| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2006;82:515-523
© 2006 The Society of Thoracic Surgeons
Division of Cardiothoracic Surgery, Department of Surgery, University of Texas Medical Branch, Galveston, Texas
Accepted for publication March 14, 2006.
* Address correspondence to Dr Kollar, Department of Surgery, University of Texas Medical Branch, 301 University Blvd, Galveston, TX 77555 (Email: ankollar{at}utmb.edu).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentThe Society of Thoracic...References |
|---|
METHODS: A retrospective search in our hospital's medical record database was done to identify patients with postoperative strokes who underwent coronary artery bypass grafting operations from January 1, 1993, until December 31, 2004. All cases were individually reviewed, and the temporal relationship between neurologic event and postoperative episodes of AF was determined. During the study period it was our consistent policy to use only Coumadin anticoagulation limited to patients who had persistent AF or were to be discharged in AF.
RESULTS: Of the 2,964 coronary artery bypass grafting operations, 576 patients (19.4%) had AF and 32 patients (1.1%) suffered stroke. Seventeen stroke patients maintained normal sinus rhythm during their hospital stay. Of the remaining 15 patients, 9 presented with neurologic deficit before the first episode of AF, with 5 having intraoperative and 4 having postoperative stroke. Of the 6 patients with AF before neurologic event, three strokes occurred within 1 week after spontaneous conversion to normal sinus rhythm. One patient with preoperative and also with intraoperative AF who underwent emergency coronary artery bypass grafting woke up with stroke. In the remaining two cases, the AF or atrial flutter episodes lasted less than 6 hours each before the neurologic event. More aggressive anticoagulation as suggested in the published guidelines could not have prevented strokes in any of these 6 patients.
CONCLUSIONS: This retrospective analysis does not support the use of aggressive anticoagulation, particularly full intravenous heparinization as a bridging therapy to decrease the already low incidence of postoperative strokes after routine coronary artery bypass grafting surgery.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentThe Society of Thoracic...References |
|---|
Although many studies have confirmed that chemical prophylaxis successfully reduces the incidence of postoperative AF [10], there is no report whether this reduction had any effect on the stroke incidence. The American College of Cardiology/American Heart Association (ACC/AHA) guidelines [9, 14] strongly recommend early aggressive anticoagulation to prevent thromboembolic strokes, particularly when cardioversion is attempted. According to one of the largest prospective studies on postoperative AF, involving 70 institutions in 17 countries, 56.2% of patients in AF received intravenous heparin while only 17.6% of patients were started on Coumadin [8]; however, there are no data to suggest that such a practice could prevent postoperative strokes in any significant number.
In our institution, we do not use aggressive anticoagulation for postoperative AF, and we rarely use electrical cardioversion except for intubated patients on the intensive care unit and for those with significant hemodynamic compromise. Full heparinization is reserved for patients with pulmonary embolism or documented deep venous thrombosis, and Coumadin therapy is introduced gradually when the patients have had AF preoperatively or remain in new, continuous AF for more than 48 hours and they are expected to be discharged in AF. The aim of this study was to examine the temporal relationship between postoperative AF and cerebrovascular accidents after CABG operations and to determine whether any of the strokes could have been prevented by a more aggressive anticoagulation protocol.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentThe Society of Thoracic...References |
|---|
Preoperative Studies and Stroke Prevention in Coronary Artery Bypass Graft Patients
In our program, during the study period carotid duplex screening was done on patients with carotid bruits and when there were symptoms or history of cerebrovascular accidents or TIA. Symptomatic patients were also evaluated by head computed tomography and vascular surgery if more than 50% internal carotid artery stenosis was noted on carotid duplex. In case of concomitant disease and high-grade (>80%) internal carotid artery stenosis, we preferred staged procedures: carotid endarterectomy first, followed by CABG. After CABG surgery, all patients were started on 325 mg of aspirin 8 hours after arrival to the surgical intensive care unit unless excessive postoperative bleeding was recorded by means of the chest tubes.
Surgical Technique
During the study period, five surgeons performed all CABG operations using essentially the same surgical technique. In our institution, intraoperative transesophageal echocardiography (TEE) has become a routine procedure during the second half of the study period, and epiaortic echocardiography is used only when ascending aortic disease was suspected by palpation and TEE was not in use. However, we adopted the routine use of single aortic cross-clamping to minimize manipulations on the ascending aorta. The operations were done with cardiopulmonary bypass using moderate hypothermia and intermittent cold potassium blood cardioplegia except for a few patients in the second half of the study period who had operations without cardiopulmonary bypass. The right atrium was cannulated with a two-stage venous cannula, and during aortic cross-clamping the ascending aorta was vented by gravity. Left atrial or ventricular venting was not routinely used. Proximal anastomoses were constructed first, followed by the distal (side-to-side or end-to-side) vein and internal thoracic artery anastomoses. All patients received two right ventricular and two right atrial temporary pacing wires, and atrial or atrioventricular sequential pacing was instituted as deemed appropriate.
Postoperative Care and Arrhythmia Detection
Most patients were transferred from the intensive care unit to a surgical telemetry unit on the first postoperative day where surface electrocardiogram and atrial electrograms are monitored simultaneously until discharge. Any arrhythmia episodes are automatically printed and recorded in the chart by the nursing staff. Our routine protocol consists of oral ß-blocker prophylaxis unless contraindicated. Episodes of AF are treated for rate control with combined digoxin and ß-blocker or calcium-channel blocker if ß-blocker is not tolerated. Amiodarone, quinidine, or procainamide are only used occasionally. Electrical cardioversion is routinely used in patients intubated on the intensive care unit and in those with significant hemodynamic compromise. Atrial flutter is typically converted back to sinus rhythm or to AF with rapid atrial pacing. The pacing wires are removed 2 or more hours before discharge.
Our practice is to institute oral anticoagulation in only those patients who have persistent AF more than 48 hours and were to be discharged in AF. Full heparinization was not used because of AF, although some received venous thromboprophylaxis levels of heparin (5,000 U subcutaneously every 8 to 12 hours). Therapeutic prothrombin and international normalized ratio levels (international normalized ratio, 2.0 to 3.0) was frequently achieved only after discharge. These policies were part of a written protocol for perioperative care that was agreed on and used by each of the surgeons with cases during the study period.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentThe Society of Thoracic...References |
|---|
The above 32 patients with stroke or TIA represent our current database. There were 19 men and 13 women with a mean age of 61 years (range, 45 to 85 years). History and clinical data are presented in Table 1, and perioperative surgical data are presented in Table 2. Head computed tomography verification was done in all but one case (TIA only). Twenty-five patients had radiographic evidence of fresh ischemic infarct. Twelve patients woke up with neurologic deficit and were defined as having suffered an intraoperative stroke. The other 20 patients were neurologically intact after extubation and presented with sudden hemiparesis, focal deficit, or signs of encephalopathy between postoperative days 2 and 35. Four patients with focal neurologic deficit and no evidence of infarct recovered completely (TIA), and all other surviving patients were discharged in improved neurologic condition after a median postoperative length of stay of 14 days (range, 6 to 90 days).
|
|
Of the 12 patients with intraoperative strokes, 6 remained in normal sinus rhythm and 5 patients had subsequent AF without additional neurologic sequel. One patient required emergency CABG for acute myocardial infarction and had preoperative AF after temporizing percutaneous transluminal coronary angioplasty for which he was fully anticoagulated with heparin. He suffered an intraoperative stroke, and a TEE on postoperative day 3 confirmed left atrial appendage thrombus (intraoperative TEE was not available at that time).
|
|
The remaining 9 patients, then, are of greatest interest for this analysis: they had postoperative strokes or TIA (days 2 to 8) and AF or atrial flutter (Table 4). In 4 patients (2 had strokes on postoperative day 2 and 2 on postoperative day 3) the first episodes of atrial fibrillation occurred 12 hours to 3 days after the stroke. Of the remaining 5 patients, one (no. 7) developed intraoperative AF, which was electrically terminated, then atrially paced for 24 hours. On postoperative day 3 he had another episode of AF that converted to sinus rhythm within 6 hours, but he suffered a stroke the same evening. Transesophageal echocardiography showed a small atrial septal defect with bidirectional shunt suggesting paradoxical embolism. One patient (no. 22) developed atrial flutter on day 2 (successfully rapid paced into sinus rhythm) and had a first recurrence of flutter with similar conversion again on postoperative day 8 and a stroke the same day. One patient (no. 24) had four short episodes (30 minutes to 6 hours) of AF and was discharged in sinus rhythm on postoperative day 7. He was readmitted in normal sinus rhythm with symptoms of stroke and shortness of breath and was confirmed to have a pulmonary embolism and a patent foramen ovale with a right-to-left shunt. One patient (no. 25) had three short episodes (30 minutes to 4 hours) of AF and was discharged home in sinus rhythm on postoperative day 6. This patient had documented normal sinus rhythm on return clinic visits (postoperative day 17 and postoperative day 31) and also at readmission to the hospital with a massive stroke and right internal carotid artery occlusion (postoperative day 35 after CABG). The last patient (no. 27) again had three short episodes (30 minutes to 4 hours) of AF until postoperative day 4. This patient's postoperative recovery was slower than usual, but he had no neurologic deficit until he had a sudden respiratory arrest on postoperative day 13. His cardiac recovery was complete (no more AF episodes), but he suffered bilateral anoxic brain damage and subsequently died.
In our series no patient with neurologic event after postoperative AF had a long enough AF episode to warrant aggressive anticoagulation suggested by the ACC/AHA guidelines.
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentThe Society of Thoracic...References |
|---|
The pathophysiology of postoperative AF is not completely understood. Apart from obvious comorbid conditions such as valvular heart disease, atrial enlargement, congestive heart failure, and history of preoperative atrial arrhythmias [1–3, 5], several other risk factors predispose cardiac surgical patients for postoperative AF. Advanced age is the strongest, followed by systemic hypertension, left ventricular hypertrophy, peripheral vascular disease, and chronic lung disease [1–6, 8]. Longer cardiopulmonary bypass time and aortic cross-clamp time have been shown to be associated with increased incidence of postoperative AF [1, 15, 16]. Similarly, various surgical manipulations on the atria (eg, left atrial venting [2, 3]), the need for prolonged inotropic or mechanical support, and the patient's own adrenergic status [5, 7, 8] appear to have an impact, although it is not clear whether these factors are true independent risk factors or they are just representative of sicker patients. Postoperative pericardial fluid collection and pericarditis have also been associated with atrial arrhythmias [17, 18], whereas technical modifications, such as the preservation of anterior epicardial fat pad, have been found to decrease the overall incidence of AF [19].
In nonsurgical patients, the relationship between chronic AF and thromboembolic stroke has been extensively studied and well established [14]. Chronic AF leads to decreased blood flow velocities, particularly in the left atrial appendage, with consequent thrombus formation and systemic embolization [20, 21]. Because the economic and health care implications of stroke are enormous, there have been many efforts to reduce stroke rates in these patient groups. It is now universally accepted that long-term anticoagulation with Coumadin significantly decreases the risk of strokes in higher risk medical patients [14]. However, studies that risk-stratified these patients have led to accepted guidelines that have defined lower risk patients with chronic AF in whom the relative risks do not favor Coumadin anticoagulation [11–13, 22–25]. On the basis of these studies, for younger patients (<65 years of age) with no or minimal other risk factors who have an annual stroke risk of less than 1.5% per year, aspirin alone is considered sufficient, and the risk of Coumadin-related bleeding may actually outweigh the benefit of therapy [22].
It is not entirely evident whether cardioversion and rhythm control is a successful strategy to prevent strokes. In the AFFIRM study [26] performed in the United States and in another study [27] from the Netherlands, medical patients with chronic AF were randomized into rate control plus anticoagulation versus rhythm control plus anticoagulation groups. After 2 to 3 years of follow up, a similar number of strokes occurred in both groups. Altogether the rhythm control strategy offered no survival advantage, and the rate control group had advantages including less frequent need for hospitalization and fewer adverse reactions to antiarrhythmic drugs.
How can these data be applied to our CABG patient population? The combined ACC/AHA/European Society of Cardiology practice guidelines (2001) for AF [14] recommend anticoagulation with heparin or oral anticoagulation in general and continuous heparin anticoagulation when cardioversion is attempted if postoperative AF persists beyond 48 hours. These guidelines also make the following statement regarding nonsurgical patients who are orally anticoagulated for chronic AF: "It is the consensus of the writing group that anticoagulation may be interrupted for a period of up to 1 week for surgical or diagnostic procedures that carry high risk of bleeding without substituting heparin" [14]. The recently published ACC/AHA 2004 guideline update for CABG surgery [9] recommends warfarin anticoagulation if the arrhythmia persists beyond 24 hours; however, in the subsequent text there is further speculation that "an aggressive anticoagulation and cardioversion philosophy may reduce the neurological complications associated with this arrhythmia," and "it may be advisable to use intravenous heparin."
Besides the obvious discordance between those guidelines, it is not quite clear why a new-onset AF occurring in the early postoperative period with its self-limited and intermittent nature represents such a high stroke risk that an aggressive anticoagulation protocol is advised within 24 hours, whereas medical patients with chronic AF undergoing noncardiac operations can be left without anticoagulation for up to 1 week without significant risk of stroke.
The cardiac surgical literature has been relatively consistent on this question, considering postoperative AF as an increased stroke risk [1–4, 7, 9]. However, in a recent single institutional study involving more than 9,000 patients, AF dropped out as a risk factor for stroke on multivariate analysis [6], lending doubt regarding the strength of this relationship. It is surprising, though, that there is only one paper in the entire English literature on this topic discussing temporal relationship between these two entities. Lahtinen and colleagues [28] from Finland analyzed data of 52 stroke patients after CABG operation and found that in 19 patients (36%) the first AF episode preceded the development of stroke by a mean of 21.3 hours (average, 2.5 AF episodes before stroke). The stroke was attributed to calcification in the ascending aorta in 13 patients (25%), and 16 patients (31%) had greater than 70% internal carotid artery stenosis. The study does not state the overall incidence of AF in their CABG population or their routine arrhythmia prophylaxis and management. Additional variables were preoperative stroke or TIA (21 cases), significant valvular disease (18 cases), and history of preoperative AF (5 cases), and 8 of the above AF 19 cases suffered a stroke in spite of anticoagulation. Similarly no information was given on other potential intraoperative risk factors, such as left atrial vent placement, that might be relevant in surgical patients.
A recent analysis from the Texas Heart Institute Cardiovascular Research Database [7] compared 994 CABG patients with postoperative AF and 5,481 patients without arrhythmia. Atrial fibrillation was found to be an independent predictor of long-term mortality at 5 years (adjusted odds ratio, 1.5). It was also associated with greater in-hospital mortality (odds ratio, 1.7), with more perioperative strokes (odds ratio, 2.02), and with prolonged hospital stay (14 versus 10 days). However, patients who developed AF were older, more often hypertensive, had chronic obstructive pulmonary disease, noncoronary vascular disease, congestive heart failure, and more severe underlying coronary artery disease, all of which are strong predictors for stroke without surgery. They were also likely to have had an intraaortic balloon pump placed and longer cardiopulmonary bypass time. Although the authors performed a case-matched subanalysis to strengthen their argument, the adequacy of the adjustment was limited and therefore the reported associations may not be truly independent, as pointed out in the editorial comment [29]. Also no temporal relationship between AF and stroke was reported in this series.
The Multicenter Study of Perioperative Ischemia Research Group and investigators of the Ischemia Research and Education Foundation have recently published their prospective study performed in 70 hospitals on 4 continents [8]. This study included more than 5,000 patients undergoing CABG operations with or without valve surgery on cardiopulmonary bypass. Patients with postoperative AF were significantly older (67.8 years versus 61.8 years), and a significantly larger number had history of AF (14.6% versus 6.0%), valvular disease (27.8% versus 14.9%), congestive heart failure (40.1% versus 32.0%), chronic obstructive pulmonary disease (14.0% versus 8.6%), and prior neurologic event (13.0% versus 9.3%). The overall incidence of postoperative AF was 32.3%, with 43% of patients having more than one episode and 22% having more than two episodes. The overall stroke rate for patients without AF was 1.2%, whereas it was 0.93% in patients with one episode of AF and 1.4% in patients with more than one episode of AF. The incidence of composite neurologic outcomes (stroke, encephalopathy, and stroke score changes postoperatively) was significantly higher for patients with more than one episode of AF as compared with patients with normal sinus rhythm, but not for patients with only one episode of AF. Patients with more than one episode of AF had an overall higher mortality (4.7% versus 2.1%) than patients without AF, and composite complication outcome was also significantly higher (22.6% versus 15.4%, respectively). In this study, 56.2% of all patients with AF were started on intravenous heparin, and 17.6% received Coumadin, although the effectiveness of anticoagulation and its effect on neurologic outcomes was not analyzed. The authors conclude that sicker patients develop AF more often and they also have a higher incidence of other postoperative complications, which, however, did not seem to apply to patients with only one episode of postoperative AF.
The Northern New England Cardiovascular Disease Group analyzed more than 11,000 CABG patients' data from 1996 to 2001 [30] and found AF was a positive predictor for stroke (odds ratio, 1.82). Interestingly, however, prolonged cardiopulmonary bypass time (longer than 114 minutes; odds ratio, 2.36) and prolonged inotropic agent use (odds ratio, 2.59), which had previously been established as predisposing factors for AF, were stronger predictors for stroke than AF by itself. In spite of their large database, a temporal relationship could only be determined in less than one third of the stroke patients, but even those cases were unconfirmed. This same group also analyzed the etiologic mechanism of stroke in 388 patients operated on from 1992 to 2000 [31, and reported that almost two thirds of all cerebrovascular events occurred within 2 days of the operation and 38% were classified as nonembolic. In these patients even early anticoagulation with intravenous heparin could not have prevented stroke.
Our 12-year surgical experience is limited compared with large institutional or pooled data and includes slightly more than 2,900 CABG cases with 19.4% AF incidence, and a stroke incidence of 1.1% (32 patients). We have not performed a detailed retrospective risk factor analysis, but according to the Society of Thoracic Surgeons database yearly reports, our overall risk profile is comparable to the national database population. Our telemetry monitoring system, however, includes simultaneous continuous monitoring of atrial electrograms and surface electrocardiograms, enabling us to identify the exact nature and onset of postoperative arrhythmias more precisely, particularly supraventricular arrhythmias, and to reliably detect all episodes of AF. Therefore, we believe that our recorded AF incidence is accurate, with atrial arrhythmias rarely if ever missed during hospitalization. Moreover, the prompt identification of the nature of the arrhythmia allowed us to establish a temporal relationship between these episodes and the neurologic complication in every case.
According to our initial analysis the overall incidence of stroke in patients with AF was higher (2.6%), than in patients who maintained normal sinus rhythm (0.7%) during their hospital stay. However, of the 15 patients who had both stroke and AF, 9 patients had their neurologic event before the first episode of arrhythmia, with 6 having intraoperative stroke. Of the 6 patients who experienced atrial arrhythmia before stroke (1.0% true incidence of stroke in AF patients), 1 was an emergency case with preoperative myocardial infarction and AF episodes who developed left atrial appendage thrombus in spite of heparin anticoagulation and suffered an intraoperative stroke (case 6). Three (nos. 24, 25, and 27) strokes occurred more than 1 week after the patients were successfully converted and stayed in normal sinus rhythm (no indication for anticoagulation). The last 2 patients (nos. 7 and 22) had preoperative history of TIA without significant carotid disease; both of these patients had two to three short episodes (<6 hours' duration) of AF or atrial flutter with successful conversion each time and again no indication (arrhythmia lasting more than 24 to 48 hours) for intravenous heparinization. Our policy of routine ß-blocker plus aspirin plus rate control for AF episodes, and delayed oral anticoagulation (started only after 48 hours of persistent or recurrent AF) yielded no strokes that could have been prevented with more aggressive anticoagulation as suggested by the ACC/AHA practice guidelines.
From our clinical experience and the literature review presented above, we suggest that there is insufficient evidence indicating a high enough incidence of strokes as a direct result of AF after CABG to justify early aggressive anticoagulation, particularly full heparinization as bridging therapy in the recently operated patients. The only data not consistent with this view are the previously described study from Finland. In that study the average time between the first AF episode and the neurologic event was 21.3 hours, less than the ACC/AHA recommended time interval to start anticoagulation for postoperative AF. In other words, these neurologic episodes may not have been preventable by anticoagulation.
We conclude that the development of AF, particularly recurrent AF in spite of routine ß-blocker prophylaxis, may simply identify sicker patients with a higher stroke risk profile, and the apparent association between these two entities may be in large part a parallel phenomenon similar to medical patients. Furthermore, there is no clear evidence to suggest that aggressive chemical or electrical cardioversion and prompt anticoagulation with intravenous heparin could prevent strokes in any significant number. Many patients with postoperative AF have an overall low risk profile for thromboembolic stroke, and considering the self-limiting nature of the postoperative arrhythmia it seems appropriate to treat these patients similar to medical patients, with aspirin alone. There is no convincing evidence to suggest that early aggressive anticoagulation, particularly intravenous heparinization as bridging therapy, will decrease the already low incidence of postoperative strokes. On the contrary, there is evidence that early postoperative heparin administration increases the risk of bleeding complications [32].
It is our opinion that if continuous AF persists, lasting more than 48 hours, and the patient is expected to remain in AF, it seems appropriate to start Coumadin, and therapeutic oral anticoagulation (international normalized ratio. 2.0 to 3.0) could be achieved gradually in an outpatient setting without an increased risk of cardiac tamponade. Such a policy should not and in our series has not inordinately delayed hospital discharge.
|
The Society of Thoracic Surgeons Policy Action Center |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentThe Society of Thoracic...References |
|---|
This website is now available at www.sts.org/takeaction.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentThe Society of Thoracic...References |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
