|Home | About | Journals | Submit | Contact Us | Français|
Since the adoption of off-pump coronary artery bypass surgery (OPCAB), numerous investigators have compared its short- and long-term results with those of on-pump coronary bypass surgery. Some reports of OPCAB were quite favorable, whereas others were critical, claiming that it resulted in incomplete revascularization and reduced venous graft patency. A potentially serious complication of OPCAB, not heretofore sufficiently confronted, is the increased incidence of early postoperative acute aortic dissection, in comparison with the more familiar intraoperative and late-occurring aortic dissection after conventional on-pump bypass surgery.
Early postoperative acute aortic dissection after OPCAB appears to be more frequent than was initially thought. Its clinical manifestations can be unusual and often neurologic in nature—rendering diagnosis difficult, causing delays in surgical intervention, and resulting in a high mortality rate.
When the physician notes unusual developments in patients after OPCAB that lead to the suspicion of aortic dissection, immediate computed tomography of the chest and surgery should occur if dissection is confirmed. If not detected early, this sequela almost certainly leads to rapid death from aortic rupture. Prevention lies in the strict control of systolic blood pressure during the performance of proximal anastomoses; avoidance of aortic clamping through the use of sequential all-arterial grafts or new-generation mechanical connectors; and, at times, aggressive replacement of the aorta with a prosthetic graft.
Herein, we present the cases of 4 patients who sustained acute aortic dissection early after OPCAB. We review the pertinent medical literature.
Aortic dissection is known to occur spontaneously, as in severe hypertension and Marfan syndrome, or it may be iatrogenic and follow a cardiac operation at any time and with no typical interval. The medical literature contains many reports of iatrogenic aortic dissections that occurred intraoperatively and that were repaired successfully.1–3 Similarly, late dissection occurring many years after a successful cardiac operation is well known.4 Conversely, little attention has been given to postoperative acute aortic dissection that occurs within hours, days, or weeks after surgery. Here, we present our experience with this complication, call attention to its potentially fatal outcome, review causal factors and clinical manifestations that can lead to its prompt recognition, and compare its frequency in off-pump coronary artery bypass surgery (OPCAB) and on-pump coronary artery bypass surgery (on-pump CAB).
In August 2002, 4 days after sustaining a myocardial infarction, a 74-year-old hypertensive man underwent quadruple OPCAB, with grafts of the left internal mammary artery (LIMA) to the left anterior descending coronary artery (LAD) and a diagonal branch sequentially, and reversed saphenous vein bypass grafts to the right coronary artery (RCA) and to a marginal branch of the circumflex coronary artery (Cx). The aorta was normal in size and texture. With use of an atraumatic partial aortic clamp, we maintained the patient's systolic blood pressure at 80 mmHg during the 2 proximal aortovenous anastomoses. On the 7th postoperative day, while preparing for his discharge from the hospital, the patient developed sudden paraparesis of the lower extremities and had to return to bed. His electrocardiogram remained unchanged, and his systolic blood pressure was stable at 110 mmHg. Bedside transthoracic echocardiography ruled out cardiac tamponade. A computed tomographic (CT) scan of the chest with intravenous contrast medium showed a dissection flap that involved the ascending aorta, aortic arch, and descending thoracic aorta, with contrast medium filling the true and false lumina—findings that were diagnostic of type A aortic dissection. During surgery, the dissection was found to have originated at the proximal anastomosis of the venous graft to the RCA. An ellipse of aortic wall that encompassed both proximal anastomoses was excised, the surrounding dissected layers of aortic wall were glued together with BioGlue® (CryoLife, Inc.; Kennesaw, Ga), and the defect was replaced with an ellipse of Hemashield® graft material (Boston Scientific Corporation; Natick, Mass). Both venous grafts were then reattached to the Hemashield graft. The patient's recovery was uneventful. Six years later, follow-up CT showed a clotted false lumen in the descending aorta and arch, and a normal aortic diameter throughout. His blood pressure was well controlled by means of medical therapy. He remained asymptomatic as of August 2009.
In May 2008, a 61-year-old hypertensive, hyperlipidemic, obese man (body mass index, 44) underwent triple OPCAB for unstable angina and severe shortness of breath precipitated by exercise. The surgery consisted of a sequential LIMA graft to a large ramus branch of the Cx and its marginal branch, and a venous graft to a large posterior descending artery (PDA) branch of the RCA. During surgery, the ascending aorta was found to be normal. A soft-jaw partial-occluding clamp was applied, and the patient's systolic blood pressure was kept at 80 mmHg during the proximal aortovenous anastomosis. On the 3rd postoperative day, he experienced syncope and was returned to bed. For a brief time, his speech was slurred, but it quickly returned to normal. Examination of his vital signs and a brief neurologic investigation revealed no unusual results. He was sent for a CT scan of the brain. Upon its completion and while he was still on the examination table, he reported back pain, and a previously palpable right radial pulse was lost. Immediate CT of the chest showed an extensive type A aortic dissection with occlusion of the left common carotid artery and a considerable pericardial effusion (Fig. 1). While the operating room was being prepared, he developed signs of cardiac tamponade and was re-explored immediately; however, a full-thickness rupture of the dissected ascending aorta into the pericardial cavity precluded further surgical intervention. The dissection appeared to originate at the proximal anastomosis of the venous graft to the aorta. The elapsed time between the syncopal episode and death was 62 minutes. Subsequent review of the CT of the brain showed multiple infarcts in the left hemisphere and a single infarct in the right hemisphere, all of which appeared to be chronic to the interpreting radiologist.
On 11 July 2008, an 81-year-old man with a history of recurrent syncope, hypertension, chronic obstructive pulmonary disease (COPD) secondary to tobacco use, and gastrectomy for peptic ulcer disease underwent cardiac catheterization after a negative electrophysiologic evaluation. Due to marked tortuosity and calcification of the RCA, a critical distal lesion could not be crossed with a guidewire despite repeated attempts. The next day, he underwent double OPCAB with a reversed saphenous vein graft to the PDA and a free right internal mammary artery (RIMA) graft to the marginal branch of the Cx. His aorta was soft, pliable, and of normal size. His systolic blood pressure was kept at less than 80 mmHg during the 2 proximal anastomoses. The patient's postoperative course was remarkably uneventful considering his comorbidities. On postoperative day 6, he was discharged to a skilled-nursing facility for rehabilitation and physical therapy. He returned to our office a week later for follow-up and was found to be recovering very well, with normal heart rate and blood pressure. He was then discharged from the nursing facility. On 31 July, he came to the emergency department upon the recommendation of a visiting nurse, who had measured his oxygen saturation at 85% on room air. The patient was feeling well and reported no symptoms. Upon examination, his oxygen saturation was 92%. Results of a chest radiograph, an electrocardiogram, and a full blood and chemistry analysis were all within acceptable ranges. The patient was to be discharged from the emergency department when it was suggested that pulmonary embolism be ruled out. A CT scan of the chest with contrast medium was negative for pulmonary embolism, and it was interpreted to be negative for aortic dissection. However, there was a 2.3-cm collection of contrast extravasation in the area of the aorta–free RIMA anastomosis medially, which was thought to be a contained leak or a pseudoaneurysm (Fig. 2). Selective aortography in the cardiac catheterization laboratory showed patent grafts and no active extravasation of contrast medium from either proximal anastomosis. On the basis of our experience with our 2 earlier patients, we decided to explore the chest in order to rule out a dissection, or to repair a pseudoaneurysm. During surgery, the ascending aorta appeared normal and not dissected. An organized, medium-sized hematoma obscured the origin of the RIMA graft medially. Cardiopulmonary bypass was instituted, and the clot was moved away from the graft, exposing a normal anastomotic site. The aorta was then cross-clamped and incised, exposing a contained dissection that started at the origin of the aorta–free RIMA anastomosis and extended to the base of the aortic arch (Fig. 3). The ascending aorta was replaced with a prosthetic tube, and both grafts were reimplanted on it. Histologic examination of the resected aorta revealed inflammation, giant cells, and thrombi. The patient's prolonged postoperative course was complicated by the need for a tracheostomy and intravenous hyperalimentation. On September 14, the patient died of respiratory arrest in a rehabilitation unit.
In April 2008, an 86-year-old woman with critical, symptomatic, left main coronary artery stenosis and hypertension underwent quadruple OPCAB. The anastomoses involved a reversed saphenous vein graft to the PDA, sequential vein grafts to the diagonal branch of the LAD and the obtuse marginal branch of the Cx, and the LIMA to the LAD. Both proximal anastomoses of the venous grafts to the grossly normal aorta were performed with use of a side-biting clamp that was applied while the patient's systolic blood pressure was kept at less than 80 mmHg. She later required left thoracentesis, and she was transferred to a skilled-nursing facility 10 days postoperatively for rehabilitation. In early December 2008, she reported to her family physician a gnawing right subscapular pain that subsided upon the administration of simple analgesics. A CT scan of the chest with intravenous contrast medium showed dilation of the ascending aorta to 4.2 cm, with a type A dissection that extended to the aortic arch but spared the great vessels and the thoracic aorta. Because of her advanced age and her lack of compelling symptoms, we elected to treat her condition conservatively as a chronic dissection, in the hope that she would be protected from rupture by periaortic scarring that was secondary to her earlier heart surgery. A repeat CT scan in May 2009 showed a stable appearance of the aorta. The patient remained asymptomatic as of August 2009.
In on-pump CAB surgery, intraoperative aortic dissection is generally a more common event than is delayed-but-early acute postoperative aortic dissection. In 8,624 patients who underwent on-pump CAB, Ruchat and colleagues1 encountered 3 postoperative dissections (occurrence range, 8–32 days after initial surgery), and 7 intraoperative. In 6,943 cases of cardiac surgery, Murphy and associates2 reported 9 dissections that occurred from 30 min to 21 days postoperatively and 15 that presented intraoperatively. Still and coworkers3 observed 4 early and 20 intraoperative dissections in 14,877 surgical patients. Chavanon and co-authors5 reported only 1 intraoperative dissection in 2,723 patients and none that occurred later. In the instances of intraoperative aortic dissection, the site of intimal tearing was generally at the site of aortic cannulation or cross-clamping. In regard to the early postoperative dissections, the intimal tear usually originated at the aortic clamp site or at the proximal anastomosis on the ascending aorta. Considering the frequency of aortic dissection after on-pump CAB, Hagl and Griepp6 speculated that the use of a partial-occlusion clamp on the pulsating, sometimes diseased aorta could increase the risk of iatrogenic aortic dissection during OPCAB.
In 2001, a Montreal Heart Institute team5 published its initial experience with OPCAB surgery in 308 patients and compared it with the results in 2,723 who underwent on-pump CAB. There were 2 instances of early postoperative aortic dissection and 1 of intraoperative dissection in the OPCAB group, and only 1 intraoperative dissection in the on-pump CAB group. This report prompted other early OPCAB proponents5,7–13 to publish their experience with this complication (Table I). The mortality rate in unrecognized early dissection after OPCAB was almost 100%, due to cardiac tamponade and rupture of the aorta in the absence of established postcardiotomy scarring. The mortality rate in recognized and surgically treated cases was 60%, far exceeding the current mortality rate of approximately 20% in the repair of spontaneous acute aortic dissection. Furthermore, in the absence of autopsies, it can be surmised that some sudden deaths after OPCAB—such as 8 of 3,866 patients studied by Buffolo and colleagues,13 and 1 among 300 patients of Demers and Cartier7—were also secondary to rupture of unrecognized early postoperative dissection and not to the terminal catastrophic events that are usually invoked (fatal ventricular arrhythmias or massive pulmonary embolism). In a discussion of Chavanon and associates' publication,5 Hagl and Griepp6 stated that the total number of iatrogenic aortic dissections is likely underestimated because some occur a few days or weeks postoperatively. Hagl and co-investigators14 estimated that early postoperative aortic dissection underlies 3% to 5% of deaths after cardiac surgery.
Intraoperative aortic dissection is easily recognized due to the presence of a rapidly expanding hematoma. The dissection can be repaired in fashion similar to acute spontaneous aortic dissection, although with a much higher mortality rate.2 Conversely, early dissection after OPCAB is not always easily diagnosed. It may be missed in patients who are experiencing no pain, and it may be confused with more common complications of coronary artery bypass surgery (for example, myocardial infarction, graft occlusion, left ventricular dysfunction, or arrhythmias), leading to crucial delay and possibly to sudden death from rupture of the aorta, as in our patient 2. Among the most ominous signs and symptoms of early dissection after OPCAB are acute neurologic deficits,15 including painless syncope, sudden cortical blindness, extremity numbness, and sudden paraplegia. The more typical tearing chest and back pain or the sudden loss of peripheral pulses after OPCAB should readily suggest the correct diagnosis. The prohibitively high mortality rate of early postoperative dissection is directly related to delays in its diagnosis and, consequently, in its surgical repair.16
Early postoperative aortic dissection can be clinically silent and be discovered only incidentally. Our literature search identified the case of a 64-year-old man who underwent on-pump CAB and was found incidentally to have experienced aortic dissection 4 months later.17 Because of his lack of symptoms, he was treated medically and was followed up for 2 years, with an uneventful result. Whether such indolent early dissection will become chronic and manifest itself many years later is controversial. Hagl and colleagues14 reviewed the cases of 109 patients with chronic aortic dissection, all of whom had undergone on-pump CAB, and concluded that the intimal tearing occurred most often at the site of the initial partial-occlusion clamp. In a subsequent case review of 11 patients who experienced late dissections of the ascending aorta after previous cardiac surgery (on-pump CAB and valvular), Stanger and colleagues18 deemed the dissections de novo events on the basis of the lack of characteristic histologic findings, the sites of the entry tears, and the times since initial surgery. They cautioned, however, that aortic dissection can follow cardiac surgery at any time, with no typical associated histologic evidence and a high in-hospital mortality rate.
Besides a high index of suspicion, timely diagnosis of early postoperative aortic dissection crucially depends upon a CT of the chest9 with intravenous contrast, in accordance with pulmonary-embolism protocol (a 5- to 10-min routine procedure in most hospitals). Typically, the CT will show a dissection flap in the ascending aorta that may extend to the aortic arch and the descending thoracic aorta, and contrast medium may be found in the true and false lumina. These findings are pathognomonic of type A or DeBakey type I acute aortic dissection and should prompt immediate surgical attention. The aortic injury can be repaired by local replacement of the dissected aortic segment with a patch graft or by replacement of the entire ascending aorta with a tube graft, followed by reimplantation of the veins in both cases.
Although transesophageal echocardiography (TEE) is routinely used in most hospitals during cardiac sur-gery, this precise diagnostic method is not always avail-able, and it may not be feasible in a patient who is un-comfortable or in unstable condition.
Intra- and postoperative acute aortic dissections usually originate at the site of an intimal tear that has been caused by aortic cannulation, insertion of a cardioplegia needle, application of an aortic cross-clamp or side-biting clamp, or at the site of a proximal anastomosis. Dissections have also occurred consequent to the use of mechanical proximal aortic connectors,10–12 which were hoped to eliminate cerebrovascular accidents from atheromatous and calcified aortas. More recently, dissections have been reported to occur retrograde, during or after endovascular stenting of the thoracic aorta.19 In several reports, the dissection has been attributed to a dilated ascending aorta at the time of initial surgery, but the authors failed to mention precise aortic measurements. Some dissections had specifically associated histologic findings, such as cystic medial necrosis; others had no specific associations other than aortic layers separated by a hematoma. Regardless, long-standing hypertension was common in most patients. Intraoperative dissection appears to occur more frequently in on-pump CAB and usually at the aortic cannulation site, whereas early postoperative dissection is more prevalent in OPCAB, probably due to a pulsatile pattern of arterial pressure during application of a side-biting clamp and performance of the proximal anastomoses.
Accordingly, prevention of early postoperative dissection in OPCAB should be directed toward averting trauma to the ascending aorta: specifically, by avoiding aortic clamping altogether and, ideally, by performing no-touch surgery with all arterial grafting. Although this technique is feasible in many younger patients, it may be impractical in an elderly insulin-dependent patient, for instance. It is therefore paramount to maintain a low systemic blood pressure by pharmacologic means during the performance of proximal anastomoses, after completion of the surgery, and over the long term thereafter. In our practice, we have routinely kept our patients' systolic blood pressure at less than 80 mmHg during proximal anastomoses, but we have avoided using a partial-occluding clamp altogether when the integrity of the aortic wall has been in question. We have also routinely used TEE and, when necessary, epicardial echocardiography before clamping the aorta. However, neither precaution has completely eliminated aortic dissection. Others18 have suggested aggressive replacement of the ascending aorta with a prosthetic graft in the presence of a dilated (more than 4 cm), calcific, ordiffusely atheromatous ascending aorta (a “hostile” aorta). Finally, despite the disappointing 1st generation of mechanical proximal aortic connectors in terms of graft patency and frequency of aortic dissection, continued technological advances in this field are expected to occur.
Early-but-delayed aortic dissection after OPCAB is a potentially catastrophic complication, the frequency of which has been generally underestimated. It can occur at any time after surgery. Its recognition can be elusive and requires a high index of suspicion, especially in the presence of atypical neurologic symptoms. Immediate CT scanning of the chest with intravenous contrast can lead to timely surgical repair that may save the patient's life.
Determining the true frequency of early dissection after OPCAB will require honest reporting of all such cases and insistence that autopsies be conducted after all delayed sudden deaths that occur after OPCAB.
Address for reprints: Imad F. Tabry, MD, Jim Moran Heart & Vascular Center, Holy Cross Hospital, 4725 N. Federal Hwy., Suite 503, Fort Lauderdale, FL 33308 E-mail: ten.tsacmoc@yrbati