Left Atrial Radiofrequency Ablation During Mitral Valve Surgery for Continuous Atrial Fibrillation a Randomized Controlled Trial

ORIGINAL CONTRIBUTION

Left Atrial Radiofrequency Ablation During Mitral Valve Surgery for Continuous Atrial Fibrillation A Randomized Controlled Trial

George Doukas, FRCSI Context Although left atrial radiofrequency ablation (RFA) is increasingly used for Nilesh J. Samani, MD, FRCP the treatment of chronic atrial fibrillation during mitral valve surgery, its efficacy to Christos Alexiou, FRCS, PhD restore sinus rhythm and any resulting benefits have not been examined in the context of an adequately powered randomized trial. Mehmet Oc, MD Objective To determine whether intraoperative RFA of the left atrium increases the Derek T. Chin, MD long-term restoration of sinus rhythm and improves exercise capacity. Peter G. Stafford, MD Design, Setting, and Patients Randomized, double-blind trial performed in a single Leong L. Ng, MD UK tertiary referral center with enrollment between December 2001 and November 2003. A total of 101 patients referred for mitral valve surgery with at least 6 months’ Tomasz J. Spyt, FRCS, MD history of uninterrupted atrial fibrillation were assessed for eligibility; 97 were en- TRIAL FIBRILLATION IS A FRE- rolled. Patients were followed up for 12 months. quent and important con- Intervention Patients were randomly assigned to undergo mitral valve surgery and tributor to symptoms and RFA of the left atrium (n=49) or mitral valve surgery alone (controls; n=48). morbidity in patients with Main Outcome Measures The primary outcome measure was presence of sinus mitralA valve disease. More than 40% of rhythm at 12 months; secondary measures were patient functional status and exercise patients referred for mitral valve sur- capacity (assessed by shuttle-walk test), left atrial contractility, and left atrial and left gery have continuous atrial fibrilla- ventricular dimension and function and plasma levels of B-type natriuretic peptide. tion.1-3 Patients who remain in atrial fi- Results At 12 months, sinus rhythm was present in 20 (44.4%) of 45 RFA patients brillation following mitral valve surgery and in 2 (4.5%) of 44 controls (rate ratio, 9.8; 95% CI, 2.4-86.3; PϽ.001). Restora- may have lower survival 3 to 5 years tion of sinus rhythm in the RFA group was accompanied by a greater improvement in ϩ ϩ later4 compared with those in sinus mean (SD) shuttle-walk distance compared with controls ( 94 [102] m vs 48 [82] rhythm, although this has not been ob- m; P=.003) and a greater reduction in the plasma level of B-type natriuretic peptide 5 (−104 [87] fmol/mL vs −51 [82] fmol/mL; P=.03). Patients randomized to receive RFA served in all studies. had similar rates of postoperative complications and deaths as control patients. For a long time, it was believed that atrial fibrillation occurs as a result of a Conclusions Radiofrequency ablation of the left atrium during mitral valve surgery multiple-circuit reentry mechanism. for continuous atrial fibrillation significantly increases the rate of sinus rhythm restoration 1 year postoperatively, improving patient exercise capacity. On the basis of its This notion is now challenged by evi- efficacy and safety, routine use of RFA of the left atrium during mitral valve surgery dence suggesting that, in some cir- may be justified. cumstances, even continuous atrial Trial Registration ClinicalTrials.gov Identifier: NCT00238706. fibrillation may be organized by 1 or a JAMA. 2005;294:2323-2329 www.jama.com small number of high-frequency sources (rotors) in the left atrium, a novel surgical mode for treatment of Author Affiliations: Departments of Cardiac Surgery (Drs especially in the vicinity of the pulmo- Doukas, Alexiou, Oc, and Spyt) and Cardiology (Drs nary veins.6 Hence, radiofrequency atrial fibrillation. It is based on the Samani, Chin, and Stafford), Glenfield Hospital, Uni- original concept of the maze proce- versity Hospitals of Leicester NHS Trust, and the Depart- ablation (RFA) of the left atrium dur- 7 ment of Cardiovascular Sciences, University of Leicester ing open-heart surgery has emerged as dure and aims to eliminate anatomi- (Drs Doukas, Samani, Ng, and Spyt), Leicester, England. cally determined reentrant circuits by Correspondence: George Doukas, FRCSI, Department creating contiguous lines of scar tissue of Cardiac Surgery, Glenfield Hospital, Groby Road, Leic- For editorial comment see p 2357. ester LE3 9QP, England (doukas_george@hotmail between the pulmonary veins and the .com).

Downloaded From: https://jamanetwork.com/ on 09/25/2021 LEFT ATRIAL RADIOFREQUENCY ABLATION IN MITRAL VALVE SURGERY

mitral valve annulus.7,8 Although RFA and the superioinferior left atrial atrium size was studied in the 2-dimen- therapy of the left atrium is being diameter (Ն60 vs Ͻ60 mm). These 3 sional mode on apical views. Mitral in- increasingly offered to patients with variables produced 8 subgroups. Ran- flow velocity was obtained by pulse- chronic atrial fibrillation who are domized blocks of between 4 and 6 wave Doppler examination from the undergoing mitral valve surgery,9-11 its were prepared for each subgroup in apical 4-chamber view. efficacy has not been tested in the con- advance by computer-generated num- text of an adequately powered ran- bers, and individual allocations were Surgery and RFA Procedure domized study. The aims of this pro- placed in sealed envelopes. Assign- Operations were performed between spective randomized clinical trial were ment took place on the day of the December 2001 and November 2003 to evaluate early and late outcomes operation by picking the next enve- by a single surgeon (T.J.S.). Access to following left atrial RFA during mitral lope for the relevant group by a person the mitral valve was obtained through valve surgery for treatment of continu- masked to previous allocations. In the left atrium. If tricuspid valve sur- ous atrial fibrillation and to assess the addition, the participants and the gery was undertaken, a transeptal functional effects of restoration of assessors of outcomes were blinded to approach was used. A Cosgrove- sinus rhythm. group assignment. Edwards annuloplasty band was rou- tinely inserted during mitral valve re- METHODS Outcome End Points pair, tricuspid valve repair, or both. Participants and Sample Size The left atrial appendage was over- The trial was conducted within the De- The primary outcome end point was the sewn from within in all patients in both partments of Cardiac Surgery and Car- presence of sinus rhythm at 12 months. groups. diology, Glenfield Hospital, Leicester, Secondary end points included pa- Radiofrequency lesions were cre- England. Patients requiring mitral valve tient functional status and exercise ca- ated endocardially with a handheld surgery and who also had a history of pacity, left atrial contractility, and left monopolar, 7-electrode, temperature- continuous atrial fibrillation were eli- atrial and left ventricular dimension and controlled probe (EP Technologies, gible. The term continuous atrial fibril- function and plasma levels of B-type na- Boston Scientific Corp, San Jose, lation denotes the presence of uninter- triuretic peptide (BNP). Calif). Radiofrequency generators rupted atrial fibrillation for at least 6 At the time of study design, avail- were set at 100 W and 70°C. Radiofre- months that showed no evidence of able data on the long-term efficacy of quency waves were delivered for 120 spontaneous reversibility to sinus RFA during mitral valve surgery sug- seconds to achieve transmural lesions, rhythm and was not possible to revert gested a reduction in postoperative as previously described.10 The left with medications or direct current car- atrial fibrillation of 64%.9 Assuming atrial incision was complemented dioversion. Patients with sick sinus syn- that after mitral valve surgery alone, by a semilunar RFA line to isolate the drome, uncontrolled hyperthyroid- 85% of those who were in atrial fibril- right pulmonary veins. The left pul- ism, permanent pacemaker, or previous lation preoperatively would continue monary veins were then encircled and cardiac surgery were excluded. The pro- to be in atrial fibrillation,12 approxi- a line was drawn connecting the 2 tocol was approved by the Leicester- mately 100 patients (50 in each encircling lines, the obliterated left shire Research Ethics Committee (ref- group) were required to detect a 30% atrial appendage, and the mitral valve erence No. 7503). Written informed reduction in atrial fibrillation with a annulus. consent was obtained from all partici- power of 90% at the .05 level of statis- pants, and the recommendations of the tical significance. This would allow for Postoperative Management revised version of the Declaration of up to 12% of patients being lost to and Follow-up Helsinki were met. follow-up or dying within the study Cardiac rhythm was continuously period. monitored for 48 hours. Thereafter, Interventions and daily 12-lead electrocardiograms Randomization Process Preoperative Assessment (ECGs) were performed for the dura- Patients were randomly assigned to Demographics and clinical data were tion of hospitalization. Per protocol, all undergo mitral valve surgery alone or prospectively recorded. The shuttle- patients were given amiodarone (or so- mitral valve surgery plus RFA of the walk test (SWT), a validated instru- talol if amiodarone was not tolerated) left atrium. Additional cardiac proce- ment for assessing functional capac- for at least 3 months. In patients who dures were performed as required. ity,13 was carried out in all patients. were in stable sinus rhythm, the anti- Blocked stratified randomization by Transthoracic echocardiography was arrhythmic agent was stopped at 3 patient characteristics was used. The performed to examine mitral valve func- months. Those who developed persis- variables used were age (Ն70 vs Ͻ70 tion, left atrium dimensions, atrial tent atrial fibrillation during the hos- years), etiology of the mitral valve dis- transport function, and left ventricu- pital stay underwent cardioversion ease (rheumatic vs nonrheumatic), lar dimension and function. The left within 24 hours. Patients who re-

Downloaded From: https://jamanetwork.com/ on 09/25/2021 LEFT ATRIAL RADIOFREQUENCY ABLATION IN MITRAL VALVE SURGERY

mained in atrial fibrillation at hospital SWT (ie, the distance traveled in Figure. Flow of Patients Through the Trial discharge had another cardioversion at- meters) was measured at the end of

tempt 4 weeks later. If this was unsuc- each effort. 101 Patients Assessed cessful, further cardioversion was at- for Eligibility tempted only if clinically indicated. BNP Measurement Patients who remained in atrial fibril- Five milliliters of venous blood was col- 4 Excluded 2 Had Sick Sinus lation after the 3-month follow-up in- lected in tubes containing 1 mg/mL of Syndrome 2 Refused to terval were treated with a variety of EDTA and 500 U/mL of aprotinin. Participate antiarrhythmic agents to achieve ap- Blood was centrifuged and the plasma propriate rate control. Warfarin was ad- stored at −70°C until assay. Plasma for ministered to all patients for the dura- BNP was extracted on C18 columns be- 97 Randomized tion of the study. Before hospital fore assay. B-type natriuretic peptide 49 Assigned to Undergo 48 Assigned to Undergo discharge, all patients underwent trans- levels were quantified using specific an- Mitral Valve Surgery Mitral Valve Surgery thoracic echocardiography. The pa- tibodies in a competitive immunolu- (With or Without Other (With or Without Other 15 Cardiac Procedure) Cardiac Procedure) tients were seen at 3, 6, and 12 post- minometric assay. and Radiofrequency Only (Control Group) operative months in the outpatient Ablation clinics, where they had a clinical ex- Statistical Analysis 48 Received Intervention 48 Received Intervention amination and a 12-lead ECG. At 6 and Analysis was undertaken on an inten- as Assigned as Assigned 12 months, a transthoracic echocar- tion-to-treat basis. Categorical vari- 1 Did Not Undergo diogram and an SWT were also car- ables were compared with ␹2 test or Mitral Valve Surgery ried out. Blood samples were col- Fisher exact test and continuous vari- 3 Operative Deaths 4 Operative Deaths lected for BNP levels preoperatively and ables with an unpaired t test or Mann- 0 Lost to Follow-up 0 Lost to Follow-up at 6 and 12 months postoperatively. Whitney test as appropriate. Predic- During follow-up, if a patient re- tors of sinus rhythm conversion within 45 Included in Analysis 44 Included in Analysis ported symptoms suggestive of dys- the RFA group were identified with uni- rhythmia either at a specified visit or variate and multiple logistic regres- between visits, a 24-hour Holter tape sion analysis. First, a series of vari- Preoperative Clinical Profile (and an ECG if appropriate) was ob- ables were screened with univariate There were no significant differences tained, and if atrial fibrillation was ob- analysis. Then, the variables that at- between the RFA and control groups served in a patient with previous si- tained a P value of Յ.05 were entered in their demographics, type of mitral nus rhythm, the patient was deemed to into a multiple logistic regression valve pathology, prevalence of tricus- have reverted to atrial fibrillation at the model. Statistical analysis was per- pid regurgitation, left atrium size, left next formal assessment. Likewise, the formed using SPSS software, version 11 ventricular function, and duration of findings on any sporadic ECGs ob- (SPSS Inc, Chicago, Ill). atrial fibrillation (TABLE 1). Patients in tained by the patients’ physicians were both groups had similar rate-control similarly taken into account. antiarrhythmic medication use. Asso- RESULTS ciated conditions were evenly distrib- Shuttle-Walk Test Study Design and Conduct uted. There were no differences in Functional capacity during exercise The design and flow of the trial is shown New York Heart Association classifica- was assessed using the SWT. This is a in the FIGURE. One hundred one con- tion, SWT distance walked, or Parson- maximal symptom-limited test with secutive patients requiring mitral valve net score (a scoring system based on 12 progressive levels14 during which surgery who also had continuous atrial preoperative risk factors that aims to patients are required to walk 10 m fibrillation were assessed for eligibil- predict the mortality risk in patients back and forth. The walking speed ity; 97 were enrolled and randomized. about to undergo coronary and heart is paced by an audio signal from a All patients underwent RFA as random- valve operations16) (Table 1). cassette that emits beeps at regular ized with the exception of 1 patient, intervals. The speed is increased each who was found unexpectedly to have Operative Procedures minute by 0.17 m/s until the next level mild mitral regurgitation during the and Clinical Outcomes is attained. The test is terminated preoperative on-table transesophageal The 2 groups underwent comparable either by the patient when he/she echocardiography and did not un- numbers and types of operations with becomes breathless maintaining the dergo mitral valve surgery. This pa- similar bypass and ischemic times required speed or by the operator tient was excluded from further analy- (TABLE 2). There were 3 hospital when the patient fails to complete sis. During the follow-up period, none deaths in the RFA group (6.1%) and 4 a shuttle in the time allowed. For of the controls crossed over to the RFA in the control group (8.3%) (P=.71). this study, the level reached in the group. There were no differences between the

Downloaded From: https://jamanetwork.com/ on 09/25/2021 LEFT ATRIAL RADIOFREQUENCY ABLATION IN MITRAL VALVE SURGERY

groups in other complications or Table 1. Baseline Patient Characteristics* lengths of stay in the intensive care Radiofrequency Ablation Group Control Group Characteristics (n = 49) (n = 48) unit or the hospital (Table 2). Male 31 (63.3) 24 (50) After hospital discharge, there were Age, mean (SD), y 67.2 (9) 67 (8) no deaths at 12 months. Two patients Hypertension 15 (30.6) 11 (23) in the RFA group had reoperations, 1 Hypercholesterolemia 5 (10.2) 7 (14.5) for recurrent mitral regurgitation due Diabetes mellitus 1 (2) 2 (4.2) to chordal rupture and 1 for mitral valve Chronic obstructive pulmonary disease 4 (8.2) 2 (4.2) stenosis following an edge-to-edge re- CVA or TIA 9 (18.3) 4 (8.3) pair. Two patients in the control group Renal impairment 3 (6.1) 2 (4.2) experienced transient ischemic at- Angina symptoms 11 (22.4) 14 (29) tacks at 4 and 9 postoperative months. ␤-Blocker use 33 (67.3) 31 (64.6) Digoxin use 38 (77.6) 40 (83.3) Antiarrhythmic Drug Therapy, Calcium channel blocker use 17 (34.6) 14 (29.2) Cardiac Rhythm, and Predictors NYHA class, mean (SD) 2.5 (0.7) 2.4 (0.6) of Sinus Rhythm Restoration Duration of atrial fibrillation, mean (SD), mo 57 (55.1) 46.7 (64.3) Mitral regurgitation 41 (83.6) 40 (83.3) Eighty-two patients (92%) were tak- Mixed mitral valve disease 8 (16.4) 8 (16.7) ing amiodarone and 7 (2 in the RFA Mitral valve disease etiology group and 5 controls) were taking so- Degenerative 36 (73) 32 (66.6) talol at hospital discharge. At 12 Rheumatic 11 (22.4) 11 (23) months, of the 22 patients who were in Ischemic 2 (4.1) 5 (10.4) sinus rhythm, none were taking amio- Moderate or severe tricuspid regurgitation 12 (24.5) 14 (29.2) darone and 4 were taking ␤-blockers for Impaired left ventricular function 13 (26) 9 (19) other indications. Of the 67 patients in Left atrial size, mean (SD), cm 5.8 (0.7) 6.0 (1.1) atrial fibrillation at 12 months, 28 were Parsonnet score† 13.7 (7) 14.4 (9.1) taking amiodarone, 39 ␤-blockers, 57 Shuttle-walk distance, mean (SD), m 281 (143) 253 (115) digoxin, and 22 calcium channel Abbreviations: CVA, cerebrovascular accident; NYHA, New York Heart Association; TIA, transient ischemic attack. *Data are reported as No. (%) unless otherwise indicated. antagonists. †Score range is 0 to 56. Information on cardiac rhythm at hospital discharge and during follow-up is shown in TABLE 3. At each Table 2. Operative Data and Early Clinical Outcomes* time point, sinus rhythm was signifi- Radiofrequency Ablation Group Control Group Parameters (n = 49) (n = 48) cantly more prevalent in the RFA group Mitral valve repair 38 (77) 34 (71) (eg, at 12 months, rate ratio, 9.8; 95% Quadrangular resection 15 (39) 13 (38) confidence interval, 2.4-86.3; PϽ.001). Sliding plasty 4 (11) 2 (6) In a multivariate logistic regression Edge-to-edge repair 5 (13) 6 (18) analysis, a left atrium larger than 6 cm Artificial chordae insertion 3 (8) 1 (3) in patients with rheumatic mitral valve Annuloplasty band only 10 (27) 12 (35) disease was the only independent pre- Mitral valve replacement 11 (23) 14 (29) dictor of persisting atrial fibrillation 12 Coronary artery bypass graft surgery 5 (10.2) 6 (12.5) months following RFA (adjusted odds Tricuspid valve repair 9 (18.4) 7 (14.6) ratio, −0.76; 95% confidence interval, Bypass time, mean (SD), min 106 (34) 99 (37) −1.70 to −0.20; P=.01). Tricuspid valve Aortic clamp time, mean (SD), min 70 (26) 64 (28) repair was a negative factor that ap- Elective procedures 44 (90) 42 (87) proached but did not reach statistical Operative mortality† 3 (6.1) 4 (8.3) Chest infection 8 (16.4) 9 (18.7) significance (P=.06). CVA or TIA 2 (4) 1 (2.1) Functional Outcomes Sepsis 3 (6.1) 2 (4.2) Hemofiltration 3 (6.1) 6 (12.5) Compared with baseline, the RFA and Intensive care unit stay, mean (SD), d 1.93 (1.96) 2.42 (2.93) control groups both achieved greater Total hospital stay, mean (SD), d 11.9 (5.5) 12.2 (7.1) SWT distances postoperatively Abbreviations: CVA, cerebrovascular accident; TIA, transient ischemic attack. (TABLE 4). At 12 months, the RFA *Data are reported as No. (%) unless otherwise indicated. †Causes of death in the radiofrequency ablation group were low cardiac output (n = 1), pneumonia (n = 1), and stroke group recorded longer SWT distances (n = 1) and in the control group were right-sided heart failure (n = 2), intestinal infarction (n = 1), and sudden cardiac than the control group (P =.02), arrest (n = 1). although the increase in SWT distance

Downloaded From: https://jamanetwork.com/ on 09/25/2021 LEFT ATRIAL RADIOFREQUENCY ABLATION IN MITRAL VALVE SURGERY

from baseline was not statistically sig- Echocardiographic Evaluation cantly lower left ventricular end- nificantly higher in the RFA group Preoperatively, the RFA and control systolic diameter and higher ejection (P=.13). Within the RFA group, there groups had similar ejection fraction, left fraction at 12 months (TABLE 5). Of the was no difference in the baseline SWT ventricular end-systolic diameter, and 22 patients (20 in the RFA group and between patients who subsequently left ventricular end-diastolic diam- 2 controls) who were in sinus rhythm converted to sinus rhythm and those eter. Compared with controls, pa- at 12 months, 19 (86.4%) (18 in the who did not. However, patients who tients who underwent RFA had signifi- RFA group and 1 control) exhibited a converted to sinus rhythm had significantly greater improvements in SWT distances at both 6 and 12 months Table 3. Postoperative Cardiac Rhythm at Various Time Intervals* compared with those who remained in Radiofrequency Ablation Group Control Group atrial fibrillation (Table 4). The con- Rhythm (n = 45) (n = 44) P Value Hospital discharge trol and the RFA groups had similar Sinus rhythm 24 (53.3) 2 (4.5) Ͻ.001 average New York Heart Association Atrial fibrillation 12 (26.6) 26 (59.1) .004 class scores 1 year postoperatively Atrial flutter 1 (2.2) 0 Ͼ.99 (Table 4). Nodal rhythm 6 (13.3) 12 (27.3) .16 Pacemaker 2 (4.4) 4 (9.1) .43 BNP Levels Three months At baseline, BNP levels were similar be- Sinus rhythm 21 (46.6) 3 (6.8) Ͻ.001 tween the RFA and control groups Atrial fibrillation 18 (40) 33 (75) .002 (Table 4). At 12 months, BNP values Atrial flutter 2 (4.4) 0 .49 had decreased in both groups. The Nodal rhythm 2 (4.4) 4 (9.1) .43 change from baseline was signifi- Pacemaker 2 (4.4) 4 (9.1) .43 Six months cantly greater in the RFA group Sinus rhythm 20 (44.4) 3 (6.8) Ͻ.001 (P=.02). Within the RFA group, there Atrial fibrillation 22 (48.9) 37 (84) .001 was no difference in baseline BNP level Atrial flutter 1 (2.2) 0 Ͼ.99 between those who converted to sinus Nodal rhythm 0 1 (2.3) Ͼ.99 rhythm and those who remained in Pacemaker 2 (2.2) 4 (9.1) .43 atrial fibrillation. At 12 months, BNP Twelve months levels were lower in those in sinus Sinus rhythm 20 (44.4) 2 (4.5) Ͻ.001 rhythm, but not significantly so. How- Atrial fibrillation 22 (48.9) 37 (84.1) .001 ever, those who converted to sinus Atrial flutter 1 (2.2) 0 Ͼ.99 rhythm had a significantly greater re- Nodal rhythm 0 1 (2.3) Ͼ.99 duction in their BNP levels from base- Pacemaker 2 (4.4) 4 (9.1) .43 line compared with those who did not *Data are reported as No. (%). The rate ratios of sinus rhythm in the radiofrequency ablation group compared with the control group were 11.7 (95% confidence interval [CI], 2.9-102.4), 6.8 (95% CI, 2.0-35.8), 6.5 (95% CI, 1.9-34.3), convert (P=.03). and 9.8 (95% CI, 2.4-86.3) at discharge, 3 months, 6 months, and 12 months, respectively.

Table 4. Functional and Biochemical Outcomes* Outcomes RFA Group Control Group P Value RFA, Sinus Rhythm RFA, Atrial Fibrillation P Value Shuttle-walk distance, m Baseline 281 (143) 253 (115) .33 313 (161) 244 (111) .11 6 mo 331 (136) 297 (114) .34 381 (128) 271 (121) .006 12 mo 359 (140) 304 (120) .02 407 (130) 292 (122) .002 Change from baseline to 12 mo 78 (94) 49 (97) .13 94 (102) 48 (82) .003 NYHA class Baseline 2.5 (0.7) 2.4 (0.6) .90 2.4 (0.7) 2.6 (0.7) .29 6 mo 1.4 (0.6) 1.5 (0.6) .67 1.3 (0.5) 1.4 (0.6) .58 12 mo 1.2 (0.5) 1.3 (0.5) .34 1.1 (0.4) 1.4 (0.5) .11 BNP level, median (IQR), fmol/mL Baseline 212 (151-319) 185 (96-294) .30 218 (156-358) 205 (141-317) .50 6 mo 155 (109-219) 152 (65-243) .72 169 (101-220) 192 (94-249) .32 12 mo 160 (103-210) 148 (81-231) .80 108 (79-173) 168 (125-209) .08 Change from baseline to 12 mo 76 (125) 30 (71) .02 −104 (87) −51 (82) .03 Abbreviations: BNP, B-type natriuretic peptide; IQR, interquartile range; NYHA, New York Heart Association; RFA, radiofrequency ablation. *Data are reported as mean (SD) unless otherwise indicated.

Downloaded From: https://jamanetwork.com/ on 09/25/2021 LEFT ATRIAL RADIOFREQUENCY ABLATION IN MITRAL VALVE SURGERY

(7%) is in keeping with previously pub- Table 5. Echocardiographic Data* lished figures.12,21,22 Radiofrequency Ablation Group Control Group Parameters (n = 45) (n = 44) P Value Functional capacity improved in both Baseline groups, reflecting a beneficial hemody- Ejection fraction, % 57 (6) 58 (7) .70 namic effect derived from correction of LVESD, cm 4.4 (0.5) 4.5 (0.7) .39 mitral regurgitation. However, within the LVEDD, cm 5.92 (0.4) 5.97 (0.6) .61 RFA group, patients converting to si- Maximum left atrial area, cm2 35 (7) 34 (9) .49 nus rhythm achieved significantly longer Minimum left atrial area, cm2 26 (7) 26 (8) .85 SWT than their counterparts remain- Six months ing in atrial fibrillation. This suggests that Ejection fraction, % 56 (7) 51 (6) .01 sinus rhythm restoration in patients un- LVESD, cm 3.96 (0.7) 4.33 (0.7) .02 dergoing successful correction of mi- LVEDD, cm 5.77 (0.6) 5.80 (0.7) .80 Maximum left atrial area, cm2 34 (8) 32.4 (9) .38 tral regurgitation further improves func- Minimum left atrial area, cm2 23 (8) 25.6 (8) .11 tional status. Although the mere Twelve months restoration of sinus rhythm may be im- Ejection fraction, % 59 (7) 54.2 (7) .004 portant, perhaps a more relevant index LVESD, cm 3.93 (0.7) 4.26 (0.6) .03 is recovery of atrial contraction. Evi- LVEDD, cm 5.65 (0.6) 5.90 (0.6) .27 dence of recovery of left atrial contrac- Maximum left atrial area, cm2 32 (6) 33.5 (7) .24 tility was observed in 86% of patients Minimum left atrial area, cm2 21 (6) 25 (7) .14 who regained sinus rhythm and could ac- Abbreviations: LVEDD, left ventricular end-diastolic diameter; LVESD, left ventricular end-systolic diameter. count for the improvements in cardiac *Data are reported as mean (SD). dimensions and function and the better exercise capacity following RFA. waves in all Doppler-traced beats with portion of patients who regained si- B-natriuretic peptide is released from values greater than 25 cm/s. nus rhythm 6 months after biatrial RFA the myocardium in response to stretch was approximately 50%, increasing to and is used to monitor the success of vari- COMMENT 82% at 12 months. In a more recent re- ous therapeutic interventions in pa- In this randomized controlled trial, we port, 30 individuals were randomized tients with heart failure.23 We noted that found that left atrial RFA for continu- to receive pulmonary vein isolation, BNP levels were reduced postopera- ous atrial fibrillation during mitral valve maze procedure, or mitral valve sur- tively in both groups as a result of cor- surgery significantly increased the gery alone and were reported to have rective valve surgery with consequent re- prevalence of sinus rhythm 12 months similar likelihood of sinus rhythm res- duction in the filling pressures. However, later without an increase in periopera- toration following pulmonary vein iso- this reduction was significantly more tive morbidity. At 6 months, restora- lation and maze procedure.18 Several prominent in those who converted to si- tion of sinus rhythm was accompa- centers have published nonrandom- nus rhythm. This is in agreement with nied by significant improvements in left ized studies describing success rates in a recent report showing significant de- ventricular dimensions and function, restoring sinus rhythm of 60% to creases in BNP levels following success- satisfactory rates of recovery of left atrial 90%.9-11,19,20 Direct comparison of our ful maze procedure.24 contractility, enhanced exercise capac- findings with those in these studies is A limitation of our study is that our ity, and a significant decrease in BNP difficult, as they lacked randomiza- analysis is based on ECGs obtained at levels. We noted, by comparing pa- tion, often used different tools and le- 3-month intervals and on Holter and tients in the RFA group who did and sion sets, and, in several cases, in- ECG assessment at other times only for did not achieve sinus rhythm, that the cluded patients with intermittent atrial symptoms. Episodes of recurrent atrial additional benefit of sinus rhythm res- fibrillation. fibrillation could have been paroxys- toration in terms of exercise capacity The lack of difference in periopera- mal and asymptomatic, although the is almost equal to that of correcting the tive morbidity and mortality between likelihood of such occurrences in pa- mitral valve pathology. the RFA and control groups in our tients who were previously in continu- To our knowledge, this is the first ad- study attests to the safety and reliabil- ous atrial fibrillation is difficult to es- equately powered randomized clinical ity of the RFA procedure. Our pa- timate. Our study cannot exclude the trial to assess the efficacy of left atrial tients included a significant propor- occurrence of such episodes; to detect RFA during mitral valve surgery in pa- tion of individuals requiring these would have required continu- tients with continuous atrial fibrilla- concomitant coronary artery bypass ous rhythm monitoring for prolonged tion. A previous randomized study in- graft surgery (11%) and/or tricuspid periods, which was not practicable. cluded 30 patients (15 in each arm) valve surgery (17%). In this context, the However, the finding that the major- who underwent biatrial RFA.17 The pro- perioperative mortality in the 2 groups ity of patients who had restored sinus

Downloaded From: https://jamanetwork.com/ on 09/25/2021 LEFT ATRIAL RADIOFREQUENCY ABLATION IN MITRAL VALVE SURGERY

rhythm were in this rhythm at dis- ficacy, our findings suggest that routine Obtained funding: Doukas, Samani, Stafford, Spyt. Administrative, technical, or material support: Doukas, charge and consistently at each fol- use of left atrial RFA during mitral valve Samani, Ng, Spyt. low-up suggests that they were in this surgery is justified. Study supervision: Samani, Alexiou, Chin, Spyt. Financial Disclosures: None reported. rhythm permanently. Author Contributions: Dr Spyt had full access to all Funding/Support: The study was funded through a The search for the optimal treat- of the data in the study and takes responsibility for project grant by the British Heart Foundation. Dr ment of atrial fibrillation accompany- the integrity of the data and the accuracy of the data Samani holds a British Heart Foundation Chair in Car- analysis. diology. ing mitral valve disease will continue, Study concept and design: Doukas, Samani, Chin, Role of the Sponsor: The sponsor had no role in study and further randomized clinical trials Stafford, Spyt. design, data collection, data analysis, data interpre- Acquisition of data: Doukas, Oc. tation, writing of the report, or the decision to sub- are needed to examine the efficacy of Analysis and interpretation of data: Doukas, Samani, mit the report for publication. bipolar ablation,25 which more consis- Alexiou, Ng, Spyt. Acknowledgment: We thank Nick Taub, research Drafting of the manuscript: Doukas, Samani, Alexiou, fellow in Medical Statistics, for providing his expert tently achieves transmural lesions, and Oc. statistical advice and Loraine Ricketts, senior echo- also the role of the less-invasive thora- Critical revision of the manuscript for important in- cardiographer, Julie Hayton, chief CCSO adult tellectual content: Samani, Alexiou, Chin, Stafford, Ng, echocardiography, and Bahar Oc, anesthetist, coscopic interventions. In the mean- Spyt. for performing and analyzing the echocardio- time, on the basis of its safety and ef- Statistical analysis: Alexiou, Oc. grams.

REFERENCES 1. Kannel WB, Abbott RD, Savage DD, McNamara sessment of results at one year. Eur J Cardiothorac Surg. fibrillation and mitral valve disease. Eur Heart J. 2002; PM. Epidemiologic features of chronic atrial fibrilla- 1999;15:851-854. 23:558-566. tion: the Framingham Study. N Engl J Med. 1982;306: 10. Benussi S, Nascimbene S, Agricola E, et al. Sur- 18. de Lima GG, Kalil RA, Leiria TL, et al. Random- 1018-1022. gical ablation of atrial fibrillation using the epicardial ized study of surgery for patients with permanent atrial 2. Benjamin EJ, Wolf PA, D’Agostino RB, Silbershatz radiofrequency approach: mid-term results and risk fibrillation as a result of mitral valve disease. Ann Tho- H, Kannel WB, Levy D. Impact of atrial fibrillation on analysis. Ann Thorac Surg. 2002;74:1050-1057. rac Surg. 2004;77:2089-2094. the risk of death: the Framingham Heart Study. 11. Mohr FW, Fabricius AM, Falk V, et al. Curative 19. Sie HT, Beukema WP, Elvan A, Ramdat Misier AR. Circulation. 1998;98:946-952. treatment of atrial fibrillation with intraoperative ra- Long-term results of irrigated radiofrequency modi- 3. Brodell GK, Cosgrove D, Schiavone W, Under- diofrequency ablation: short-term and midterm results. fied maze procedure in 200 patients with concomi- wood D, Loop F. Cardiac rhythm and conduction dis- J Thorac Cardiovasc Surg. 2002;123:919-927. tant cardiac surgery: six years experience. Ann Tho- turbances in patients undergoing mitral valve surgery. 12. Large SR, Hosseinpour AR, Wisbey C, Wells FC. rac Surg. 2004;77:512-517. Cleve Clin J Med. 1991;58:397-399. Spontaneous cardioversion and mitral valve repair: a 20. Sueda T, Imai K, Ishii O, Orihashi K, Watari M, 4. Lim E, Barlow C, Hosseinpour AR, et al. Influence role for surgical cardioversion (Cox-maze)? Eur J Car- Okada K. Efficacy of pulmonary vein isolation for the of atrial fibrillation on outcome following mitral valve diothorac Surg. 1997;11:76-80. elimination of chronic atrial fibrillation in cardiac repair. Circulation. 2001;104:I59-I63. 13. Morales FJ, Martinez A, Mendez M, et al. A valvular surgery. Ann Thorac Surg. 2001;71:1189- 5. Chua YL, Scaff HV, Orszulak TA, Morris JJ. Out- shuttle walk test for assessment of functional capac- 1193. come of mitral valve repair in patients with preopera- ity in chronic heart failure. Am Heart J. 1999;138: 21. Izumoto H, Kawazoe K, Kitahara H, Kamata J. Op- tive atrial fibrillation: should the maze procedure be 291-298. erative results after Cox/maze procedure combined combined with mitral valvuloplasty? J Thorac Cardio- 14. Singh SJ, Morgan MD, Scott S, et al. Develop- with mitral valve operation. Ann Thorac Surg. 1998;66: vasc Surg. 1994;107:408-415. ment of a shuttle walking test of disability in patients 800-804. 6. Jalife J, Berenfeld O, Mansour M.. Mother rotors with chronic airways obstruction. Thorax. 1992; 22. Yau TM, Farag-el-Ghoneimi YA, Armstrong S, Iva- and fibrillatory conduction: a mechanism of atrial 47:1019-1024. nov J, David TE. Mitral valve repair and replacement fibrillation. Cardiovasc Res. 2002;54:204-216. 15. Ng LL, O’Brien RJ, Demme B, Jennings S. A non- for rheumatic disease. J Thorac Cardiovasc Surg. 2000; 7. Cox JL, Boineau JP, Schuessler RB, Kater KM, Lap- competitive immunochemiluminometric assay for car- 119:53-61. pas DG. Five-year experience with the maze proce- diotrophin-1 detects elevated plasma levels in hu- 23. Levin ER, Gardner DG, Samson WK. Natriuretic dure for atrial fibrillation. Ann Thorac Surg. 1993;56: man heart failure. Clin Sci. 2002;102:411-416. peptides. N Engl J Med. 1998;339:321-328. 814-823. 16. Parsonnet V, Dean D, Bernstein AD. A method 24. Watanabe M, Murakami M, Furukawa H, Naka- 8. Khargi K, Deneke T, Haardt H, et al. Saline- of uniform stratification of risk for evaluating the re- hara H, Tanaka H, Sunamori M. Decreased plasma brain irrigated, cooled-tip radiofrequency ablation is an ef- sults of surgery in acquired adult heart disease. natriuretic peptide levels after a successful maze fective technique to perform the maze procedure. Ann Circulation. 1989;79(6 pt 2):I3-I12. procedure. J Heart Valve Dis. 2003;12:287-291. Thorac Surg. 2001;72:S1090-S1095. 17. Deneke T, Khargi K, Grewe PH, et al. Efficacy of 25. Khargi K, Hutten BA, Lemke B, Deneke T. Surgi- 9. Melo J, Adragao P, Neves J, et al. Surgery for atrial an additional MAZE procedure using cooled-tip ra- cal treatment of atrial fibrillation: a systematic review. fibrillation using radiofrequency catheter ablation: as- diofrequency ablation in patients with chronic atrial Eur J Cardiothorac Surg. 2005;27:258-265.

Downloaded From: https://jamanetwork.com/ on 09/25/2021