Predictors of Adverse Outcome and Transient Neurological Dysfunction After Ascending Aorta/Hemiarch Replacement

Transcription

1 Predictors of Adverse Outcome and Transient Neurological Dysfunction After Ascending Aorta/Hemiarch Replacement Marek P. Ehrlich, MD, M. Arisan Ergin, MD, PhD, Jock N. McCullough, MD, Steven L. Lansman, MD, PhD, Jan D. Galla, MD, PhD, Carol A. Bodian, DrPH, Anil Z. Apaydin, MD, and Randall B. Griepp, MD Departments of Cardiothoracic Surgery and Biomathematics, The Mount Sinai Medical Center, New York, New York Background. This study was undertaken to determine predictors of adverse outcome and transient neurological dysfunction after replacement of the ascending aorta with an open distal anastomosis. Methods. All 443 patients (300 male, median age 63) undergoing replacement of the ascending aorta with an open distal anastomosis between 1986 and 1998 were included in the analysis. The ascending aorta alone was replaced in 190 (42.9%); 253 (57.1%) also had proximal arch replacement. Median hypothermic circulatory arrest (HCA) time was 25 minutes (range 12 to 68). Either death or permanent neurological dysfunction were considered adverse outcome (AO). Results. Adverse outcome occurred in 11.5% (51 of 443) of patients overall: in 7.4% of elective (20 of 269) or urgent (4 of 54) operations, but in 17% (19 of 113) of emergencies. Multivariate analysis of the group as a whole revealed that significant (p < 0.05) independent preoperative predictors of AO were age greater than 60 [odds ratio (OR) 2.2], hemodynamic instability (OR 2.7), and dissection (OR 1.9). For the 435 operative survivors, procedural variables predictive of AO were contained rupture (OR 2.8) and HCA time (OR 1.03/min). When only the 271 elective patients were analyzed separately, the need for a concomitant procedure (p 0.009, OR 3.6) and HCA time (p 0.002, OR 1.06/min) were the only predictors of AO in multivariate analysis. Transient neurological dysfunction (TND) occurred in 86 of 392 patients (22%). Significant predictors of TND for all patients without AO were age (OR 1.06/y), HCA time (OR 1.04/min), coronary artery disease (OR 2.2), hemodynamic instability (OR 3.4), and acute operation (OR 2.2). Survival of discharged patients was 93% at 1 year and 83% at 5 years. Conclusions. Early elective operation and shorter HCA time during ascending aorta/hemiarch surgery will reduce both AO and TND. (Ann Thorac Surg 2000;69: ) 2000 by The Society of Thoracic Surgeons Despite numerous advances over the last 30 years, aneurysms of the ascending aorta and aortic arch still remain a major challenge for cardiac surgeons. Since the first successful replacement of the aortic arch with perfusion of the head vessels by De Bakey and colleagues [1], various methods have been employed to preserve cerebral function during aneurysm operations. Currently, the most frequently used technique for cerebral protection during ascending aorta/aortic arch surgery is deep hypothermia and circulatory arrest (HCA). Although this technique was initially implemented in aortic arch surgery by Borst and colleagues in 1963 [2], it was a decade later, with the demonstration that prolonged HCA could radically simplify replacement of the aortic arch, that use of this technique became widespread in patients who had previously been considered inoperable Presented at the Forty-sixth Annual Meeting of the Southern Thoracic Surgery Association, San Juan, Puerto Rico, Nov 4 6, Address reprint requests to Dr Griepp, Department of Cardiothoracic Surgery, The Mount Sinai Medical Center, One Gustave L. Levy Place, Box 1028, New York, NY 10029; marekehrlich@hotmail.com. [3]. With much better survival of these patients in subsequent years came the realization that a considerable number of these high-risk often elderly patients suffered adverse neurological sequelae, but only a few studies have looked carefully at the impact of use of HCA on permanent and temporary neurological outcome, as well as on both early and late survival [4, 5]. The current study was undertaken to examine both preoperative and procedural predictors of adverse outcome after operations on the ascending aorta with an open distal anastomosis during a period of HCA. We defined adverse outcome as either death or permanent neurological disability, since a successful operation should result not only in survival, but a good quality of life. For most patients, the recitation of a series of separate estimates of the risk of various complications and of mortality is relatively meaningless: what they want to know are the odds that they will emerge from the operation more or less as they were before, but without aneurysm-related symptoms or the threat of imminent aortic rupture. We also examined factors affecting the incidence of 2000 by The Society of Thoracic Surgeons /00/$20.00 Published by Elsevier Science Inc PII S (00)

2 1756 EHRLICH ET AL Ann Thorac Surg ASCENDING AORTA/HEMIARCH REPLACEMENT 2000;69: Table 1. Clinical Characteristics Acute Dissections (n 121) Chronic Dissections (n 90) Nondissecting Aneurysms (n 232) All (n 443) p Value Median age (y) a Age 60 y (%) Male sex (%) CAD (%) History of hypertension (%) Diabetes (%) Neurological history (%) Marfan s syndrome (%) Hemodynamic compromise (%) Rupture (%) Emergency operation (%) Reoperation (%) Any concomitant procedures (%) a Kruskal-Wallis; all other p values from chi-square analysis indicating any differences among the 3 subgroups of patients: acute dissections versus chronic dissections versus nondissecting aneurysms. CAD coronary artery disease. temporary neurological dysfunction because this assessment seems to reflect the adequacy of cerebral protection, and therefore can help in determining how further to improve cerebral recovery following aortic surgery [5]. In addition, recent evidence has suggested that the occurrence of this syndrome is frequently associated with long-lasting albeit subtle cognitive impairment [4]. and Methods Patient Population All 443 consecutive patients who underwent operations on the ascending aorta and proximal aortic arch with HCA as the principal method of cerebral protection from January 1986 to March 1998 were reviewed retrospectively using information gathered contemporaneously in our departmental database and supplemented, as necessary, from patient records. with isolated aortic arch aneurysms were excluded from analysis because they were felt to constitute a separate group clinically, with greater diversity in clinical symptoms, surgical approaches, and methods of cerebral protection. There were 300 men and 143 women whose ages ranged from 8 to 88 years (mean years); 244 (55.1%) were older than 60 years at the time of surgery. Of the entire cohort, 271 patients (61.2%) were operated on electively, 12.6% urgently (within 72 hours of admission), and 26.2% on an emergent basis (within 24 hours of admission). A total of 211 patients (47.6%) had type A aortic dissections (121 acute and 90 chronic), 143 had degenerative aortic disease, and 47 had atherosclerotic aortas. Thirty-four patients in this series had Marfan s syndrome. Eight patients had miscellaneous other etiologies such as mycotic aneurysms. As seen in Table 1, a history of hypertension was the most common preoperative finding, although there was also a relatively high proportion of patients undergoing reoperation and experiencing rupture. If one compares patients with either acute or chronic dissection with patients with degenerative or atherosclerotic aneurysms, it is apparent not only that urgent and emergent operation, rupture, and hemodynamic compromise were more frequently seen among patients with dissection (all p 0.001), largely because of the inclusion of patients with acute dissections, but also that patients with dissection were older ( p 0.09), more often had a history of hypertension ( p 0.001), and were more likely to have had previous operations ( p 0.001). Interestingly, no difference was found in the incidence of dissection and nondissecting aneurysm in patients with Marfan s syndrome. Surgical Technique The surgical technique and the application of profound HCA throughout the study period were standard. The method of profound hypothermia and circulatory arrest has previously been described in detail [5]. Briefly, it consists of core cooling during cardiopulmonary bypass to produce profound total body hypothermia to an average core temperature of 12 to 15 C measured in the esophagus. The proximal repair is accomplished during this period of core cooling, which, in an average adult, requires at least 30 to 40 minutes. The part of the procedure that requires interruption of cerebral blood flow is then done during the period of circulatory arrest. The head is packed in ice to prevent warming of the central nervous system during prolonged circulatory arrest. During the later years of the study, 28 patients felt to be at high risk of embolization underwent retrograde cerebral perfusion (RCP): in 17 patients, an interval of RCP was utilized with the hope of enhancing cerebral protection; in 11, RCP was carried out very briefly, just long enough to flush the cerebral circulation before resumption of antegrade perfusion.

3 Ann Thorac Surg EHRLICH ET AL 2000;69: ASCENDING AORTA/HEMIARCH REPLACEMENT 1757 Table 2. Intraoperative and Postoperative s (Operating Room Deaths Excluded) Acute Dissections (n 119) Chronic Dissections (n 86) Nondissecting Aneurysms (n 230) All (n 435) p Value CPB (min) 201 (94,406) 223 (93,457) 201 (84,452) 204 (84,457) 0.03 a HCA (min) 31 (12,66) 29 (12,67) 22 (12,68) 25 (12,68) a Extent of replacement (%) Hemiarch Aortic valve replacement (%) None Biological Mechanical Bentall (%) No Button Cabrol Classic Concomitant procedures (%) 0.01 None CABG Others Any complications (%) a Kruskal-Wallis; all other p values from chi-square analysis indicating any differences among the 3 subgroups of patients: acute dissections versus chronic dissections versus nondissecting aneurysms. CPB and HCA times are median values (and ranges). CABG coronary artery bypass grafting; CPB cardiopulmonary bypass; HCA hypothermic circulatory arrest. Operative Procedures As outlined in Table 2, 190 patients had only ascending aorta replacement; 253 (57.1%) had hemiarch replacement. The aortic valve was replaced with a mechanical one in 225 patients (50.8%); a biological valve was implanted in 70 patients (15.8%), and the aortic valve was untouched in 148 patients (33.4%). In more than half of all 443 patients (63.5%), an aortic valveascending aorta conduit was implanted utilizing some variant of the Bentall procedure. The button-bentall was the most common way of implanting the coronary arteries, followed by the Cabrol technique and the classic Bentall. Fourteen patients had aortic valve replacement with a separate ascending aorta conduit (Wheat procedure). Coronary bypass grafting was the most common concomitant procedure and was performed in 84 patients (19%) overall. Other additional procedures such as mitral valve replacement were done in 10.2% of the patients. Table 2 shows other important intraoperative variables. Median circulatory arrest time (HCA) in this series was 25 minutes (range, 12 to 68). Circulatory arrest time was kept below 30 minutes in 68.9% of all patients, and only 25 patients had HCA intervals exceeding 50 minutes. Definitions of Adverse Outcome and Neurological Complications Adverse outcome was defined as intraoperative or hospital death, or the occurrence of permanent neurological injury. The presence of neurological dysfunction at the time of discharge from the hospital, whether focal injury (stroke) or global (coma), was considered permanent neurological injury. Transient neurological dysfunction (TND), which could only be assessed in operative survivors without permanent neurological dysfunction, was defined as the occurrence of prolonged postoperative confusion, agitation, or transient delirium. Since 1993, the severity of TND has been graded on a five-point scale, but since this more detailed assessment was not available for all patients, only the presence or absence of TND was scored. Computed tomography, when performed on patients with temporary dysfunction, was usually normal. Statistical Methods Initially, all patients were treated as a single group, but subsequent analysis indicated that patients with acute dissection differed in significant ways from those with chronic dissection and nondissecting aneurysms, and that these latter groups also differed from one another. Analysis for risk factors was therefore carried out not only for the group as a whole, but separately for patients with acute dissection and for patients with either chronic dissection or chronic nondissecting aneurysms; some comparison of patients before and after 1993, when a more detailed assessment of occurrence of TND was begun, was also carried out. All patients were included in the analysis of preoperative factors impacting on adverse outcome; patients who died in the operating room were excluded from analysis of intraoperative variables, and patients with adverse outcome were excluded from the analysis of transient neurological dysfunction. In addi-

4 1758 EHRLICH ET AL Ann Thorac Surg ASCENDING AORTA/HEMIARCH REPLACEMENT 2000;69: Table 3A. Univariate Analysis of Preoperative Risk Factors for Adverse Outcome in All (n 443) of With AO (%) p Value a Sex 0.42 Male (11) Female (13) Age Age 60 y (7) Age 60 y (15) Coronary artery disease 0.4 Yes (14) No (11) Etiology Degenerative (3) Atherosclerotic 47 8 (17) Dissection (16) Marfan s 11 1 (9) Other 31 4 (13) Acute dissection 0.02 Yes (17) No (9) Hemodynamic compromise Yes 30 9 (30) No (10) Rupture Yes (24) No (9) Urgency of operation Elective (8) Urgent 56 6 (11) Emergent (20) tion, an analysis of likelihood of adverse outcome following elective surgery was also carried out. All putative factors were analyzed initially by chisquare or Kruskal-Wallis tests, as appropriate. Each univariate analysis was followed by stepwise logistic regression to determine independent risk factors. A p value less than 0.05 was considered significant. Results Adverse Outcome Adverse outcome (AO) death or permanent neurological deficit occurred in 51 of 443 patients overall (11.5%). Adverse outcome was significantly lower following elective (20/269) and urgent (4/54) operations (both 7.4%) than after emergency procedures where it was observed in 17% (19/113) of patients ( p 0.01). Univariate analysis of the group as a whole revealed a number of preoperative, intraoperative, and immediate postoperative factors which were associated with adverse outcome (Table 3). The preoperative risk factors were: age greater than 60 years, hemodynamic compromise, emergent operation, acute dissection, and rupture. If one compares patients with acute or chronic dissection and Table 3B. Univariate Analysis of Operative and Postoperative Risk Factors for Adverse Outcome in All Surviving the Operation (n 435 ) of With AO (%) p Value a HCA time HCA time 30 min (7) HCA time 30 min (16) Extent of replacement 0.85 Ascending aorta (10) Hemiarch (10) Aortic valve 0.04 None (15) Mechanical (7) Biological 68 5 (7) Bentall procedure 0.01 No (4) Button (6) Cabrol 27 5 (19) Classic 11 0 Concomitant procedures 0.38 None (9) CABG (14) Others 44 5 (11) Any complication Yes (21) No (2) a p Values indicate differences among the listed categories for each variable with regard to adverse outcome. AO adverse outcome; CABG coronary artery bypass grafting; HCA hypothermic circulatory arrest. those without dissection, it is apparent that the distribution of age was significantly different: There was a preponderance of elderly patients with acute and especially with chronic dissections (Table 1). There were also significant differences between etiological groups with regard to a history of hypertension and coronary artery disease, and in the incidence of reoperation. Not surprisingly, comparison of acute and chronic dissection with nondissecting aneurysms also revealed that significantly more patients with acute dissection had emergency operations, hemodynamic compromise preoperatively, and evidence of rupture intraoperatively. Among the intraoperative factors predicting a poor outcome in the group as a whole (Table 3B) was a duration of HCA exceeding 30 minutes. Again, if one compares the intraoperative data from the 3 groups of patients (Table 2), the longest cerebral ischemic times were seen in patients with acute dissections (31 minutes); cerebral ischemic time was greater with chronic dissections (29 minutes) than with nondissecting aneurysms (22 minutes, p ). There were also significant differences among the groups with regard to the kinds of operations carried out: a much higher percentage of patients with acute dissections did not have replacement of the aortic valve and had no concomitant procedures. in whom the aortic valve was not replaced had a

5 Ann Thorac Surg EHRLICH ET AL 2000;69: ASCENDING AORTA/HEMIARCH REPLACEMENT 1759 Table 4A. Multivariate Analysis of Preoperative Factors Responsible for AO (n 443) Odds Ratio p Value Age 60 y Hemodynamic instability Dissection Table 4B. Multivariate Analysis of Preoperative and Procedural Factors Responsible for AO (n 435) Odds Ratio p Value Contained rupture HCA time 1.03/min 0.02 AO adverse outcome; HCA hypothermic circulatory arrest. higher incidence of AO than those who had aortic valve replacement, and patients who had a valved conduit implanted using the Cabrol technique had a higher incidence of AO than those with a classic or button Bentall operation. These observations can both be explained by noting the significant preponderance of patients with acute dissection in both higher risk groups. In univariate analysis of the group as a whole, the occurrence of any postoperative complication also increased the chance of adverse outcome: infection, renal failure, bleeding (requiring reoperation for hemostasis), cardiac problems, or respiratory compromise (prolonged intubation or tracheostomy; Table 3B). With the exception of those related to the heart, all postoperative complications occurred in a significantly higher percentage of patients with acute dissection ( p 0.001) and also tended to occur more frequently in patients with chronic dissection than in those with nondissecting aneurysms. Multivariate analysis of the group as a whole revealed that statistically significant independent preoperative predictors of adverse outcome were age greater than 60, hemodynamic instability, and presence of dissection (Table 4A). For the 435 operative survivors, additional procedural variables predictive of adverse outcome were contained rupture and HCA duration (Table 4B). When only patients with chronic dissections or nondissecting aneurysms were considered, excluding patients with acute dissection, the preoperative risk factors for adverse outcome in multivariate analysis were female sex [ p 0.06, odds ratio (OR) 2.1], emergency operation ( p 0.06, OR 4.1), and any concomitant procedure ( p , OR 4.1). Chronic dissection was not an independent risk factor for adverse outcome. Separate analysis of acute dissections did not reveal any independent preoperative risk factors for adverse outcome. The 271 elective cases were then considered separately to determine which variables were associated with adverse outcome in this group of patients, in whom accurate assessment of operative risk is most relevant. Univariate assessment revealed a history of coronary artery disease ( p 0.03), etiology of the aneurysm ( p 0.01), need for a concomitant procedure ( p 0.001), coronary artery bypass surgery ( p 0.002), cerebral ischemic time ( p ), duration of cardiopulmonary bypass ( p 0.02), myocardial ischemic time ( p 0.03), and presence of any complication ( p 0.001) to be associated with adverse outcome. As might have been expected from the considerable overlap with the group comprising all chronic dissections and nondissecting aneurysms, multivariate analysis of all elective patients revealed three risk factors for adverse outcome: performance of a concomitant procedure ( p 0.009, OR 3.6), duration of cerebral ischemia ( p 0.002, OR 1.06/minute), and occurrence of any postoperative complication. Transient Neurological Dysfunction Transient neurological dysfunction, which was evaluated only in patients surviving operation without permanent neurological dysfunction, occurred in 86 of 392 patients (22%). TND occurred in 37% of patients with acute dissection, 25% of patients with chronic dissection, and 14% of patients with nondissecting aneurysms, p Since 1993, when a graded scale of symptoms was introduced and encouraged more careful assessment, 27% of 296 patients were identified as having TND, compared with 12% in the earlier interval, p Univariate analysis showed that some of the same preoperative factors which were important contributors to adverse outcome were also implicated in the occurrence of TND, as shown in Table 5A: age, etiology, hemodynamic compromise, urgency of operation, and presence of acute dissection. In addition, the presence of coronary artery disease was significant: 41% of patients with a history of coronary artery disease and 42% with concomitant coronary artery bypass operation experienced TND. In this regard, it should be noted that the incidence of coronary artery disease was significantly higher in patients with chronic dissection (28%) than in the other groups. Among intraoperative variables (Table 5B), an HCA time exceeding 30 minutes was present in 29% of patients without adverse outcome, of whom 34% presented with TND. In addition to a higher incidence of TND with concurrent coronary artery bypass operation, patients with hemiarch replacement had more than twice the incidence of TND than those with just ascending aorta replacement. Use of a biological rather than a mechanical valve for aortic valve replacement was associated with a higher incidence of TND (33%) than use of a mechanical valve (15%), and the Cabrol variant of the Bentall procedure was also associated with a high risk of TND, but it should be noted that biological valves were more frequently used in older patients, and the Cabrol variant of the Bentall procedure was more common among patients with acute dissection. When risk factors for TND were considered only in the 292 patients with chronic dissections or aneurysms, a number of preoperative, intraoperative, and immediate postoperative factors were significant in univariate analysis. In addition to age and a history of coronary artery

6 1760 EHRLICH ET AL Ann Thorac Surg ASCENDING AORTA/HEMIARCH REPLACEMENT 2000;69: Table 5A. Univariate Analysis of Preoperative Risk Factors for TND in All Without AO (n 392) of With TND (%) p Value a Sex 0.21 Male (20) Female (26) Age Age 60 y (11) Age 60 y (32) Coronary artery disease Yes (41) No (18) Etiology Degenerative (15) Atherosclerotic 39 6 (15) Dissection (32) Marfan s 10 0 Other 27 3 (11) Acute dissection Yes (37) No (17) Hemodynamic compromise Yes (62) No (20) Rupture Yes (38) No (19) Urgency of operation Elective (17) Urgent (22) Emergent (35) Table 5B. Univariate Analysis of Intraoperative and Postoperative Risk Factors for TND in All Without AO (n 392) of With TND (%) p Value a HCA time HCA time 30 min (15) HCA time 30 min (34) Extent of replacement Ascending aorta (12) Hemiarch (29) Aortic valve None (27) Mechanical (15) Biological (33) Bentall procedure 0.05 No (27) Button (18) Cabrol 22 8 (36) Classic 11 1 (9) Concomitant procedures None (18) CABG (42) Others 39 7 (18) Any complication Yes (44) No (10) a p Values indicate differences among the listed categories for each variable with regard to TND. AO adverse outcome; CABG coronary arterial bypass grafting; HCA hypothermic circulatory arrest; TND transient neurological dysfunction. disease (both p 0.001), a history of hypertension ( p 0.04) or of diabetes ( p 0.01) was an adverse risk factor, but the presence of Marfan s syndrome significantly lowered the risk of TND ( p 0.03). Intraoperative factors associated with a high risk of TND in patients with chronic lesions included not only cerebral ischemic time ( p 0.001) but also duration of cardiopulmonary bypass ( p 0.02), duration of myocardial ischemia ( p 0.06), and the use of RCP (p 0.04). Other factors associated with higher risk of TND in this subgroup were hemiarch rather than ascending aorta repair ( p 0.002), use of a biological valve ( p 0.001), and coronary artery bypass grafting ( p 0.001). The occurrence of any postoperative complications was highly correlated with TND ( p 0.001). Especially renal, respiratory, or infectious complications (all p 0.001), but also cardiac complications ( p 0.01) and bleeding ( p 0.03) were associated with a high incidence of TND; it is admittedly more difficult to evaluate TND in the presence of complications which may require additional sedation for their treatment. Significant predictors of TND in multivariate analysis of the group as a whole were age, coronary artery disease, hemodynamic instability, acute operation, and HCA duration (Table 6). When only patients with chronic dissections or aneurysms were considered, factors predicting TND in multivariate analysis were age ( p 0.002, OR 1.07/year), coronary artery bypass grafting ( p 0.006, OR 3.1), HCA duration ( p 0.02, OR 1.05/minute), and the presence of any postoperative complication ( p , OR 6.8). Survival Hospital mortality was 9% (40 patients), but only 8 patients died in the operating room. Survival of discharged patients was 93% at 1 year and 83% at 5 years Table 6. Multivariate Analysis of Preoperative and Operative Factors Responsible for TND (n 392) Odds Ratio p Value Age 1.06/y Hemodynamic instability Coronary artery disease Acute operation HCA time 1.04/min HCA hypothermic circulatory arrest; dysfunction. TND transient neurological

7 Ann Thorac Surg EHRLICH ET AL 2000;69: ASCENDING AORTA/HEMIARCH REPLACEMENT 1761 Fig 1. Kaplan-Meier survival curve of all patients after discharge following ascending aorta operation with an open distal anastomosis (with standard error bars). (Fig 1). Long-term survival was not significantly different in patients with acute dissection, chronic dissection, and nondissecting aneurysms. Comment Although results of operations on the ascending aorta and proximal aortic arch have steadily improved in recent years, there is still a relatively high mortality, especially in emergency cases, and an incidence of cerebral complications which has been reported to vary between 7 and 35% [6, 7]. This has led to a reexamination of several aspects of these operations, including increasing questioning of the use of hypothermic circulatory arrest as the principal means of cerebral protection during arrest of the antegrade circulation [8 12]. The perceived shortcomings of HCA have led to the adoption of alternate methods of cerebral protection by a number of aortic surgeons [8 10, 12, 13]. This trend prompted us to reexamine our experience with use of HCA during ascending aorta/hemiarch surgery in order to enable comparisons with recent reports using selective antegrade cerebral perfusion and retrograde cerebral perfusion [9, 10, 12, 13]. We elected to combine permanent neurological injury and death as adverse outcome to increase our sensitivity in identifying preoperative and intraoperative risk factors, and because the threat of permanent neurological injury is often as worrisome to prospective patients as the chance of dying, and is often associated with early mortality [14]. We also wanted to reexamine the syndrome we have termed transient neurological dysfunction because previous studies have consistently shown TND to correlate with duration of cerebral ischemia, in contrast with permanent neurological injury, which is predominantly focal, and for which prolonged HCA has not invariably emerged as a risk factor [5]. In addition to being a reflection of suboptimal cerebral protection, recent evidence has also suggested that TND seems to be a marker of long-lasting but subtle cognitive impairment [4]. Measures aimed at reducing the incidence of TND should therefore contribute toward improving cerebral protection and outcome following ascending aorta/aortic arch operations. The finding that advanced age is a risk factor for both adverse outcome and transient neurological dysfunction is not unexpected, and is in accord with our own earlier findings and the observations of others [4, 5, 15, 16]. Similarly, the idea that hemodynamic instability, rupture, emergency operation, and the presence of acute dissection increase the risk of surgery on the ascending aorta and aortic arch is also not surprising in light of earlier studies and common sense [17]. The absence of Marfan s syndrome as a risk factor for adverse outcome is symbolic of the success which has been achieved in a particular group of high-risk patients by early elective intervention, reducing the need for emergency operations with their associated higher mortality and morbidity. We are somewhat surprised that duration of HCA emerged in this study as a risk factor for adverse outcome as well as for transient neurological dysfunction. Our previous studies, as well as studies by others in which smaller numbers of patients were surveyed, did not identify duration of HCA as a risk factor for stroke: We attributed this to the notion that stroke is predominantly embolic in origin [18]. The fact that we see an impact of prolonged HCA on the incidence of adverse outcome in this study may be a result of increased statistical power derived from combining death and permanent neurological dysfunction as adverse outcome, and of having examined a larger number of patients: others have previously documented a relationship of prolonged HCA and HCA/RCP with both mortality and stroke [10, 16, 19]. The identification of HCA duration as a risk factor for adverse outcome in this series characterized by thorough cooling to low temperatures and a mean interval of cerebral ischemia well below the consensus estimate of a safe period of circulatory arrest underscores the continuing need to reduce HCA duration. The idea that use of RCP can reduce the morbidity and mortality of these operations is not borne out by our data, in which RCP was used in 28 admittedly high-risk patients without evidence of improved outcome. Although several authors have reported excellent results using RCP, comparisons have usually been made to historical controls, and the contribution of other factors to improved results with ascending aorta/aortic arch surgery in recent years has not been acknowledged [9, 13, 20 22]. Some reports have documented that prolonged RCP is a risk factor for stroke and mortality in much the same way as prolonged HCA [19]. Other recent experimental and clinical studies have cast doubt on the ability of RCP to provide meaningful metabolic sustenance even at deeply hypothermic temperatures, leading to the conclusion that the major benefit of RCP is to assure maintenance of deep hypothermia during arrest of antegrade circulation [23, 24]. In addition to its contribution to adverse outcome, the duration of HCA is a concern because of its unquestionably strong impact on the occurrence of transient neurological dysfunction. A recent study documented that

8 1762 EHRLICH ET AL Ann Thorac Surg ASCENDING AORTA/HEMIARCH REPLACEMENT 2000;69: Fig 2. Estimated probability of adverse outcome following elective procedures on the ascending aorta/proximal aortic arch for patients with and without need for concomitant procedures, depending upon anticipated duration of HCA, which can be estimated by an experienced surgeon from the extent and complexity of the projected repair. Data are from the 271 patients in this series who underwent elective ascending aorta/aortic arch surgery. TND was much more prevalent among patients with persistent loss of cognitive function on psychometric testing 6 weeks after surgery than among patients with prompt recovery of test scores, suggesting that TND may not be entirely a transient phenomenon, but may be a marker for long-lasting cognitive deficits [4, 25]. Since formal psychometric testing is an expensive and cumbersome tool for assessing the adequacy of cerebral protection during aortic surgery, we urge use of the relatively simple clinical evaluation of TND in the immediate postoperative period as a way of comparing the efficacy of different protocols for cerebral protection. The emergence of coronary artery disease as an independent risk factor for the occurrence of TND is interesting. We speculate that the presence of coronary artery disease may be an indication of cerebrovascular atherosclerosis in these patients, which puts them at higher risk of cerebral dysfunction postoperatively. It should be noted that more than half of the patients with chronic dissection in this series were undergoing reoperation, often after coronary artery bypass grafting, reinforcing the need to avoid clamp injury during coronary artery procedures to avert the possibility of iatrogenic dissection in these high-risk patients. Overall, however, the role of prolonged HCA has far less of an impact on adverse outcome and on the occurrence of TND than the need for emergency surgery for acute dissection or rupture. The results of this study reinforce dramatically the notion that elective operation is far safer than emergency surgery on the ascending aorta/proximal aortic arch, and support the idea that elective replacement of the ascending aorta should be undertaken early in patients at risk for rupture or dissection. Earlier elective operation would also help reduce the contribution of age to the risk both of adverse outcome and of TND. Our relatively low incidence of adverse outcome in elective operations encourages this approach, and a recent series of ascending aortic operations suggests that mortality and stroke can each be reduced to below 2% in elective cases [26]. A rational approach to advising patients with chronic aneurysms is to try to balance the likelihood of rupture against the risk of operative intervention, bearing in mind that most patients who experience rupture do not reach a suitable hospital in time to undergo emergency operation. On the basis of data gathered from a large number of patients with chronic ascending aortic aneurysms whom we are following, we estimate that the risk of rupture of an aneurysm with a diameter of 5 cm is 5% to 7% per year; the risk increases to 10% to 15% per year with a6cmaneurysm. We therefore currently recommend that patients with few risk factors associated with adverse outcome younger patients without coronary artery disease, with an uncomplicated lesion likely to require a relative short duration of HCA undergo elective resection of an ascending aortic aneurysm at a diameter of 5 cm. with a higher operative risk older patients requiring concomitant procedures and with more complex or extensive lesions should probably be advised to undergo elective ascending aortic aneurysm repair at a diameter of 6cm. In future, we will attempt to construct a more individualized estimate of the risk of adverse outcome based on the results of the multivariate analyses of the patients in this series, relying especially on those risk factors we have identified as important in patients operated on electively, using the characteristics of the aortic lesion to estimate the likely duration of cerebral ischemia, and making an educated guess as to whether or not the patient will require a concomitant procedure. For example, we can use Figure 2 to estimate that the risk of AO following a straightforward uncomplicated elective ascending aneurysm resection would be only 6% if no concomitant procedure were required, versus 18% if coronary artery bypass grafting, mitral valve replacement, or other adjunctive procedure were necessary, whereas a more complex operative repair necessitating a longer cerebral ischemic time would entail a risk of AO of 10% without, and 25% with a concomitant procedure. It is our hope that more refined understanding of operative risk will lead to earlier elective intervention in suitable patients, with fewer emergency operations and, therefore, improved outcomes following resection of ascending aortic aneurysms. References 1. DeBakey ME, Crawford ES, Cooley DA, Morris GC. Successful resection of a fusiform aneurysm of aortic arch with replacement by homograft. Surg Gynecol Obstet 1957;105: Borst HG, Schaudig A, Rudolph W. Arteriovenous fistula of the aortic arch: repair during deep hypothermia and circulatory arrest. J Thorac Cardiovasc Surg 1964;48: Griepp RB, Stinson EB, Hollingsworth JF, Buehler D. Prosthetic replacement of the aortic arch. J Thorac Cardiovasc Surg 1975;70:

9 Ann Thorac Surg EHRLICH ET AL 2000;69: ASCENDING AORTA/HEMIARCH REPLACEMENT Ergin MA, Uysal S, Reich DL, et al. Temporary neurological dysfunction after deep hypothermic circulatory arrest: a clinical marker of long-term functional deficit. Ann Thorac Surg 1999;67: Ergin MA, Galla JD, Lansman SL, et al. Hypothermic circulatory arrest in operation on the thoracic aorta: determinants of operative mortality and neurological outcome. J Thorac Cardiovasc Surg 1994;107: Lansman SL, Raissi S, Ergin MA, Griepp RB. Urgent operation for acute transverse aortic arch dissection. J Thorac Cardiovasc Surg 1989;97: Livesay JJ, Cooley DA, Duncun JM, et al. Open distal anastomosis: improved results in the treatment of aneurysms of the aortic arch. Circulation 1982;66(Suppl 1): Kipfer B, Striffeler H, Gersbach P, et al. Surgery for acute ascending aortic dissection: closed versus open distal aortic repair. Eur J Cardiothorac Surg 1995;9: Raskin SA, Coselli JS. Retrograde cerebral perfusion: overview, techniques and results. Perfusion 1995;10: Moshkovitz Y, David TE, Caleb M, et al. Circulatory arrest under moderate systemic hypothermia and cold retrograde cerebral perfusion. Ann Thorac Surg 1998;66: Veeragandham RS, Hamilton IN, O Connor C, et al. Experience with antegrade bihemispheric cerebral perfusion in aortic arch operations. Ann Thorac Surg 1998;66: Byrne JG, Fitzgerald DJ, Aranki SF. Simultaneous selective cerebral perfusion and systemic circulatory arrest through the right axillary artery for aortic surgery. J Card Surg 1998; 13: Safi HJ, Letsou GV, Iliopoulos DC, et al. Impact of retrograde cerebral perfusion on ascending aortic and arch aneurysm repair. Ann Thorac Surg 1997;63: Goossens D, Schepens M, Hamerlijnck R, et al. Predictors of hospital mortality in type A aortic dissections: a retrospective analysis of 148 consecutive surgical patients. Cardiovasc Surg 1998;6: Okita Y, Takamoto S, Ando M, et al. Predictive factors for postoperative cerebral complications in patients with thoracic aortic aneurysms. Eur J Cardiothorac Surg 1996;10: Svensson LG, Crawford ES, Hess KR, et al. Deep hypothermia with circulatory arrest: determinants of stroke and early mortality in 656 patients. J Thorac Cardiovasc Surg 1993;106: Santini F, Pessotto R, Restuivo S, et al. Factors influencing outcome after emergency surgical repair of acute type A aortic dissection. G Ital Cardiol 1999;29: Okita Y, Takamoto S, Ando M, et al. Mortality and cerebral outcome in patients who underwent aortic arch operations using deep hypothermic circulatory arrest with retrograde cerebral perfusion: no relation of early death, stroke, and delirium to the duration of circulatory arrest. J Thorac Cardiovasc Surg 1998;115: Wong CH, Bonser RS. Does retrograde cerebral perfusion affect risk factors for stroke and mortality after hypothermic circulatory arrest? Ann Thorac Surg 1999;67: Nguyen B, Muller M, Kipfer B, et al. Different techniques of distal aortic repair in acute type A dissection: impact on late morphology and reoperation. Eur J Cardiothorac Surg 1999; 15: David TE, Armstrong S, Ivanov J, Barnard S. Surgery for acute type A aortic dissection. Ann Thorac Surg 1999;67: Stowe CL, Baertlein MA, Wierman MD, Rucker M, Ebra G. Surgical management of ascending and aortic arch disease: refined techniques with improved results. Ann Thorac Surg 1998;66: Higami T, Kozawa S, Asada T, et al. Retrograde cerebral perfusion versus selective cerebral perfusion as evaluated by cerebral oxygen saturation during aortic arch reconstruction. Ann Thorac Surg 199;67: Cheung AT, Bavaria JE, Weiss SJ, et al. Neurophysiologic effects of retrograde perfusion used for aortic reconstruction. J Cardiothorac Vasc Anesth 1998;12: Reich DL, Uysal S, Sliwinski M, et al. Neuropsychological outcome after deep hypothermic circulatory arrest in adults. J Thorac Cardiovasc Surg 1999;117: Cohn LH, Rizzo RJ, Adams DH, et al. Reduced mortality and morbidity for ascending aneurysm resection regardless of cause. Ann Thorac Surg 1996;62: DISCUSSION DR ALVAN W. ATKINSON (Raleigh, NC): I just wondered, did you use circulatory arrest without retrograde cerebral perfusion or did you use it in some patients, and if so, did you notice any change in your statistics? DR MCCULLOUGH: Actually our group is using retrograde cerebral perfusion in a much more limited fashion over the last couple of years. We tend to use retrograde cerebral perfusion more as a washout technique, a brief interval at the end of the procedure to wash out gross macroscopic debris from the arch. So, the number of patients in this series that had any significant retrograde cerebral perfusion was small and was actually correlated with a worse outcome, but we did not make too much of that in this paper because our group used retrograde cerebral perfusion in patients who we thought were at higher risk. So, yes, it was used in a small portion of the patients and it was associated with a worse outcome. DR WILLIAM A. BAUMGARTNER (Baltimore, MD): I would like to congratulate Dr Ehrlich and his group. As we all are aware, their institution is one of the leading ones in this field of aortic surgery. I think your work has helped us to understand a lot about the risks and methods of these complicated procedures. I have a couple of questions regarding transient neurologic dysfunction. As you know, we have been very interested in this aspect of it because I think it is the one area that we, perhaps, might be able to control some day. What was the incidence of transient neurologic dysfunction? I know you have done work on cognitive evaluation of patients in the past, and I wonder if any of these patients had pre and postoperative testing performed. DR MCCULLOUGH: Transient neurological dysfunction defined in this series of patients is the surgeon s observation of postoperative delirium, prolonged confusion, any nonfocal and ultimately grossly resolved neurological dysfunction during their hospitalization. This group did not have any systematic application of neuropsychiatric or neuropsychological testing as in our previous paper, but our previous paper did show in patients that underwent circulatory arrest, if they performed poorly on their early neuropsych testing, that was a strong predictor of a poor performance later. So, transient neurological dysfunction would probably correlate with poor performance on early neuropsych testing, and that would be associated with poorer late neuropsych performance, particularly in the domains of fine motor function and memory.