Author + information
- Received December 17, 1996
- Revision received July 21, 1997
- Accepted August 12, 1997
- Published online November 15, 1997.
- ↵*Dr. Brian Olshansky, Section of Electrophysiology, Division of Cardiology, Loyola University Medical Center, 2160 South First Avenue, Maywood, Illinois 60153.
Objectives. We sought to evaluate the risk of thromboembolic events in the presence of chronic atrial flutter and to determine the impact of anticoagulation therapy, if any, on this risk.
Background. Thromboembolic events are thought to be rare after cardioversion of atrial flutter.
Methods. This study was a retrospective analysis of 110 consecutive patients referred to the electrophysiology laboratory for cardioversion of chronic atrial flutter from 1986 to 1996. Atrial flutter was present for at least 6 months. Of the 110 patients reviewed, 100 had adequate information available regarding the effectiveness of anticoagulation (mean age 64 years, range 27 to 86; 75 men, 25 women; mean left ventricular ejection fraction 42%).
Results. Thirteen patients (13%) had a thromboembolic event. Of these, seven were attributable to causes other than atrial flutter. In the remaining six patients (6%), thromboembolic events occurred during a rhythm of atrial flutter or after cardioversion to sinus rhythm. Other causes of thromboembolism were excluded. Effective anticoagulation was associated with a decreased risk of thromboembolism (p = 0.026).
Conclusions. Patients with chronic atrial flutter are at an increased risk of thromboembolic events. Effective anticoagulation may decrease this risk.
Atrial flutter was first described almost 90 years ago. It is a frequently observed cardiac arrhythmia, almost as common as all other regular supraventricular tachycardias combined . Of those patients admitted with a diagnosis of supraventricular tachycardia between 1985 and 1990, 77% had atrial fibrillation and 10% had atrial flutter . In 1990, 21,000 patients were discharged with a primary diagnosis of atrial flutter, whereas 86,000 patients had atrial flutter as one of seven discharge diagnoses .
Atrial flutter is usually symptomatic, with the most common symptom being palpitations. It is often less well tolerated than atrial fibrillation . This may be due to the abrupt, striking increase in the ventricular response rate that can occur with little physical exertion with atrial flutter. Atrial flutter may cause hypotension, angina or congestive heart failure depending on the ventricular response rate, persistence of the atrial flutter and underlying ventricular function. After return to sinus rhythm, immediate improvement in symptoms and ventricular function are possible [5, 6]. Tachycardia-mediated cardiomyopathy, which can result from chronic atrial flutter, may completely resolve over time [7–9]. Cardioversion is therefore often recommended.
After cardioversion of atrial flutter, thromboembolic events are thought to be rare and are less likely to occur than after cardioversion of atrial fibrillation. Anticoagulation has therefore been considered unnecessary . Surprisingly, the true risk of clinically apparent thromboembolic events occurring in patients with chronic atrial flutter has not been addressed carefully. The major clinical trials of anticoagulation therapy for atrial fibrillation excluded patients with atrial flutter [11–16]. In several small published trials of cardioversion, patients with atrial fibrillation, atrial flutter and supraventricular tachycardias were pooled for analysis [17–20], making interpretation difficult.
To further evaluate the risk of thromboembolic events in the presence of chronic atrial flutter, we retrospectively reviewed the results of all patients referred to the electrophysiology laboratory for cardioversion of chronic atrial flutter from 1986 to 1996.
This study was a retrospective analysis of 110 consecutive patients referred to the electrophysiology laboratory for cardioversion of chronic atrial flutter from January 1986 through May 1996. Of the 110 patients reviewed, 100 had adequate information available regarding the effectiveness of anticoagulation, documented persistence of atrial flutter and underlying heart disease to be included in the study. Twenty-eight patients had a history of cardioversion before referral. All patients included were successfully converted to normal sinus rhythm with either overdrive right atrial pacing (16%) or direct current cardioversion (84%). Thirty-six patients underwent repeat cardioversion for recurrent atrial flutter during the study period. Atrial flutter was the predominant rhythm for a minimal duration of 6 months, but was present up to 5 years before presentation. Sixty-five patients had coronary artery disease (CAD); 17 patients had valvular heart disease; 10 patients had other forms of heart disease; and 10 patients had no structural heart disease. The mean left ventricular ejection fraction was 42%.
1.2 Study Protocol
1.2.1 Definition of Cardiac Rhythm
Atrial flutterwas defined as an organized atrial tachycardia at a rate of ≥240 beats/min. Using established criteria, atrial flutter was diagnosed by the 12-lead electrocardiogram (ECG) in all patients included in this study [4, 21, 22].
1.2.2 Definition of Effectiveness
Anticoagulation was considered effectiveif the international normalized ratio (INR) was ≥2.0 and the prothrombin time (PT) was ≥16 s consistently. Determination of the effectiveness of anticoagulation was made based on all PT/INR values obtained in the 4 weeks before cardioversion. Any subtherapeutic values obtained during this time placed the patient in the ineffectively anticoagulated group. In the case of spontaneous thromboembolism, the last PT/INR value obtained before presentation, as well as the PT/INR value obtained immediately after the events, was evaluated in a similar manner. Some patients considered to be effectively anticoagulated before cardioversion had episodes of ineffective anticoagulation at times other than those specified previously. Some patients were not anticoagulated at all. These patients were included in the group of ineffectively anticoagulated patients for the purpose of this study.
1.2.3 Thromboembolic Events
The determination of a thromboembolic event was based on clinical findings and supported by radiologic studies, including a computerized axial tomographic scan of the head and cerebral angiography. All patients who experienced a thromboembolic event underwent a screening two-dimensional echocardiogram and carotid Doppler studies, as is the usual practice in our institution.
1.3 Data Collection
Various clinical variables were assessed, including age, gender, anticoagulation status, presence and type of heart disease, ejection fraction, history of hypertension (HTN), history of diabetes, previous cardioversion, episodes of transient atrial fibrillation and thromboembolic events. Clinical examination, diagnostic evaluation and the results of pertinent tests performed after any thromboembolic events were also noted. Persistence of flutter was documented by evaluation during patient examinations, ECG recordings and patient history. Episodes of transient atrial fibrillation were documented by ECG or Holter recordings. No long-term rhythm monitoring was performed in any patient involved in the study.
1.4 Statistical Analysis
Data were analyzed using the two-tailed Fisher exact test. Patients who had effective anticoagulation were compared for risk of thromboembolism with those who had ineffective anticoagulation using the definitions previously described.
One hundred patients were studied; fifty four received effective anticoagulation; 30 had ineffective anticoagulation; and 16 had no anticoagulation before cardioversion or the thromboembolic event, or both.
Patients were grouped by anticoagulation status into two subgroups. The first group included patients with effective anticoagulation before cardioversion or the thromboembolic event, or both. The second group included patients with either ineffective or no anticoagulation before cardioversion or the thromboembolic event, or both. A comparison of the baseline characteristics of the patients in the effective, ineffective and no anticoagulation groups are presented in Table 1. These groups were similar with respect to age, gender, left ventricular ejection fraction and history of HTN. Notably, the groups differed with respect to the presence and type of cardiac disease. More patients with valvular heart disease received effective anticoagulation; more patients without heart disease had ineffective or no anticoagulation. The implications of this disparity are addressed subsequently. Four patients from the effective anticoagulation group and two from the ineffective or no anticoagulation group had transient atrial fibrillation detected during the study period.
2.1 Thromboembolic Events
Thirteen patients experienced a thromboembolic event during the study. Effective anticoagulation was associated with a significantly decreased risk of thromboembolic events (p = 0.026 by two-tailed Fisher exact test). No thromboembolic events occurred in the 54 patients who had effective anticoagulation (Table 2). Seven patients had thromboembolic events that were probably not related to atrial flutter (Table 3). In these patients, either the thromboembolic events occurred before the onset of atrial flutter or another likely source of embolism was detected.
Three patients developed clinical manifestations of a cerebrovascular accident (CVA) before the documented onset of atrial flutter. A fourth patient was found to have significant carotid artery atherosclerotic disease documented by carotid Doppler studies after a retinal artery embolus. A left ventricular thrombus was detected on transthoracic echocardiography after the thromboembolic event in the remaining three patients.
For the remaining six patients with clinically apparent thromboembolic events, atrial flutter was considered the most probable causal factor. In all cases, carotid Doppler studies and transthoracic echocardiograms were unremarkable for a source of emboli. All CVAs were verified by a computed tomographic scan.
2.1.1 Patients With Thromboembolic Events Attributed to Atrial Flutter
patient1 was a 27-year old woman with hypertrophic cardiomyopathy. Coumarin therapy was withheld before radiofrequency catheter ablation of atrial flutter. Forty-eight hours after successful ablation, the patient had a pulmonary embolism verified by pulmonary angiography. Venographic studies of the lower extremities revealed no evidence of thrombosis.
patient2 was an 84-year old woman without a history of cardiac disease who developed two transient ischemic attacks during a period of ineffective anticoagulation.
patient3 was a 68-year old woman with rheumatic heart disease who had ineffective anticoagulation before cardioversion of atrial flutter. The patient developed amaurosis fugax 24 h after cardioversion to sinus rhythm. An ophthalmologic examination revealed a retinal artery thromboembolus.
patient4 was a 59-year old man without cardiac disease who had no anticoagulation for chronic atrial flutter. The patient had a right basal ganglia cerebral infarct within 48 h after cardioversion to sinus rhythm.
patient5 was a 66-year old man with a history of ischemic cardiomyopathy and chronic atrial flutter. The patient had a right temporal lobe CVA during a period of ineffective anticoagulation.
patient6 was a 58-year old man with CAD, a left ventricular ejection fraction of 35% and chronic atrial flutter. He had a right occipital-temporal lobe CVA during a period of ineffective anticoagulation.
On the basis of our findings, it appears that chronic atrial flutter is associated with a remarkably high risk of clinically apparent thromboembolism. Effective anticoagulation appears to reduce this risk.
Considering the common occurrence of atrial flutter, surprisingly little information is available to determine the true risk of a thromboembolic event. As noted previously, the data from most published studies have suggested that the incidence of thromboembolism in atrial flutter is low [10–20].
Arnold et al. reported no embolic events in 122 cardioversions of patients with atrial flutter, 26% of whom received anticoagulation. It was suggested that patients with atrial flutter do not need anticoagulation. The reason our results differed from those of Arnold et al. may lie in the nature of the groups studied. Our study included only patients with chronic atrial flutter and excluded those with recent onset flutter and postoperative atrial flutter. In the study of Arnold et al. , patients with postoperative fibrillation and flutter represented the majority (67%) of the study group. Also, most patients included in this study (73%) had acute or recent onset of atrial arrhythmia, with a duration <30 days. This difference in etiology and duration of atrial flutter may be responsible for the differences observed in the risk of thromboembolism between the two studies. Patients with postoperative and acute or recent onset atrial flutter may have a different and quite possibly lower risk of thromboembolism than those with chronic atrial flutter.
Chalasani et al. reviewed 98 direct current cardioversions of atrial flutter in 85 patients. Only 12% of the patients had received anticoagulation before cardioversion. No thromboembolic events were reported up to 24 h after cardioversion. It is unknown whether any patient developed thromboembolic complications >24 h after cardioversion, as was observed in two patients in our study. Also of note, the duration of atrial flutter was unknown in 73 of 98 episodes analyzed and was not reported in the remaining 25 episodes. Thus, the applicability of this study to patients with chronic atrial flutter is unknown.
Thromboembolic risk is thought to be low in atrial flutter owing to the organized atrial contraction that may occur. This synchronous activity is presumed to prevent blood stasis and thrombus formation. Atrial function can be altered during atrial flutter or after cardioversion to sinus rhythm, similar to atrial fibrillation. Jordaens et al. showed that atrial systolic function can be diminished in atrial flutter and that the return of atrial function can be delayed for days to weeks after cardioversion. Atrial contribution to ventricular filling was measured using echocardiographic Doppler-estimated diastolic transmitral flow in 22 patients who were converted to sinus rhythm using either direct current shock or rapid pacing. Atrial contribution to ventricular filling was found to be minimal or absent after conversion to sinus rhythm. Atrial dysfunction lasted up to 2 weeks after cardioversion. As atrial dysfunction occurred with either direct current cardioversion or atrial pacing, “stunning” due to direct current cardioversion did not appear to be the cause. Rather, atrial dysfunction appeared to be due to the atrial flutter itself.
O’Neill et al. found that the return of atrial mechanical function after cardioversion is variable and often delayed. They reported that a decrease in atrial mechanical function could cause stasis of blood and formation of atrial thrombi.
Transesophageal echocardiography has greatly enhanced the ability to identify patients with atrial arrhythmia who are at greater risk of thromboembolism . Spontaneous echocardiographic contrast medium (“smoke”) and left atrial appendage thrombi have been reported by transesophageal echocardiographic studies in up to 35% of nonanticoagulated patients with atrial flutter [28, 29]. Spontaneous echo contrast medium, thought to represent stasis of blood in the atrium, has been associated with an increased risk of thromboembolism [30, 31].
Another risk factor for thromboembolic events in patients with chronic atrial flutter is the potential for transition, possibly undiagnosed, between atrial fibrillation and atrial flutter. Because atrial fibrillation has a well established risk of thromboembolism, conversion from flutter to fibrillation may heighten the risk of thromboembolism. The alternation between atrial fibrillation and flutter may be asymptomatic and can occur in up to 25% of patients with documented chronic atrial flutter [32, 33]. Therefore, some investigators recommend Holter monitoring of all patients with chronic atrial flutter before cardioversion to exclude intermittent transitions to atrial fibrillation .
Altered mechanical function of the atria and oscillation between atrial flutter and fibrillation may explain the development of atrial thromboemboli in patients with chronic atrial flutter. The mechanisms responsible for thromboembolism in chronic atrial flutter are, however, most likely multiple and complex.
In our study, chronic atrial flutter was associated with a substantial risk of thromboembolism, with 6 of 100 patients developing a thromboembolic event attributed to atrial flutter. This surprisingly high rate of thromboembolism has not been previously reported in a large series of patients with atrial flutter. No thromboembolic events were observed in the effectively anticoagulated group. The fact that, at presentation of thromboembolism, patients had subtherapeutic anticoagulation or were not anticoagulated strongly supports a cause and effect relation. This appears to confirm a protective benefit of anticoagulation therapy in patients with chronic atrial flutter.
As noted previously, most of the patients with valvular heart disease (16 of 17) had effective anticoagulation, whereas a majority of those patients without underlying heart disease (7 of 10) received no anticoagulation. This noted disparity may be a reflection of our clinical practice, which has been influenced by previous reports [19, 35]that have placed patients with valvular heart disease and atrial arrhythmias at particularly high risk for thromboembolism.
The task of determining the efficacy of anticoagulation therapy is difficult. Most if not all patients with long-term anticoagulation therapy have some periods of “subtherapeutic” PT/INR values. Patients who experienced thromboembolic events were placed in the ineffective anticoagulation group if their PT/INR value was subtherapeutic at the time of presentation. It is very possible that many patients considered to have “effective anticoagulation” had undetected periods of ineffective anticoagulation not associated with an event. Because these patients would have been considered to have “effective” anticoagulation, our data would therefore underestimate ineffective anticoagulation. Even assuming this to be the case, an overall high incidence of thromboembolism in patients with chronic atrial flutter was noted in our study.
Our data are comparable to those of Wood et al. , who reviewed 86 patients referred for radiofrequency catheter ablation of chronic atrial flutter. In their report, 6 (7%) of 86 patients were found to have a history of thromboembolism after cardioversion. The risk of thromboembolism in chronic atrial flutter may be underestimated and is likely much higher than previously reported.
3.1 Study Limitations
Our study was a nonrandomized, retrospective analysis of patients referred to a university hospital electrophysiologic laboratory for treatment of chronic atrial flutter. This is reflected in the high frequency of cardiac disease, HTN and diabetes observed in this group. It is possible that referral bias may have resulted in a patient group at greater risk of thromboembolic complications than the “typical” patient group with atrial flutter.
Because the study was retrospective, the duration of anticoagulation, the frequency of PT/INR determination and the therapeutic range “targeted” were determined by the referring physicians caring for these patients. Our determination of the efficacy of anticoagulation therapy was therefore limited by the data available on individual patients.
The study is also limited by the ability of the investigators to establish the source of thromboembolism. It is possible that the actual cause of thromboembolism was undetected in the six cases attributed to atrial flutter. As noted previously, it is not uncommon for atrial flutter to convert spontaneously and often symptomatically to atrial fibrillation. Alternatively, it is possible that atrial flutter may have been the true source of thromboembolism in any of the seven patients in whom thromboembolism was attributed to other causes. In either case, the incidence of thromboemboli in patients with chronic atrial flutter who did not receive adequate anticoagulation is high.
These data suggest that patients with chronic atrial flutter are at a substantially increased risk of thromboembolism. These thromboembolic events can occur after cardioversion to sinus rhythm or after spontaneous conversion to either sinus rhythm or atrial fibrillation. Effective anticoagulation was associated with a decreased risk of thromboembolism in this study (p = 0.026). We therefore recommend full anticoagulation for chronic atrial flutter until large-scale, prospective, randomized trials become available.
We thank Padmavathi Pagadala, MD for assistance.
☆ This study was supported financially by Loyola University.
- coronary artery disease
- cerebrovascular accident
- electrocardiogram, electrocardiographic
- international normalized ratio
- prothrombin time
- Received December 17, 1996.
- Revision received July 21, 1997.
- Accepted August 12, 1997.
- The American College of Cardiology
- ↵Jolly WA, Ritchie WJ. Auricular flutter and fibrillation. Heart 1910–1911;2:177–221.
- Bialy D,
- Lehmann MH
- Graves EJ
- Resenkov L
- Kliev JR,
- Sacrez A,
- Sacello A,
- et al.
- Stroke Prevention in Atrial Fibrillation Study Group Investigators
- Connaly SJ,
- Laupacis A,
- Gent M
- Lown B
- Navab A,
- La Due JS
- Wells JL Jr.,
- MacLean WAH,
- James TN
- Waldo AL,
- MacLean WAH,
- Karp RB
- Arnold AZ,
- Mick MJ,
- Mazurek RP,
- Loop FD,
- Trohman RG
- Black IW,
- Hopkins AP,
- Lee LC,
- Walsh WF
- Irani WN,
- Wilett DL,
- Grayburn PA,
- Brickner E,
- Afridi I
- Fatkin D,
- Kelly RP,
- Feneley MP
- Wood K,
- Eisenberg S,
- Kalman J,
- et al.