Individualized surgical strategy for the reduction of stroke risk in patients undergoing coronary artery bypass grafting

Background . This study was designed to evaluate the efficacy of a protocol of systematic screening of the ascending aorta and internal carotid arteries and individualization of the surgical strategy to the ascending aorta and internal carotid arteries status in reducing the stroke incidence among patients undergoing coronary artery bypass grafting.

Methods . On the basis of a pre- and intraoperative screening of the ascending aorta and internal carotid arteries, 2,326 consecutive patients undergoing coronary artery bypass grafting were divided in low, moderate, and high neurologic risk groups. In the high-risk group dedicated surgical techniques were always adopted and the reduction of the neurologic risk was considered more important than the achievement of total revascularization.

Results . The incidence of perioperative stroke in the high-risk group was similar to those of the other two groups (1.1 versus 1.3 and 1.1%, respectively; p = not significant); however, angina recurrence was significantly more frequent in the high-risk group.

Conclusions . The described strategy allows a low rate of perioperative stroke in high-risk patients undergoing coronary artery bypass grafting. Whether the reduction of the neurologic risk outweighs the benefits of complete revascularization remains to be determined.

 

 

Postoperative stroke represents one of the most severe complications of coronary artery operations and, as a result of the progressive aging of the population referred for cardiac operations, it will probably tend to increase in frequency in the near future. For this reason, the identification of the surgical strategies that can allow a reduction of postoperative neurologic events in the high-risk population is a priority objective of clinical research in cardiac surgery.

In this report we analyze our 4 years of prospective experience with the use of an integrated protocol aimed at reducing the incidence of cerebrovascular events in patients undergoing coronary artery bypass procedures by means of the systematic screening of the aorta and internal carotid arteries and adoption of dedicated surgical strategies in case of atherosclerotic disease of one or both vascular districts.

This study was prospectively started in October 1993 to evaluate the possibility of reducing the incidence of cerebrovascular complications in patients undergoing coronary artery bypass grafting by means of a protocol of systematic screening of the ascending aorta and the internal carotid arteries and modification of the surgical strategy on the basis of the status of these two vascular districts.

For the purpose of this report, data collection ended in January 1997. All cases submitted to isolated coronary artery bypass (ie, not associated with any other cardiac procedures) on elective or urgent basis at our institution in this time frame were considered.

Emergency cases were not included due to the impossibility of performing a complete preoperative evaluation of the carotid arteries. Twenty-one elective or urgent patients who could not complete this evaluation were also excluded from the data analysis. Overall, 2,326 patients were included.

During the study period no major modifications in the surgical and anesthesiologic technique were adopted.

Following a protocol in use at our Institution since 1990, all the 2,326 patients were submitted to the following tests:

Extracranic carotid arteries were evaluated by color flow duplex scanning study; B-mode and color flow images of the common, external, and internal carotid arteries were obtained in longitudinal and transverse planes. Presence of plaques was noted and reduction in the cross-sectional area of the lumen was calculated. Doppler velocity spectra were recorded from each of these vessels, maintaining the angle of insonation as close to 60 degrees as possible. The highest peak systolic velocity and end-diastolic velocity were calculated. According to the flow velocity criteria, severity of internal carotid artery stenosis was graded as <Рисунок: LE> 50% when the peak systolic velocity was <Рисунок: LE> 120 cm/s, between 50% and 80% when the peak systolic velocity was > 120 cm/s, and <Рисунок: GE> 80% when the peak systolic velocity was > 120 cm/s and the end-diastolic velocity was > 120 cm/s [1] .

Patients were then classified in four groups: no disease, slight, moderate, or severe carotid artery disease (see ). In patients with severe stenosis a preoperative carotid angiography was always performed.

To identify the presence of atherosclerotic lesions of the ascending aorta the preoperative chest roeutgengram and the coronary angiography were carefully evaluated. In addition, intraoperative digital palpation of the aorta was always performed following a described methodology [2] and the patients were classified in four groups according to the scale proposed by Mills and Everson [2] (see ): no disease, mild, moderate, or severe ascending aorta atherosclerosis. In 715 patients operated during the last part of the series intraoperative ultrasonographic examination of the ascending aorta was also performed and the atherosclerotic lesions were classified according to the criteria described by Wareing and coauthors [3] (see ).

Patients with a clinical history of peripheral vasculopathy were also submitted to examination of the iliac and common femoral arteries following the described Echo-Doppler methodology.

The preoperative clinical and angiographic characteristics of the patients of the three categories are shown in . [Table 1]

The moderate and high-risk patients were significantly older than low-risk patients. In the high-risk group there was also a greater incidence of diabetes, smoking history, unstable angina, peripheral vasculopathy, and previous cerebrovascular accidents. The incidence of severe aortic and internal carotid artery disease in relation to the age of patients is summarized in . [Figure 1] As expected, older patients had a significantly higher incidence of severe atherosclerosis of both the ascending aorta and the carotid arteries.

All the operations were performed by five surgeons using standardized techniques. Patients of the different categories were treated using different surgical strategies.

After median sternotomy and pericardiotomy, cardiopulmonary bypass was established by cannulating the right atrium and the ascending aorta. Normothermic systemic perfusion was used and myocardial protection was achieved by intermittent isothermic anteretrograde blood cardioplegia; patients were actively rewarmed if their nasopharingeal temperature fell below 34°C and mean nasopharingeal temperature in this group was 36.3 ± 0.8°C. Mean arterial pressure during cardiopulmonary bypass was maintained between 50 and 70 mm Hg; volume infusion or vasopressors (phenylephrine) were used in case of hypotension. Proximal and distal anastomoses were always performed during a single period of aortic cross-clamping.

These patients were also operated using the conventional technique described, however, due to the reported concerns on the neurologic safety of normothermic cardiopulmonary bypass and retrograde cardioplegia [4] [5] , patients of this group were always operated using mildly hypothermic systemic perfusion (28°C) and isothermic intermittent antegrade blood cardioplegia; mean nasopharingeal temperature in this group was 27.4 ± 0.9°C. In cases with moderate aortic disease care was paid to avoid the atherosclerotic areas during cannulation and cross-clamp.

We adopted dedicated surgical strategies for these patients. In particular, in patients with severe monolateral internal carotid artery stenosis simultaneous carotid endoarterectomy and coronary artery bypass was performed in case of unstable angina (n = 91), whereas a staged strategy (carotid endoarterectomy performed before myocardial revascularization) was adopted in patients with stable angina (86 patients). During simultaneous operations the carotid endoarterectomy was always carried out before median sternotomy; for both concomitant and staged procedures an intraluminal carotid shunt was used and myocardial revascularization was performed in condition of hypothermic systemic perfusion. Among patients who underwent staged operations the mean interval between the carotid and coronary operations was 14.9 ± 8.7 days. Patients with bilateral carotid artery stenosis and stable angina (n = 15) were treated by carotid endoarterectomy of the most diseased site followed by combined operation on the other carotid and the coronary arteries (performed within 2 to 3 weeks), whereas the patients with bilateral carotid disease and unstable cardiac symptoms (n = 4) were treated by simultaneous carotid endoarterectomy of the most severe lesion and coronary artery bypass, followed after 2 to 3 weeks by surgical treatment of the other carotid artery.

In patients with severely atherosclerotic ascending aorta we adopted the no-touch technique described by Mills and Everson [2] and Suma [6] when an arterial cannulation side (aortic arch or femoral artery) could be found (74 patients). In this instance hypothermic cardiopulmonary bypass was established between the right atrium and the aortic arch (n = 21) or the common femoral artery (n = 53) and a left ventricle vent was inserted through the right superior pulmonary vein. Aortic cross-clamp and cardioplegic delivery were avoided and ventricular fibrillation was induced electrically. The target vessels were occluded at the level of the anastomotic site using 5−0 Prolene (Ethicon Inc, Sommerville, NJ) or silicone elastomer stitches passed around the artery and pedicled arterial grafts were used. In 15 patients myocardial revascularization had to be completed with a saphenous vein proximally anastomosed to the innominate artery. When aortic arch and femoral cannulation was impossible (63 patients) the operation was performed without the use of cardiopulmonary bypass, following the technique described by Buffolo and colleagues [7] . In these patients all-arterial revascularization was performed using the in situ internal thoracic and gastroepiploic arteries and the radial artery proximally anastomosed to an internal thoracic artery graft at Y. In 7 patients an integrated approach (surgical grafting of the left anterior descending and subsequent percutaneous angioplasty of the other target vessels) had to be used due to the technical impossibility of using more than one arterial graft. The 11 patients in whom severe disease of the ascending aorta and internal carotid arteries coexisted had all unstable angina and underwent simultaneous carotid endoarterectomy and myocardial revascularization using either the no-touch or the no-pump technique (7 and 4 patients, respectively).

Main operative data of the different groups are reported in . [Table 2] As a direct consequence of our policy of privileging the reduction of the neurologic risk at expense of the completeness of the myocardial revascularization, the high-risk patients received a significantly lower number of grafts and had an higher rate of incomplete coronary grafting.

Postoperative neurologic complications were classified according to the definitions given in the ; patients with transient ischemic attacks were not considered in the data analysis.

Following a protocol in use at our institution since 1990, a neurologic evaluation was always performed by a cardiac anesthesiologist (blinded to the pre- and intraoperative evaluation) at the moment of the awakening of the patient from the anesthesia. In cases of neurologic abnormalities a complete neurologic examination was performed by a consultant neurologist immediately after the onset of symptoms, 24 hours later, at regular interval of 1 or more days (depending on the clinical status), and at discharge. In case of stroke the neurologic outcome was assessed using the Glasgow Outcome Scale [8] . The Glasgow Outcome score was assigned in concomitance with the neurologic evaluations by a neurologist and a cardiac anesthesiologist; disagreements were resolved by consensus after a common reevaluation. In patients with clinical evidence of neurologic complications of any type brain computed tomographic scans were performed immediately after the evidence of symptoms, 24 to 48 hours later, after any change in the neurologic status, and the day before discharge. All scans were performed prospectively and independently evaluated by two expert neuroradiologists; disagreements were resolved by common reexamination. In case of computed tomographic evidence of recent stroke, lesion size was measured by a ruler to the nearest half millimeter in two dimensions and multiplied by the thickness of the slice following a previously described methodology [9] [10] . For multiple or irregularly shaped lesions the sum of the single ischemic areas was calculated. All neurologic measurements were prospectively entered in a computerized database.

All patients were regularly submitted to clinical examination at our institution 1 and 6 months after operation and then every year thereafter. A stress myocardial scintigraphy was obtained in all patients 6 months postoperatively and every year thereafter. In case of clinical or instrumental suspect of residual myocardial ischemia or angina recurrence, reangiography was always proposed to the patient.

Statistical analysis

Data are expressed as mean value ± 1 standard deviation. For statistical analysis qualitative data were compared using the <Рисунок: chi> 2 test with Bonferroni's correction. Parametric quantitative data were compared by two-way analysis of variance; for p level less than 0.05 comparisons between groups were carried out using the unpaired t Student's test with Bonferroni's correction. Nonparametric quantitative data were compared by Friedman's test; for p level less than 0.05 comparisons between groups were carried out using the test Mann-Whitney U test with Bonferroni's correction. A p value less than 0.05 was considered significant.

Mortality and morbidity data of the patients of the different risk categories are summarized in . [Table 3] In the low-risk group the causes of death were stroke (8 patients), myocardial infarction (9), multiorgan failure (2), gastrointestinal complications and pulmonary embolism (1 patient each).

In the moderate-risk category the 6 deaths were due to stroke (2 patients), perioperative myocardial infarction, renal failure, untractable coagulopathy, and massive intestinal infarction (1 patient each), whereas among the high-risk patients the 9 deaths were due to myocardial infarction (3 patients), stroke (1), multiorgan failure (2), sepsis, hepatic insufficiency, and intestinal infarction (1 patient each).

No significant difference in mortality was found between the three groups; renal insufficiency was the only postoperative complications significantly more frequent in the high-risk population. Postoperative mortality and morbidity of patients of the high-risk group in relation to the different surgical strategy used are shown in . [Table 4]

Overall, 27 patients suffered a perioperative stroke (1.1%). Twenty-one of the 27 strokes occurred intraoperatively and 6 (22.2%) in the postoperative period. The stroke incidence was not significantly different between the three risk groups ; [Table 5] in fact, 18 events were reported in the low-risk group (1.1%; 15 intraoperative and 3 postoperative), five in the moderate-risk group (1.3%; 4 intraoperative and 1 postoperative), and four in the high-risk group (1.1%; 2 intra- and 2 postoperative).

The mean Glasgow Outcome Scale score and the computed tomographic extension of the brain lesion were similar in the three risk categories ( p = not significant).

In the moderate-risk group all four episodes of intraoperative stroke occurred in patients with moderate aortic disease (4 of 116, 3.4%; ), [Figure 2] whereas the only postoperative stroke was reported in 1 of the 25 patients of concomitant carotid and aortic disease (1 of 25, 4.0%); there were no case of perioperative stroke among patients with moderate carotid stenosis (n = 239).

The difference in incidence of intraoperative stroke between patients with aortic and carotid disease was statistically significant ( p < 0.01), whereas no differences was found with regard to the postoperative events.

In the high-risk group the two intraoperative stroke occurred in 1 of the 137 patients with severe ascending aorta atherosclerosis (0.7%) and in 1 of the 196 patients with severe carotid stenosis (0.5%), whereas the two postoperative events were reported in 1 patients with severe carotid disease (again 0.5%) and in 1 patients with severe aortic atherosclerosis (again 0.7%); there were no case of perioperative stroke among patients with concomitant carotid and aortic disease (n = 11). In this group no significant difference in the incidence of intra- and postoperative events between patients with [Figure 3] carotid and aortic disease could be demonstrated .

In the high-risk series the four strokes occurred in 1 patient with simultaneous carotid endoarterectomy and myocardial revascularization (1 of 95, 1.0%; intraoperatively), in 1 patient who underwent no-pump revascularization (1 of 63, 1.5%; postoperatively), in 1 patient operated using the no-touch technique (1 of 74, 1.3%; intraoperatively), and in one case of staged carotid endoarterectomy and myocardial revascularization (1 of 101, 0.9%; postoperatively) . [Table 6]

No cases of perioperative stroke were reported among patients who underwent simultaneous no-pump or no-touch and carotid procedures, carotid endoarterectomy followed by carotid and coronary operation, or staged carotid and coronary procedure followed by endoarterectomy of the other carotid artery.

Follow-up

            Clinical follow-up was 98.3% complete (2,253 of 2,290 patients). Mean follow-up time for the overall population was 35 ± 11 months without significant differences in length between the three groups (38 ± 12 months in the low, 39 ± 8 in the moderate, and 38 ± 10 in the high-risk group).

During the follow-up there were 37 noncardiac death (7 in the low, 16 in the moderate, and 14 in the high-risk group), mainly from cancer (15 patients, 4 in the low, 6 in the moderate, and 5 in the high-risk group) and cerebrovascular accidents (8 patients, 3 in the low, 3 in the moderate, and 2 in the high-risk group).

The rate of freedom from clinical or instrumental angina recurrence and cardiac death was significantly higher in the low and moderate-risk categories: 98.4% in low-risk patients (1,549 of 1,574), 98.0% in the moderate (351 of 358), and 85.3% in the high-risk cases (274 of 321) ( p < 0.01).

In 1993 we started this prospective study with the aim of evaluating the possibility of reducing the incidence of neurologic events in patients undergoing isolated coronary artery bypass grafting by means of a protocol of systematic pre- and intraoperative evaluation of the aorta and the carotid arteries and adoption of dedicated surgical strategies in cases of severe atherosclerosis of one or both vascular districts. Using this approach we achieved a 1.0% global stroke incidence and a rate of cerebrovascular complications in the high-risk group not significantly superior to that of the low-risk cases and considerably inferior to that expected on the basis of the current literature.

In fact, Roach and coauthors [11] in a recent large prospective multicenter trial found that perioperative cerebrovascular accidents occurred in 7.8% of patients with carotid disease versus 2.5% of cases without this characteristic and the Buffalo Cardiac Cerebral Study Group estimated a 6.01 odds ratio for postoperative stroke in patients with carotid stenosis <Рисунок: GE> 50% [12] . Likewise, McKhann and colleagues [13] reported perioperative cerebrovascular accidents in 15.5% of their patients in whom a carotid bruit was identified preoperatively and Reed and associates [14] showed in a case-control study that the presence of clinically diagnosed carotid pathology increases the risk of perioperative stroke or transient ischemic attack by 3.9−fold. On the other hand, Mickleborough and colleagues [15] reporting on 1,631 consecutive patients undergoing coronary artery bypass procedures reported a stroke incidence of 10% in patients with and 0.9% in patients without intraoperative evidence of atherosclerotic ascending aorta ( p < 0.01). Similarly, Gardner and co-workers [16] found that perioperative stroke occurred in 14% of patients with aortic disease compared to 3% of cases without evidence of diseased aorta and Lynn and associates [17] , using logistic regression analysis, identified aortic calcification as an highly significant risk factor for postoperative stroke in a series of 1,000 consecutive coronary patients. The low incidence of cerebrovascular events in all groups of our series (despite the fact that our protocol did not take into consideration any other risk factors for perioperative stroke) suggests that atherosclerotic disease of the ascending aorta and internal carotid arteries, besides being the only factors susceptible of surgical manipulation, is probably the major determinant of the stroke incidence in patients undergoing coronary artery bypass grafting.

A limitation of our study relies in the method used to evaluate the ascending aorta. In the majority of our patients the assessment of the aortic status was performed by intraoperative palpation, a technique that has already been shown to be quite insensitive in the demonstration of ascending aorta atherosclerosis [18] [19] . In fact, when analyzing the neurologic outcome of the moderate-risk population in relation to the type and severity of the atherosclerotic lesions, it is evident that all the cases of intraoperative strokes of this group occurred in patients with (presumed) moderate aortic atherosclerosis . It seems likely that in these patients a more severe disease of the ascending aorta has been unrecognized and the patients have been undertreated; the adoption of dedicated strategies even in these instances could have probably led to a further reduction of the rate of perioperative neurologic complications. For this reason we decided to use routine intraoperative echographic examination of the ascending aorta in the last part of our study.

The technical solutions that we applied have been extensively described and are part of the knowledge of all cardiac surgeons; however, we concentrate our effort on the careful evaluation of each case and on the systematic individualization of the surgical approach to the clinical and anatomic condition of every single patient. In this perspective some patients at high risk were treated using integrated techniques (surgical and percutaneous interventions) and, more in general, the completeness of surgical revascularization (once that the left anterior descending was protected by a mammary graft) was judged less important that the reduction of the neurologic risk. As a consequence, patients of the high-risk category received a significantly inferior number of grafts, had an higher rate of incomplete revascularization , and an inferior degree of freedom from ischemia at midterm follow-up. Whether the reduction of the perioperative neurologic risk outweighed the benefits of complete myocardial revascularization (especially in these older patients with reduced functional capacities) and whether an incomplete grafting limited to the most important vessels can be accepted in these instances is an important question that requires further extensive investigation. In this regard, it is important to underscore that a considerable part of the cases of ischemia recurrence in the high-risk group were detected only instrumentally (by means of stress myocardial scintigraphy) and affected only marginally the patients' quality of life (whereas a perioperative stroke would probably have more severe consequences on the patients' physical and social status). It is possible that in a near future the no-pump grafting of all the target vessels with pedicled arterial conduits will become a real possibility, and will represent the ideal surgical solution in this setting. However, at present technical and technological improvements are needed before a safe introduction of this strategy into the everyday clinical practice.

Our policy in dealing with combined monolateral carotid and coronary disease was to adopt simultaneous carotid endoarterectomy and myocardial revascularization in patients with unstable angina and a staged procedure (the carotid endoarterectomy performed before the coronary revascularization) in patients in more stable cardiac conditions. Patients with concomitant bilateral carotid artery stenosis and stable angina were treated by carotid endoarterectomy of the most diseased site followed by combined operation on the other carotid and the coronary arteries, whereas the patients with bilateral carotid disease and unstable cardiac symptoms were treated by simultaneous carotid endoarterectomy of the most severe lesion and coronary artery bypass, followed by surgical treatment of the other carotid artery.

This approach allowed a safe and effective treatment of concomitant carotid and coronary disease, leading to acceptable rates of perioperative stroke and myocardial infarction in all groups of patients .

In conclusion, our experience demonstrates that it is possible to achieve a low rate of perioperative cerebrovascular complications in high-risk patients undergoing isolated coronary artery bypass grafting with the use of an integrated protocol of systematic evaluation of the ascending aorta and internal carotid arteries and adoption of dedicated surgical strategies in case of atherosclerotic disease of one or both of these vascular districts. It is likely that in the next decade the adoption of this type of policy and a continuous effort to tailor the surgical strategy to the individual characteristics of the patient will play a key role in the necessary effort of providing acceptable risk procedures to an increasingly older population with a growing rate of associated noncardiac pathologies.

 

 

 

1.Roederer G.O., Langlois Y.E., Jager K.A., Lawrence R.J., Primozich J.T., Phillips D.J.. A simple spectral parameter for accurate classification of severe carotid disease. Bruit 1984;8:174−178.

2.Mills N.L., Everson C.T.. Atherosclerosis of the ascending aorta and coronary artery bypass. J Thorac Cardiovasc Surg 1991;102:546−553.

3.Wareing T.H., Davila-Roman V.G., Daily B.B.. Strategy for the reduction of stroke incidence in cardiac surgical patients. Ann Thorac Surg 1993;55:1400−1408.

4.Martin T.D., Craver J.M., Gott J.P.. A prospective, randomized trial of retrograde warm blood cardioplegia. Ann Thorac Surg 1994;57:298−304.

5.Guyton R.A., Mellitt R.J., Weintraub W.S.. A critical assessment of neurological risk during warm heart surgery. J Card Surg 1995;10:488−492.

6.Suma H.. Coronary artery bypass grafting in patients with calcified ascending aorta. Ann Thorac Surg 1989;48:728−730.

7.Buffolo E., Andrade J.C.S., Branco J.N.R.. Myocardial revascularization without extracorporeal circulation. Eur J Cardio-thorac Surg 1990;4:504−508.

8.Jennett J., Bond M.. Assessment of outcome after severe brain damage. Lancet 1975;1:480−484.

9.Petersen P., Madsen E.B., Brun B., Pedersen F., Gyldensted C., Boysen G.. Silent cerebral infarction in chronic atrial fibrillation. Stroke 1987;18:1098−1100.

10.Petersen P., Pedersen P., Johnsen A.. Cerebral computed tomography in paroxysmal atrial fibrillation. Acta Neurol Scand 1989;79:482−486.

11.Roach G.W., Kanchuger M., Mangano C.M.. Adverse cerebral outcome after coronary bypass surgery. N Engl J Med 1996;335:1857−1863.

12.Ricotta J.J., Faggioli G.L., Castilone A., Hassett J.M.. Risk factors for stroke after cardiac surgery. J Vasc Surg 1995;21:359−364.

13.McKhann G.M., Goldsborough M.A., Borowicz L.M.. Predictors of stroke risk in coronary artery bypass patients. Ann Thorac Surg 1997;63:516−522.

14.Reed G.L., Dinger D.E., Picard E.H., DeSanctis R.W.. Stroke following coronary artery bypass surgery. N Engl J Med 1988;319:1246−1250.

15.Mickleborough L.L., Walker P.M., Takagi Y., Ohashi M., Ivanov J., Tamariz M.. Risk factors for stroke in patients undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg 1996;112:1250−1259.

16.Gardner T.J., Horneffer P.J., Manolio T.A.. Stroke following coronary artery bypass surgery. Ann Thorac Surg 1985;6:574−581.

17.Gee W., Lynn G.M., Stefanko K., Reed J.F., Nicholas G.. Risk factors for stroke after coronary artery bypass. J Thorac Cardiovasc Surg 1992;104:1518−1523.

18.Barzilai B., Marshall W.G., Saffitz J.E., Kouchoukos N.. Avoidance of embolic complications by ultrasonic characterization of the ascending aorta. Circulation 1989;80:275−279.

19.Katz E.S., Tinick P.A., Rusinek H., Ribakove G., Spencer F.C., Kronzon I.. Protruding aortic atheromas predict stroke in elderly patients undergoing cardiopulmonary bypass. J Am Coll Cardiol 1992;20:70−77.

.