Thromboembolic complications after electrical cardioversion in patients with atrial flutter

CLINICAL STUDIES

Thromboembolic Complications after ElectricalCardioversion in Patients with Atrial Flutter

Abdou Elhendy, MD, PhD, Federico Gentile, MD, Bijoy K. Khandheria, MD,Bernard J. Gersh, MD, Kent R. Bailey, PhD, Samantha C. Montgomery, MSc,

James B. Seward, MD, A. Jamil Tajik, MD

PURPOSE: To determine the incidence of thromboemboliccomplications after cardioversion in patients with atrial flutter.SUBJECTS AND METHODS: We reviewed 615 electrical car-dioversions performed electively in 493 patients with atrial flut-ter. Embolic complications were evaluated during the 30 daysafter cardioversion. Follow-up data were obtained by follow-upvisits and by contacting the treating physician.RESULTS: Anticoagulants had been administered in 415 car-dioversions (67%). Cardioversion was successful in 570 proce-dures (93%). Three embolic events (in 3 patients) occurred inthe 30 days after 550 successful cardioversions with completedfollow-up (0.6% of successful procedures; 95% confidence in-

terval, 0.1% to 1.6%). Two of the 3 patients had not been anti-coagulated, whereas the third patient had subtherapeutic oralanticoagulation. No embolic event occurred in procedures per-formed with adequate anticoagulation. The incidence of embo-lism in patients regardless of subtherapeutic anticoagulationwas 1% (3 of 303 successful cardioversions).CONCLUSIONS: We observed a low (0.6%) incidence ofpostcardioversion thromboembolic complications in patientswith atrial flutter. Embolic events did not occur in patientswith adequate anticoagulation. Am J Med. 2001;111:433– 438.�2001 by Excerpta Medica, Inc.

Electrical cardioversion is used widely to restore si-nus rhythm in patients with atrial flutter (1). Al-though anticoagulant therapy is recommended be-

fore and after electrical cardioversion in patients withatrial fibrillation (2–5), there are no guidelines regardingthe need for anticoagulation in patients with atrial flutterwho have no history of atrial fibrillation. Data on the riskof postcardioversion thromboembolism in patients withatrial flutter are conflicting, with reports indicating min-imal or no risk (4) to a variable magnitude of risk (6 –11).This lack of consensus may be related to the small num-ber of observational studies demonstrating a temporalassociation between embolization and cardioversion inpatients with atrial flutter.

Patients with atrial flutter have organized atrial electri-cal activity and contraction and are thought to have lessblood stasis, and therefore a lower risk of thromboembo-lism, compared with patients with atrial fibrillation.However, recent studies have reported left atrial thrombiin a number of patients with atrial flutter (12–14). Resto-ration of sinus rhythm by electrical cardioversion hasbeen shown to induce myocardial stunning, with the for-mation or worsening of spontaneous echo contrast, as

previously demonstrated in patients with atrial fibrilla-tion (15–19).

Our aim was to determine the incidence of thrombo-embolic complications after cardioversion in patientswith atrial flutter.

MATERIAL AND METHODS

Study PopulationThis was a retrospective analysis of 615 electrical cardio-versions performed electively in 493 patients at the MayoClinic, Rochester, Minnesota, from 1990 through 1994.Patients were identified consecutively through a registryof cardioversions performed at our institution. Atrialflutter was defined as an organized atrial tachycardia of240 beats/min or more. The duration of atrial flutter, de-fined by documentation of arrhythmias on echocardio-graphic recordings, was less than 1 month in 439 cardio-versions (71%), 1 to 6 months in 125 (20%), greater than6 months to 1 year in 18 (3%), and greater than 1 year in33 (5%). Diagnosis was made by 12-lead electrocardiog-raphy with the use of standard criteria (12).

In every procedure, patient records were reviewed foranticoagulation status. Parameters analyzed were pro-thrombin time, international normalized ratio, and acti-vated partial thromboplastin time at cardioversion; dura-tion of anticoagulation; and anticoagulant medications atdischarge. Patients were considered to be in the therapeu-tic range of anticoagulation if the duration of anticoagu-

From the Division of Cardiovascular Diseases and Internal Medicine(AE, FG, BKK, BJG, JBS, AJT) and the Section of Biostatistics (KRB,SCM), Mayo Clinic, Rochester, Minnesota.

Requests for reprints should be addressed to Bijoy K. Khandheria,MD, Division of Cardiovascular Diseases and Internal Medicine, MayoClinic, 200 First Street SW, Rochester, Minnesota 55905.

�2001 by Excerpta Medica, Inc. 0002-9343/01/$–see front matter 433All rights reserved. PII S0002-9343(01)00902-0

lation was 3 weeks or longer with an international nor-malized ratio greater than 2, prothrombin time longerthan 20 seconds (longer than 17 seconds before April1993), or activated partial thromboplastin time longerthan 60 seconds in men or longer than 58 seconds inwomen (20,21). Hypertension was defined as systolicblood pressure 140 mm Hg or greater, diastolic bloodpressure 90 mm Hg or greater, or by use of antihyperten-sive medication. Diabetes mellitus was defined as a fastingblood glucose concentration 140 mg/dL or greater or byrequirement for insulin or orally administered hypogly-cemic agents. The diagnosis of previous myocardial in-farction was based on the standard criteria of chest pain,serial electrocardiography, and cardiac enzyme changes.Dilated cardiomyopathy was defined as global left ven-tricular systolic dysfunction in the absence of a historysuggestive of coronary artery disease. Substantial valvularheart disease was considered in patients with moderate orsevere stenosis/regurgitation in 1 or more cardiac valves,as determined by clinical examination and echo-Dopplerevaluation.

CardioversionAll cardioversion procedures were performed with elec-trocardiographic monitoring and full equipment for car-diopulmonary resuscitation. Digitalis preparations werewithheld before the procedure. A synchronized direct-current shock was administered after intravenous seda-tion. The electrocardiogram was recorded before and af-ter the procedure to verify the changes in cardiac rhythm.Cardioversion was considered successful if sinus rhythmwas restored and the patient was discharged from the car-dioversion unit in sinus rhythm. Patients were followedfor 30 days after cardioversion for evaluation of embolicevents. Follow-up data were obtained by follow-up visitsand by contacting the treating physician.

Statistical AnalysisUnless otherwise specified, data are presented as meanvalues � SD or frequency percentages. Chi-squared testswere used to compare differences between proportions.Student t test was used for analysis of continuous data.Differences were considered significant if the null hy-pothesis could be rejected at the 0.05 probability level.Rates of embolic events were calculated by dividing thenumber of events by the number of procedures. Confi-dence intervals (CI) were calculated on the basis of thePoisson distribution.

RESULTS

The mean age of the study patients was 66 � 15 years.There were 362 men (73%). Associated diseases were hy-pertension in 218 patients (44%), diabetes mellitus in 87(18%), coronary artery disease in 214 (43%), valvularheart disease in 126 (26%), and cardiomyopathy in 90

(18%). Previous atrial fibrillation was documented in 283procedures (46%).

Procedures that were performed without anticoagula-tion were associated with a shorter duration of atrial flut-ter, lower left atrial and left ventricular end-diastolic di-ameters, lower prevalence of valvular heart disease, and ahigher prevalence of medication with aspirin (Table 1).

Successful CardioversionCardioversion was successful in 570 of the 615 proce-dures (93%). Atrial fibrillation was induced in 23 proce-dures, whereas atrial flutter persisted after cardioversionin the remaining 22 procedures. The success rate was 94%(414 of 439) in atrial flutter of 1 month’s duration or lessversus 89% (156 of 176) in atrial flutter of longer than 1month’s duration (P � 0.02) Successful cardioversionwas associated with younger age (65 years � 15 vs. 71 �10 years, P � 0.01). No significant difference was foundbetween successful and failed procedures in terms of sex[75% (426) vs. 64% (29) men, P � 0.1], hypertension[44% (249) vs. 56% (25), P � 0.1], diabetes mellitus [17%(96) vs. 13% (8), P � 0.5], valvular heart disease [25%(145) vs. 27% (12), P � 0.9], coronary artery disease[43% (245) vs. 38% (17), P � 0.5], or cardiomyopathy[18% (101) vs. 22% (10), P � 0.5]. The success rate ofcardioversion was 94% (44 of 47) in procedures withatrial flutter of less than 48 hours’ duration, 94% (370 of392) for atrial flutter duration of 48 hours to 1 month,90% (113 of 125) for 1 to 6 months, 89% (16 of 18) for 6months to 1 year, and 82% (27 of 33) for 1 year.

Adequacy of Anticoagulation in SuccessfulCardioversionsPatients were anticoagulated in 383 of the 570 successfulcardioversions (67%). The treatment regimen consistedof oral anticoagulation only in 188 procedures (49%),heparin only in 118 (31%), and a combination of oralanticoagulation and heparin in 77 (20%). Anticoagula-tion was considered adequate in 187 procedures (49%)and inadequate in 131 (34%); the adequacy was unknownin 65 procedures (17%) because of the lack of appropriatelaboratory values. Oral anticoagulants were administeredafter hospital discharge in 281 of successful procedures(49%). In patients with successful cardioversion whowere receiving anticoagulants, the median duration oforal anticoagulation before cardioversion was 65 days(range, 2 days to 5.5 years), whereas the median durationof heparin therapy was 3 days (range, 11 days before toimmediately after cardioversion). Bleeding complica-tions (gastrointestinal bleeding in 1 patient and urinarytract bleeding in the other) occurred in 2 anticoagulatedpatients (incidence � 0.5% of those who received antico-agulants). In both patients, bleeding was managed with-out the need for blood transfusion or surgical interven-tion.

Thromboembolic Complications after Electrical Cardioversion in Patients with Atrial Flutter/Elhendy et al

434 October 15, 2001 THE AMERICAN JOURNAL OF MEDICINE� Volume 111

Follow-upFollow-up for 30 days after cardioversion was completedafter 592 procedures (96%) consisting of 550 successfuland 42 failed cardioversions. Follow-up was less than 30days after 23 procedures (4%) consisting of 20 successfuland 3 failed cardioversions. The mean duration of fol-low-up was 6 � 6 days after these procedures. These car-dioversions were uneventful to the last date of follow-up.

Five deaths occurred within 1 month after cardiover-sion; none were related to embolic complications. Thecauses were heart failure in 3 patients, coronary arterydisease in 1 patient, and advanced cerebral disease in 1patient with multiple previous strokes who had no evi-dence of a new embolic event after cardioversion.

Embolic EventsThree embolic events (in 3 patients) occurred in the 30days after successful cardioversion (0.6% of successfulprocedures with completed follow-up; 95% CI: 0.1% to1.6%). There were no embolic events after failed cardio-versions. One patient with dilated cardiomyopathy had atransient neurologic deficit the day after cardioversion.Carotid ultrasound revealed no substantial abnormali-ties, and computed tomographic scanning revealed anold cortical infarct and no hemorrhage. Another patienthad a stroke 3 days after cardioversion. Computed tomo-graphic scanning revealed early changes consistent withan infarction in the distal left middle cerebral artery dis-

tribution without hemorrhage; carotid ultrasound wasnot performed. The third patient had pulmonary embo-lism documented by pulmonary angiography 13 days af-ter cardioversion. No Doppler studies of the peripheralveins were performed. None of the embolic events wasfatal. Duration of atrial flutter was longer than 1 month in1 patient and less than 48 hours in 2 patients (in 1 of these2 patients, atrial fibrillation was the initial rhythm beforechanging to atrial flutter).

Relation between Anticoagulation and EmbolicEventsIn one of the three procedures complicated by emboliza-tion (Figure), the patient had been anticoagulated, butthe anticoagulation was subtherapeutic (internationalnormalized ratio � 1.4). The 2 other patients with embo-lization had not been anticoagulated. None (0%) of the182 successful procedures performed with adequate an-ticoagulation was complicated by embolization (95% CI:0% to 2.0%), whereas embolization occurred in 3 (1%) of303 successful procedures with documented inadequateor absent anticoagulation (95% CI, 0.2% to 2.9%; P �0.3). The embolization rate without anticoagulation was1.1% (2 of 188 procedures; CI, 0.1% to 3.8%).

Echocardiographic FindingsPatients with successful cardioversion. Transthoracicechocardiography was performed before 453 of the suc-

Table 1. Clinical Features and Echocardiographic Data Relevant to 615 Cardioversions Performedin Patients with Atrial Flutter

Characteristic

WithAnticoagulation

(n � 415)

WithoutAnticoagulation

(n � 200) P Value

Mean � SD or Number (%)

Age (year) 66 � 14 64 � 18 0.7Male sex 303 (73) 152 (76) 0.5Hypertension 186 (45) 88 (44) 0.8Diabetes mellitus 70 (17) 34 (17) 1Previous myocardial infarction 106 (26) 55 (28) 0.6Cardiomyopathy 87 (21) 24 (12) 0.01Valvular disease 116 (28) 41 (21) 0.05Left atrial diameter (mm) 49 � 8 46 � 9 �0.01Left ventricular end-diastolic diameter (mm) 53 � 9 51 � 10 0.06Left ventricular end-systolic diameter (mm) 39 � 12 38 � 12 0.4Ejection fraction (%) 46 � 18 47 � 16 0.09Duration of atrial flutter �0.01�48 hours 15 (4) 32 (16)�48 hours to 1 month 253 (61) 139 (69)�1 month to 6 months 107 (26) 18 (9)�6 months to 1 year 15 (4) 3 (2)�1 year 24 (6) 9 (5)

Medication with aspirin 112 (27) 87 (43) �0.01Medication with dipyridamole 24 (6) 15 (8) 0.4Anticoagulants at discharge 302 (73) 9 (5) 0.001


October 15, 2001 THE AMERICAN JOURNAL OF MEDICINE� Volume 111 435

cessful cardioversions. Intracardiac thrombi were de-tected in six procedures (1.3%). Four of the 6 patientsunderwent subsequent transesophageal echocardiogra-phy, which excluded thrombi in all of them. All patientswith intracardiac thrombi were receiving anticoagulantsbefore cardioversion. Transesophageal echocardiogra-phy was performed in 152 successful procedures and re-vealed left atrial spontaneous echo contrast in 45 proce-dures (30%), as well as a left atrial appendage thrombusin 2 patients (1.3%). The overall prevalence of spontane-ous echo contrast was 28% (49 of 177) in all (successfuland failed) cardioversions. None of the patients whodemonstrated intracardiac thrombi by transthoracic ortransesophageal echocardiography had subsequent em-bolism.

Patients with embolization. Transthoracic echocardi-ography was performed before cardioversion in 2 of the 3patients with embolization. No intracardiac thrombiwere detected. Both patients had left atrial enlargementand reduced ejection fraction. Transesophageal echocar-diography was not performed in any of the patients whohad embolization.

DISCUSSION

In this study, successful restoration of sinus rhythm afteratrial flutter was achieved in 93% of the cardioversionprocedures. Failure of cardioversion was significantly as-sociated with older age (P � 0.01) and longer duration ofatrial flutter (P � 0.02). Diabetes mellitus, valvular heartdisease, cardiomyopathy, and coronary artery disease hadno impact on the success of cardioversion, suggesting thatcardioversion should not be deferred in patients present-

ing with atrial flutter on the basis of underlying heartdisease.

There were three embolic events within 30 days after550 successful cardioversions with completed follow-up(incidence of 0.6%; 95% CI: 0.1% to 1.6%). Patients wereanticoagulated in 67% of cardioversions. Anticoagulantswere given at the physicians’ discretion without random-ization of treatment. Procedures with anticoagulationwere more likely to be performed in patients who had amore remote onset of atrial flutter and who had higherclinical and echocardiographic risk profiles, such aslonger duration of atrial flutter, higher prevalence of val-vular heart disease, and larger left atrial and left ventric-ular dimensions. Therefore, the incidence of thrombo-embolism in this study population may have been re-duced by the administration of anticoagulants in themajority of these patients, particularly in those with ahigher risk profile. None of the 182 successful proceduresperformed with adequate anticoagulation was compli-cated by embolization. Only 1 patient who received oralanticoagulants had subsequent embolization. However,the anticoagulation level in this patient was subtherapeu-tic. Embolization occurred in 2 of the 31 patients (6.5%)without anticoagulation in whom atrial flutter had begunless than 48 hours earlier. It is possible that the acute onsetof atrial flutter represents a recurrence of a previouslyundetected atrial flutter, or atrial fibrillation, with alteredintra-atrial hemostasis.

Transthoracic echocardiography was performed be-fore 453 of the successful cardioversions. Intracardiacthrombi were detected in six procedures (1.3%). Trans-esophageal echocardiography was performed in 177 pro-cedures and revealed left atrial spontaneous echo contrastin 49 procedures (28%), whereas 2 patients (1.1%) had a

Figure. Distribution of embolic events in relation to the administration and the adequacy of anticoagulation in successful cardio-versions with completed follow-up.



left atrial appendage thrombus. The high prevalence ofspontaneous echo contrast in patients with atrial flutterin this study suggests that atrial flutter is associated withblood stasis in the left atrium, which may be a predispos-ing factor or an indicator of a high probability for throm-bus formation before and after cardioversion.

Some studies have shown that cardioversion of atrialflutter is associated with atrial stunning and formation ofspontaneous echo contrast. Weiss et al. (15) studied 28patients with atrial flutter by transesophageal echocardi-ography before and after cardioversion. Left atrial spon-taneous echo contrast developed de novo or worsened in12 patients (43%) after restoration of sinus rhythm. Iraniet al. (12) reported that 28% of patients with atrial flutterhad absent atrial mechanical activity immediately afterrestoration of sinus rhythm by cardioversion. It has alsobeen observed that left atrial appendage stunning occursin patients with atrial flutter (16), although to a lesserdegree than in those with atrial fibrillation, suggestingthat patients with atrial flutter are at risk of thromboem-bolic events after cardioversion, although this risk is mostlikely lower than that in patients with atrial fibrillationbecause of better preserved left atrial appendage function.

The reported incidence of embolization after cardio-version in patients with atrial flutter ranges from 0% to7% (4,7–12) (Table 2). There were no embolic eventsreported in Arnold and associates’ (4) retrospective anal-ysis that included 73% of patients with recent-onset atrialflutter. In Seidl and associates’ study (8), acute embolism(less than 48 hours) occurred in 4 of 191 patients. Lan-zarotti and Olshansky (7) reported a 6% incidence of em-boli that could not be attributed to causes other thanatrial flutter. Densem (10) recently pooled the data ofeight studies involving 457 patients with atrial flutter whounderwent cardioversion without adequate anticoagula-tion (international normalized ratio less than 2.0). Acutecardioversion-related thromboembolic events occurredin 2.2% of patients collectively.

The limitations of our study are the retrospective de-sign and the lack of randomization for administration ofanticoagulants. We did not assess the effect of cardiover-sion on left atrial appendage function and spontaneousecho contrast. We could not ascertain if the thromboem-bolic complications in the 3 patients were caused by car-dioversion. Other mechanisms, such as deep venousthrombosis, large artery atheroma, and dilated cardiomy-opathy, may have been the underlying mechanism forsuch embolic events.

This study reports a low (0.6%) incidence of thrombo-embolic events after cardioversion in patients with atrialflutter. Embolic events did not occur in patients with ad-equate anticoagulation. However, the role of anticoagu-lation in the prevention of embolic complications couldnot be demonstrated because of the low event rate.

REFERENCES1. Van Gelder IC, Tuinenburg AE, Schoonderwoerd BS, et al. Phar-

macologic versus direct-current electrical cardioversion of atrialflutter and fibrillation. Am J Cardiol. 1999;84:147R–151R.

2. Main ML, Klein AL. Cardioversion in atrial fibrillation: indications,thromboembolic prophylaxis, and role of transesophagealechocardiography. J Thromb Thrombolysis. 1999;7:53– 60.

3. Bjerkelund CJ, Orning OM. The efficacy of anticoagulant therapyin preventing embolism related to D.C. electrical conversion ofatrial fibrillation. Am J Cardiol. 1969;23:208–216.

4. Arnold AZ, Mick MJ, Mazurek RP, et al. Role of prophylactic anti-coagulation for direct current cardioversion in patients with atrialfibrillation or atrial flutter. J Am Coll Cardiol. 1992;19:851– 855.

5. Dalen JE, Hirsh J, eds. Third American College of Chest PhysiciansConsensus Conference on Antithrombotic Therapy. Chest. 1992;102(suppl):303S–549S.

6. Wood KA, Eisenberg SJ, Kalman JM, et al. Risk of thromboembo-lism in chronic atrial flutter. Am J Cardiol. 1997;79:1043–1047.

7. Lanzarotti CJ, Olshansky B. Thromboembolism in chronic atrialflutter: is the risk underestimated? J Am Coll Cardiol. 1997;30:1506–1511.

8. Seidl K, Hauer B, Schwick NG, et al. Risk of thromboembolic eventsin patients with atrial flutter. Am J Cardiol. 1998;82:580–583.

Table 2. Studies of Thromboembolic Complications of Cardioversion in Patients with Atrial Flutter

Study (Reference)Patients orProcedures Anticoagulation

AdequateAnticoagulation

EmbolicEvents

Embolic Eventswithout EffectiveAnticoagulation

Number (%)

Arnold et al. (4) 122 32 (26) NA 0 (0) 0 (0)Wood et al. (6) 56 NA NA 3 (5)* NAMehta and Baruch (9) 41 2 (5) NA 3 (7) 3/39 (8)Irani et al. (12) 36 5 (14) NA 0 (0) 0 (0)Seidl et al. (8) 138 67 (49) NA 3 (2) NALanzarotti and Olshansky (7) 100 84 (84) 54 (54) 6 (6) 6/46 (13)This study 615 415 (67) 205 (33) 3 (0.5) 3/303 (1)Polled 1108 605/1052 (58) 259/715 (36) 18/1108 (1.6) 12/388 (3.1)

* Patients not receiving aspirin or warfarin.NA � not available.


October 15, 2001 THE AMERICAN JOURNAL OF MEDICINE� Volume 111 437

9. Mehta D, Baruch L. Thromboembolism following cardioversion of“common” atrial flutter. Risk factors and limitations of transesoph-ageal echocardiography. Chest. 1996;110:1001–1003.

10. Densem CG. Patients undergoing cardioversion of atrial fluttershould be routinely anticoagulated. [letter] Am J Cardiol. 1999;83:140–141.

11. Moreyra E, Finkelhor RS, Cebul RD. Limitations of transesopha-geal echocardiography in the risk assessment of patients beforenonanticoagulated cardioversion from atrial fibrillation and flutter:an analysis of pooled trials. Am Heart J. 1995;129:71–75.

12. Irani WN, Grayburn PA, Afridi I. Prevalence of thrombus, sponta-neous echo contrast, and atrial stunning in patients undergoingcardioversion of atrial flutter. A prospective study using trans-esophageal echocardiography. Circulation. 1997;95:962–966.

13. Bikkina M, Alpert MA, Mulekar M, et al. Prevalence of intraatrialthrombus in patients with atrial flutter. Am J Cardiol. 1995;76:186–189.

14. Santiago D, Warshofsky M, Li Mandri G, et al. Left atrial appendagefunction and thrombus formation in atrial fibrillation-flutter: atransesophageal echocardiographic study. J Am Coll Cardiol.1994;24:159–164.

15. Weiss R, Marcovitz P, Knight BP, et al. Acute changes in spontane-ous echo contrast and atrial function after cardioversion of persis-tent atrial flutter. Am J Cardiol. 1998;82:1052–1055.

16. Grimm RA, Stewart WJ, Arheart K, et al. Left atrial appendage“s-tunning” after electrical cardioversion of atrial flutter: an attenu-ated response compared with atrial fibrillation as the mechanismfor lower susceptibility to thromboembolic events. J Am CollCardiol. 1997;29:582–589.

17. Black IW, Hopkins AP, Lee LC, Walsh WF. Evaluation of trans-esophageal echocardiography before cardioversion of atrial fibril-lation and flutter in nonanticoagulated patients. Am Heart J. 1993;126:375–381.

18. Grimm RA, Stewart WJ, Maloney JD, et al. Impact of electricalcardioversion for atrial fibrillation on left atrial appendage functionand spontaneous echo contrast: characterization by simultaneoustransesophageal echocardiography. J Am Coll Cardiol. 1993;22:1359–1366.

19. Manning WJ, Silverman DI, Katz SE, et al. Impaired left atrial me-chanical function after cardioversion: relation to the duration ofatrial fibrillation. J Am Coll Cardiol. 1994;23:1535–1540.

20. Hirsh J, Dalen JE, Deykin D, Poller L. Heparin: mechanism of ac-tion, pharmacokinetics, dosing considerations, monitoring, effi-cacy, and safety. Chest. 1992;102(suppl):337S–351S.

21. Hirsh J, Dalen JE, Deykin D, Poller L. Oral anticoagulants. Mecha-nism of action, clinical effectiveness, and optimal therapeuticrange. Chest. 1992;102(suppl):312S–326S.



Documents

Thromboembolic complications after electrical cardioversion in patients with atrial flutter