Intraoperative electroencephalographic (EEG) monitoring was used in 21 of 26 patients (80.8%). A shunt was used in only one of 26 patients. That decision was made preoperatively because of multiple artery involvement. RESULTS There were no strokes and no TIAs. There was one intraoperative death (3.8%). A 54-year-old woman with multiple episodes of right amaurosis fugax and syncope, who was known to have coronary Long-term results Follow-up data were available in 23 patients (92%). The average length of follow-up was 43 months, ranging from 7 to 148 months. There was one death that occurred 18 months after aortoinnominate grafting in a 70-year-old woman. The exact cause was unknown but was probably cardiac in origin. She had a history of coronary artery disease and arrhythmias and required placement of a pacemaker ~ preoperatively. She had had no recurrences of her earlier TIAs before her death. There were three recurrences of symptoms associated with failure of reconstruction, for a recurrence or repeat stenosis rate of 13.0%. The recurrence of symptoms and physical examination prompted repeat transfemoral arch arteriography in each of these patients. One of nine patients monitored after endarterectomy had recurrence of symptoms (11.1%), and two of 14 patients with grafts had recurrence of symptoms (14.3%). Both patients with Takayasu's arteritis had recurrence of symptoms. In one of these the origins of the limbs of a bifurcated graft placed 16 months previously stenosed (Fig. 1). At repeat operation, her graft was adherent to the sternum, and the graft limbs tightly compressed to one another by fibrous scar tissue. The mediastinum was tightly compacted with prosthetic material and scar tissue. A single limb graft to the right common carotid artery was placed and a side branch to the left common carotid sutured end to end to that graft. She has done well for 45 months. The other patient with Takayasu's arteritis had two separate aortic-origin graftings to the right common and left common carotid arteries. He came for examination again 51 months later with mild dizziness and was found to have a high-grade stenosis of his right distal anastomosis (Fig. 2). His symptoms were milder than they had been previously, and he was reluctant to undergo operation. He was treated with sodium warfarin (Coumadin). Since that time he has not been available for follow-up. The recurrence rate in the patients with athero-

4 Volume 9 Number 5 May 1989 Comparison of endarterectomy and bypass grafts 721 Table III. Late results* Long-term Patients* Death relief of symptoms Endarterectomy Graft , Two patients not available for follow-up. sclerosis was much lower at 4.8% (1 of 21). The one patient returned 26 months after innominate endarterectomy with distal stenosis of her endarterectomy site and recurrences of her disease in the distal innominate artery, right common carotid artery, and right subclavian artery (Fig. 3). She was still smoking three packs of cigarettes per day at the time of her recurrence. She underwent aortic-origin grafting to her right common carotid artery with a side limb to her right subclavian artery. She has done well for 20 months. Long-term durability of these operations was good. Of the 23 patients available for follow-up, 20 (87.0%) had either relief of their original symptoms, or remained without symptoms after their initial operations. All patients but the three with documented stenoses were relieved of their original symptoms or remained without symptoms. Eight of nine endarterectomy patients (88.9%) and 12 of 14 graft patients (85.7%) reported relief of symptoms (Table iii). DISCUSSION The infrequency with which innominate artery stenoses requiring operative repair occur makes it highly unlikely that prospective, randomized trials will ever be possible. As a consequence, knowledge of the natural history and optimal management of these patients must be based on retrospective clinical reviews. Such reports have documented excellent results for innominate endarterectomy, 3 for innominate brachiocephalic grafting, 5,8 and for combinations of the two? '6'7 Carlson et al? found only three patients out of 37 not suitable for endarterectomy and determined that it was the procedure of choice. In contrast, Crawford et al.9 believed that grafting was superior to endarterectomy. Both groups had excellent results. We were interested in whether one procedure proved superior to the other at our institution. We have concluded from the data presented that both methods of direct reconstruction give equal results and have minimal morbidity and mortality associated with them. However, unlike Wylie et al.,3 we have found bypass grafting rather than endarterectomy to Fig. 2. Arteriogram of a 45-year-old Asian man with Takayasu's arteritis shows distal right anastomotic stenosis 51 months postoperatively. be better suited for most of our patients. Patients with Takayasu's arteritis should have grafts placed, since the inflammatory involvement of the vessel wall makes endarterectomy impossible. Secondary innominate reconstructions, which are a small portion of any vascular practice, 1 also require aortoinnominate grafting. In essence, the decision to perform innominate endarterectomy is based on two considerations, as first elaborated by Carlson et al?: (1) the proximity of the left common carotid origin to the innominate artery and (2) the condition of the aortic arch. The relationship, of the innominate artery to the left common carotid artery may be judged quite well from the preoperative arteriogram. If the origins are separated by less than 1.5 cm, we favor grafting, since the space necessary for satisfactory damping and endarterectomy is not available without impinging on the left common carotid artery orifice. There are severn ramifications of this concept. A common brachiocephalic trunk mandates graft reconstruction. Although endarterectomy is possible, it unnecessarily complicates the operation with its requirement of

5 722 Cherry et al. Journal of VASCULAR SURGERY Fig. 3. A, Arteriogram of a 45-year-old white woman shows atherosderofic innominate stenosis subsequently treated by innominate endarterectomy. B, Arteriogram shows recurrent stenosis in the same patient 2 years later. Table IV. Direct innominate reconstruction No. of Perioperative Relief of Institution patients TIA or stroke (%) ' Mortality (%) symptoms (%) University of California, San Fransisco, Cleveland Clinic, Baylor, University of Michigan, Massachusetts General Hospital 29 Oregon State University, St Anthony, aortocarotid shunting and the needless possibility of left hemispheric neurologic events. If the patient has a common brachiocephalic trunk and if the initial problem is atheroembolic, grafting of the asymptomatic left common carotid artery, in addition to that of the innominate artery, is necessary to prevent lodgment of atheroembolic debris in the left common carotid distribution. Were the left common carotid artery not reconstructed, it would be the remaining patent outflow tract for this atheroembolic source. The degree of atherosclerosis or calcification of the aortic arch or both may be ascertained from the preoperative arteriogram, but it is often not fully appreciated until the time of intraoperativc examination. Calcification and noncalcific atherosclerosis increase the risk of intimal cracking, dissection, or embolic showering, thereby precluding either safe or hemostatic clamping. Consequently, calcification of the aortic arch at the base of the brachiocephalic origin mandates grafting rather than endarterectomy. Aortic-origin grafting is applicable to all patients undergoing direct innominate repair, whereas end- arterectomy can be used only in selected patients, that is, patients with atherosclerosis and suitable anatomy. We believe, as do Brewster et al., 6 that innominate endarterectomy is technically more demanding. In inexperienced hands, it is probably not a good choice, whereas it is as safe and durable as bypass grafting in patients who are properly selected when performed by surgeons trained in its use. Moreover, it has the advantage of avoiding prostheses in these relatively young people. Little has been written about operations for recurrent innominate disease. We were thus prompted to examine our repeat operations. In reviewing our three stenoses, we felt that for two of them no factor could be identified that, if changed, might have prevented the problem. The third occurred in a patient with Takayasu's disease in whom reconstruction was with a bifurcated graft. We believe that the choice of a bifurcated graft in this small woman was a poor one and that the failure of her primary graft was preventable. Crawford ct al. 9 showed years ago that straight grafts, with side limbs added as needed, per-

6 Volume 9 Number 5 May 1989 Comparfson of endarterectomy and bypass grafts 723 formed better than bifurcated grafts in this location. The use of straight grafts also aids in debulking the mediastinum. The site of our patient's stenoses were ha the graft limbs themselves rather than at anastomotic sites. For that reason, we felt the arteritis probably played no part in the advent of the stenoses. However, because the healing characteristics of grafts in patients with Takayasu's arteritis have not been well described, it may be possible that the reasons for her stenoses were multiple and other than graft compression. There may be more turbulence of flow in grafts originating from the ascending aorta in an end-to-side manner than in those originating elsewhere in the aorta or in an end-to-end manner. The short mink of this graft may have made the limb origins particularly susceptible to flow disturbances. We prefer direct innominate reconstruction to extraanatomic grafting in patients who are good candidates. Median sternotomy is a safe incision, with little inherent morbidity or mortality. It is also an incision marked by little postoperative pain or respiratory compromise. Aortic dissection, potentially a catastrophic problem, has only rarely been reported after direct innominate reconstruction, s Several series, this one included, report acceptable stroke and death rates. 3-8 One recent series from the Cleveland Clinic reported a higher mortality rate of 14.7%, but a large number of their patients had concomitant coronary or valvular heart disease. 4 Direct revascularization can be performed safely with the reasonable expectation of long-term durability. In our experience operative mortality was 3.8%, and 87.0% of patients remained without symptoms during the follow-up period. Others have reported similar findings, with predictable early relief of symptoms and long-term relief ranging from 82% to 100% (Table IV). We have drawn the following conclusions: (1) Direct repair of innominate lesions is the procedure of choice in patients who are good candidates with acceptable mortality and morbidity. (2) Innominate endarterectomy should be reserved for those patients with atherosclerosis whose anatomy permits the procedure. (3) Innominate endarterectomy and aortoinnominate grafting are equally effective in patients who are properly selected. (4) Aortoinnominate grafting is applicable to all patients who are good candidates. (5) Aortoinnominate grafting is the procedure of choice for recurrent innorninate artery disease and for Takayasu's arteritis. (6) Single limb grafts with added arms are probably preferable to bifurcation grafts. REFERENCES 1. Bentley FR, Hollier LR, Batson RC. Axilloaxillary bypass for subclavian and innorninate artery revascularization. Am Surg 1982;48: Moore WS, Malone JM, Goldstone J. Extrathoracic repair of branch occlusions of the aortic arch. Am J Surg 1976; 132: Carlson RE, Ehrenfeld WK, Stoney RJ, Wylie EJ. Innominate artery endarterectomy. A 16-year experience. Arch Surg 1977; 112: , 4. Vogt DP, Hertzer NR, O'Hara PJ, Beven EG. Brachiocephalic arterial reconstruction. Ann Surg 1982;196: Crawford ES, Stowe CL, Powers RW Jr. Occlusion of the innominate, common carotid, and subclavian arteries: longterm results of surgical treatment. Surgery 1983;94: Brewster DC, Moncure AC, Darling RC, Ambrosino lj, Abbott WM. Innominate artery lesions: problems encountered and lessons leamed. J VAsc SURG 1985;2: Zelenock GB, Cronenwett JL, Graham LM, et al. Brachiocephalic arterial occlusions and stenoses. Arch Surg 1985; 120: Evans WE, Williams TE, Hayes JP. Aortobrachiocephaiic reconstruction: Am J Surg 1988;156: Crawford ES, DeBakey ME, Morris GC Jr, Howell JF. Surgical treatment of occlusion of the innominate, common carotid, and subclavian arteries: a 10-year experience. Surgery 1969;65: Kieffer E, Petitjean C, Bahnini A. Surgery of failed brachiocephalic reconstructions. In: Bergan JJ, Yao JST, eds. Reoperative arterial surgery. New York: Grune & Stratton, 1986: DISCUSSION Dr. Thomas Schwarcz (Chicago, Ill.). I recently performed an innominate endarterectomy and, to get adequate control of the aorta and the innominate origin, I dissected a fair circumference of the aortic arch. Postoperatively the blood pressure was extremely labile for about 24 hours. Have you had this problem, and do you think that it might be caused by disturbing the aortic baroreceptor mechanism? Dr. Cherry. We have not encountered that problem, and I do not believe that the group from San Francisco has either. I do not think that my guess would be any better than yours as to what might have caused it. Dr. Richard Fowl (Cincinnati, Ohio). I want to know whether you have any suggestions or recommendations regarding the timing of surgery in patients with Takayasu's disease, since you mentioned that both of your patients with this diagnosis developed recurrences. Do you man-

7 724 Cherry et al. Journal of VASCULAR SURGERY age them with steroids, or do you have any other guidelines? Dr. Cherry. As you know, we have been very fortunate to have consultants like John Joyce at the Mayo Clinic who have been interested in this problem, but yes, we like to operate on these patients when their disease is inactive or at least under control with steroids. The first patient whom I presented required an operation during an active phase of her disease because she had such global ischemia that she could not raise either her arm or her head from the bed without fainting. The other patient was in remission and had his recurrence more than 4 years later. I do not know what we could have done to prevent either of these complications. Dr. James DeBord (Peoria, Ill.). In patients who have embolic symptoms to the right cerebral hemisphere and have both a carotid bifurcation lesion and an innominate lesion, could you comment on how you determine which is responsible for these symptoms and how you would stage the patient fo r surgical repair? Dr. Cherry. That is a good question. We tend to fix both lesions at the same time, but we try to use common sense. If the patient had a tight bifurcation stenosis assodated with a very minor innominate lesion that did not restrict flow, I think most of us would repair the bifurcation lesion to see whether that alone corrected the symptoms. If it did not, then we would proceed with innominate reconstruction. If, on the other hand, the innominate artery contained the dominant lesion or both were equally severe, we probably would fix the innominate artery and the carotid artery at the same time. Dr. Philip A. Vogt (Appleton, Wis.). I notice that you use intraoperative EEG monitoring. Is that to permit selective shunting? Do you use shunts during innominatc endarterectomy, and is it even feasible with replacement grafts? Dr. Cherry (closing). If we plan to perform an innominate endarterectomy, we apply the proximal clamp and simply observe the EEG tracing for i or 2 minutes to make certain it remains stable. If it changed dramatically, we probably would prefer to do a grafting operation because that is much easier to manage. The only EEG abnormalities in our series occurred in the middle of an innominate endarterectomy, but we thought it was better to finish the procedure expeditiously rather than to attempt to insert a shunt through an arteriotomy that extended nearly onto the aorta. Fortunately, this patient awakened with no neurologic deficits.