Outcome of mitral valve repair in patients with preoperative atrial fibrillation

Outcome of mitral valve repair in patients with preoperative atrial fibrillation Should the maze procedure be combined with mitral valvuloplasty? To examine late outcome of mitral valve repair in patients with preoperative atrial fibrillation, we reviewed the cases of 323 consecutive patients who underwent mitral valvuloplasty for mitral regurgitation from 1980 to 1991; average age of 215 men and 108 women was 64 years (range 14 to 88 years), and 224 patients (70 %) were in New York Heart Association class ill or IV before operation. The main indications for operation were severe mitral regurgitation (76 %), coronary artery disease with associated mitral regurgitation (15 % and aortic valve disease (6 %). At the time of mitral valve repair, coronary artery bypass grafting was done in 35 % of patients, aortic valve replacement was done in 7 %, and multiple other procedures were done in 10 %. For all patients, the 30-day mortality rate was 2.5% (70% confidence limits 1.6% to 3.4%) and survivorships at 3 and 5 years were 81 % and 76%, respectively. Before operation, 216 patients were in sinus rhythm and 97 patients had atrial fibrillation; in the latter group, 11 had recent onset of atrial fibrillation within 3 months preceding mitral valve repair. Comparing patients with preoperative atrial fibrillation to those with sinus rhythm, we found no significant difference in operative mortality (3 % versus 1.9 %) or 5-year survivorship (74.3 % ± 6.3 % versus 76.9 % ± 4.0 %). At late follow-up, atrial fibrillation was present in 5 % of patients with preoperative sinus rhythm, 80 % of patients with preoperative chronic atrial fibrillation, and 0% of patients with preoperative recent onset atrial fibrillation (p < 0.01). The left atrial size by echocardiography was larger in patients with preoperative atrial fibrillation compared with that in those with sinus rhythm (59 ± 1.4 mm versus 50.9 ± 0.7 mm; P < 0.05). There was, however, only a weak correlation between preoperative left atrial size and late atrial fibrillation. Further, age, gender, and associated coronary artery disease did not correlate with presence of atrial fibrillation at late follow-up. Prevalence of late thromboembolic events was similar in patients with preoperative sinus rhythm compared with that in those with atrial fibrillation. These data suggest that mitral valve repair should be done before or soon after the onset of atrial fibrillation to maximize the chance of postoperative sinus rhythm and avoid long-term anticoagulation with warfarin. However, the early and late results of mitral valve repair in patients with chronic atrial fibrillation are good, and concomitant operation for supraventricular arrhythmia must have negligible morbidity and no adverse effect on operative mortality. (J THORAe CARDIOVASC SURG 1994;107:408-15) Yeow L. Chua, MD (by invitation), Hartzell V. Schaff, MD, Thomas A. Orszulak, MD, and James J. Morris, MD (by invitation), Rochester, Minn. From the Section of, Mayo Clinic and Mayo Foundation, Rochester, Minn. Read at the Seventy-third Annual Meeting of The American Association for Thoracic Surgery, Chicago, Ill., April 25-28, 1993. Address for reprints: Hartzell V. Schaff, MD, Division of Thoracic and, Mayo Clinic, Rochester, MN 55905. Copyright @ 1994 by Mosby-Year Book, Inc. 0022-5223/94 $1.00 +.10 12/6/51140 Mitral valve repair is the preferred procedure for most patients with severe mitral valve regurgitation as a result of degenerative disease. 1-3 Preservation of the native valve and its apparatus confers many advantages over valve replacement including reduced early mortality and morbidity, and a particular benefit is low risk of thromboembolism with avoidance of anticoagulation in many patients. 3,4 Becauseof these advantages and the predictably good hemodynamic results of mitral valve 408

The Journal of Thoracic and Volume 107, Number 2 Chua et at. 4 0 9 Table I. Etiology of regurgitation in patients undergoing mitral valve repair Mitral diagnosis No. of patients Ruptured chordae 137 Prolapse 91 Ischemic 55 Endocarditis 15 Rheumatic 11 Cleft 10 Ruptured papillary muscle 2 Trauma I Failed repair (hemolysis) Total 323 Includes patients with idiopathic annular dilation. Percent 42.4 28.2 17.0 4.6 3.4 3.1 0.6 0.3 0.3 Table II. Operations done in conjunction with mitral valve repair No. of Concomitant procedures patients Percent CABG 113 35 AVR 21 6.5 Tricuspid valve annuloplasty 13 Tricuspid valve replacement 6 1.9 ASD closure 5 1.7 ICD implant 3 I LV myectomy 3 1 LV aneurysmectomy 2 0.7 AVR, Aortic valve replacement; ASD, atrial septal defect; lcd, implantable cardioverter defibrillator; L V, left ventricular. repair without the need for prosthetic replacement, many clinicians now offer valvuloplasty to patients with severe mitral regurgitation (MR) early in the course of their disease to prevent progressive left ventricular dilation and development of atrial fibrillation. Despite such strategy, reports suggest that 40% to 60% of patients who undergo mitral valve operation are in atrial fibrillation at the time of operation, and experience has shown that many patients remain in atrial fibrillation after operation despite reduction or elimination of MR. 5 Chronic atrial fibrillation is associated with increased risk of thromboembolism in the presence or absence of valvular heart disease. Little is known, however, a bout the extent to which atrial fibrillation compromises early and late results of mitral valve repair. Advances in the understanding of the mechanisms of atrial fibrillation and development of the maze procedure for surgical treatment have radically altered the prognosis for many patients with chronic atrial fibrillation. 6-8 The low risk of mitral valvuloplasty and the encouraging preliminary results of operation to prevent atrial reentry tachycardia have stimulated interest in combining the procedures for patients with MR. However, little is known about the influence of atrial fibrillation on early and late results of valve repair for mitral insufficiency. Therefore, we undertook the present study to determine the prevalence of preoperative atrial fibrillation among patients undergoing mitral valve repair in the current era; the investigation focused on predictors of return to sinus rhythm after operation, and we evaluated the effect of atrial fibrillation on early and late mortality and morbidity. Patients and methods From July 12, 1979, through Nov. 20,1991,323 consecutive patients underwent mitral valve repair with or without associatedprocedures by two oftheauthors (H.V.S. and T.A.o.). For this study, we reviewed operative notes, clinical case histories, and laboratory investigations including electrocardiograms, echocardiograms, and cardiac catheterization data. Status of patients was determined by clinic records of outpatient visits and correspondence with referring physicians, and follow-up focused on vital status, cardiac rhythm, and occurrence of neurologic events after operation. All autopsy reports or death certificates were reviewed, and a total of 79 clinical, hemodynamic, electrocardiographic, and echocardiographic variables were entered into a computerized database and analyzed. Patients not seen within 3 months were contacted with letters and telephone review. Recent follow-up of patients (within 3 months of data analysis) was obtained in 95.1 % of patients (16 patients were not available for late follow-up, and 57 patients [17.6% 1 died). Postoperative neurologic events included all possible thromboembolic episodes, both transient ischemic attacks and strokes. The mean follow-up interval was 2.6 years (range 3 months to 10 years). Statistical analysis. Data are expressed as mean plus or minus one standard deviation, and selected proportions are presented with their 70% confidence limits (CL). Continuous variables in patients with and without preoperative atrial fibrillation were compared by unpaired or paired Student's t test (two-sided) and analysis of variance, and discrete variables were compared by the x 2 test. Logistic regression analysis was used to determine the relationship between early mortality and univariate predictors of early death. A similar method was used to analyze predictors of return to sinus rhythm during follow-up. Survival estimates were made with the Kaplan-Meier product limit method and expressed with the standard error. Comparisons between survival distributions were made with the logrank and Gehan's Wilcoxon tests. Influence of multiple covariates on survival were analyzed with Cox's proportional hazards regression model. Differences were considered statistically significant when p :::; 0.05. Results Patients. The average age of 215 men and 108 women was 64 years (range 14 to 88 years), and 224 patients (69.3%) were in New York Heart Association (NYHA) class III or IV before operation. The principal indication for operation was severe MR in 76.2% of patients, coronary artery disease with associated severe MR in 14.6%,

4 1 0 Chua et al.,... -. _, -II-._-. ------ -._-.... 1_- T 80 - _ --1..- --1" --'---1 - as 60 1'----- > > 40 -...- <67 yr """ :J P<.0001 en 20 -.-?67 yr The Journal of Thoracic and February 1994 o ------------------------ o 1 2 3 4 5 6 Time postop, yr Fig. 1. Survival of patients after mitral valve repair stratified according to age at operation. This and subsequent survival curves include all early and late deaths. Table III. Characteristics of patients with preoperative sinus rhythm and atrial fibrillation who underwent mitral valve repair Preop. sinus rhythm Percent Preop. AF Percent No. of patients Sex (male) Mean age (yr)* CABG No. of other procedures NYHA class III and IV Preop. EF (%) Postop. EF (0/0) Preop. LA size (mm) Postop. LA size (mm) Mortality IABP Periop. stroke Hospital days, median Five-year survival Late atrial fibrillation* Stroke, per patient-year 215 140 62.7 ± 13.6 79 46 145 57.2 ± 14.2t 50.3 ± 12.3* 51.1 ± 8.1 t 46 ± 7.9t 4 7 10 9 65 36.7 21.4 67.4 1.9 3.2 4.7 77 5 4.6 97 67 67.9 ± 10.6 27 25 73 55.7 ± 13.2t 48.2 ± 12.5* 58.9 ± 12.1 t 55.4 ± l1.3t 3 4 3 10 69 27.8 25.6 75 3.0 4.1 3.1 75 70 3.6 AF, Atrial fibrillation, Preap., preoperative; Pastap., postoperative, IABP, intra aortic balloon pump; Periop., perioperatively. *p < 0.05 (between sinus and atrial fibrillation group). tp < 0.05 (between preoperative and postoperative measurements). and aortic valve disease with associated severe MR in 5.9%. Before operation, cardiac rhythm was sinus in 216 patients (66.6%) and atrial fibrillation in 97 patients (30%,70% CL 27.4% to 32.7%); eight patients had permanent pacemakers inserted at various intervals before mitral valve repair, and in two patients the rhythm before operation was indeterminate. Among the patients with preoperative atrial fibrillation, 11 had recent onset of atrial fibrillation within the 3 months immediately before mitral valve repair. The various mitral valve pathologic conditions are listed in Table I, and concomitant cardiac surgical procedures are listed in Table II. Twenty-three patients had previous cardiac operation, of which 16 (69.6%) had coronary artery bypass grafting (CABG) and five patients (21.7%) had aortic valve replacement; one patient had

The Journal of Thoracic and Volume 107, Number 2 Chua et al. 4 1 1.-,..., - --&-T -- 1.11 -- --I T!!... --, ---_ T ' ----, 1, - T - CIS 60 > > 40 """ en 80 ----I '.. - 20 -.. - No CAB -.- CAB, ---i------ T -t P<.0001 o -------------------- o 1 2 3 4 5 6 Time postop, yr Fig. 2. Survival of patients after mitral valve repair with or without concomitant coronary artery bypass (CAB). previous excision of left atrial myxoma, and one had CABG and mitral valve repair. Three patients had two previous cardiac operations. The patient with previous CABG and mitral valve repair had prior operation elsewhere and underwent reoperation because of severe hemolytic anemia. Operations were done with the use of standard cardiopulmonary bypass with moderate hypothermia (25 0 C to 28 0 C), and cold crystalloid or blood cardioplegia was used for myocardial protection. Beginning in 1991, normothermic cardiopulmonary bypass and warm cardioplegia has been used. The mitral valve was approached through a standard left atriotomy made just posterior to the interatrial groove. Methods of valve repair were individualized as necessary. Leaflet and chordal abnormalities were corrected first. Fifty-five patients (17%) had plication of redundant (floppy) anterior mitral valve leaflets, and 162 (49.5%) had plication of redundant posterior mitral valve leaflets. Portions of posterior mitral valve leaflets were excised in 62 patients (19.2%), and a portion of redundant anterior mitral valve leaflet was excised in one patient. Floppy leaflets were treated by chordal shortening in five patients. Annuloplasty was done to reduce annular dimension and protect leaflet repair in 311 patients (96.3%). Most annuloplasties (206, 63.8%) used a flexible (Duran) or rigid (Carpentier) prosthetic ring. 9, 10 The most recent 93 patients (28.8%) in which a flexible prosthesis was used had only a partial (posterior) ring inserted. Anulus dimension was reduced by plication of one or both commissures in 105 patients (32.6%), and most of these patients had operation during the early portion of this series. All patients received systemic anticoagulation with warfarin (Coumadin) for the first 6 weeks after operation, and this was usually begun on the second or third postoperative day. We recommended to patients and their referring physicians that anticoagulation be discontinued after the early (6-week) postoperative period if the rhythm was sinus and there had been no thromboembolic events. Atrial fibrillation commonly developed in patients during hospitalization; patients who were in sinus rhythm before operation underwent cardioversion to sinus rhythm before hospital dismissal. The decision to electively perform cardiovesion in patients late after operation was made by the primary physician or cardiologist. Comparison of patients with and without atrial fibrillation. Clinical and hemodynamic data for patients with and without preoperative atrial fibrillation are summarized in Table III. It is notable that the mean age of patients in the preoperative atrial fibrillation group was higher than that of patients with sinus rhythm (68 ± 10.5 years versus 62.7 ± 13.7 years; p < 0.014). The two groups were similar as regards gender distribution and degree of preoperative disability (NYHA class), however, slightly more patients with sinus rhythm had concomitant CABG (36.7% versus 27.8% in atrial fibrillation, p = 0.12). Before operation, left atrial dimension was larger in patients with atrial fibrillation compared with that in those with sinus rhythm (58.7 ± 12.1 mm versus 51 ± 7.9 mm,p < 0.001), and left atrial size was significantly lower in both groups at the time of hospital dismissal after correction of mitral insufficiency (Fig. 1). For all patients, the left ventricular ejection fraction as

4 1 2 Chua et al. The Journal of Thoracic and February 1994 80 *i'.,.i"... -... -... -I,-------'':' ----I _-._!-..L.. _.---... 60... T - fts > a.. > 40 -...- EF;:;45% :::::s en 20 -.- EF<45% T..,-------'.- A 1 ----_, -t P<.0001 o o 1 2 3 4 5 6 Time postop, yr Fig. 3. Survival of patients after mitral valve repair stratified by preoperative ejection fraction (EF). measured by echocardiography fell from 56.8% ± 13.7% to 49.5% ± 12.2% (p < 0.05). No patients died intraoperatively, and the overall 30-day mortality rate was 2.5% (70% CL 1.6% to 3.4%); early mortality was similar for patients in preoperative sinus rhythm to the early mortality for those with atrial fibrillation (1.9% and 3.0%). Operative risk was not increased in patients with previous operation; the only death in this group occurred in a patient who had two previous operations for CABG. Occurrence of postoperative low cardiac output was not influenced by preoperative atrial fibrillation as evidenced by similar prevalence of intraaortic balloon use (3.2% of patients in preoperative sinus rhythm and 4.1 % of patients with atrial fibrillation). Major complications, including all degrees of neurologic complications, were not significantly different between the two groups. For all patients, 3-, and 5-year survival estimates were 81.2% ± 2.6% and 75.7% ± 3.4%, respectively. As shown in Fig. 1 through 3, late survival was strongly influenced by advanced age at operation, need for CABG at the time of mitral valve repair, and low ejection fraction before operation. Cox multivariate analysis demonstrated that these variables were independently predictive of late death. In contrast, late survival was similar in patients with preoperative sinus rhythm compared with that in those with atrial fibrillation (5-year survival 76.9% ± 4.0% and 74.3% ± 6.3%, respectively; Fig. 4). The incidence of late thromboembolic events was 4.6% per patient-year for patients with preoperative sinus rhythm and 3.6% per patient-year for those with preoperative atrial fibrillation. We used a liberal definition of thromboembolic events, which included all transient events such as weakness, dizziness, and vismil field changes. Exactly half of all late events in each group could be classified as minor or transient events. The most important determinant of rhythm at followup was preoperative rhythm. At late follow-up, atrial fibrillation was present in 5% of patients with preoperative sinus rhythm, 80% of patients with preoperative chronic atrial fibrillation, and 0% of patients with preoperative recent-onset atrial fibrillation (p < 0.01). Age, gender, and associated coronary artery disease did not correlate with presence of atrial fibrillation at late followup. Preoperative left atrial diameter was only weakly predictive of late sinus rhythm on multivariate analysis (p = 0.054); among patients with preoperative atrial fibrillation, left atrial diameter was 55.2 ± 10.2 mm in those patients who continued to have atrial fibrillation compared with 53.2 ± 8.6 mm in patients who achieved sinus rhythm (p = 0.32). The prevalence of reoperation during follow-up was 2.5% for the entire group and was similar in patients with or without preoperative atrial fibrillation. All patients in whom reoperation was done because of hemodynamically significant MR had repair of the anterior mitral valve leaflet at the initial operation. One patient had reoperation to correct severe hemolysis caused by a regurgitant jet hitting the annuloplasty ring. Discussion In this study, 30% of patients were in atrial fibrillation at the time of mitral valve repair. This prevalence is slightly lower than might be expected from prior reports

The Journal of Thoracic and Volume 107, Number 2 Chua et al. 4 1 3 80 C; 60 > 40 ::I en 20.,,--. I... -_._...-......; -A- Sinus -.- AF ------1. o o 1 2 3 4 5 6 Time postop, yr Fig. 4. Survival of patients after mitral valve repair stratified according to preoperative rhythm. AF, Atrial fibrillation. of mitral valve operations, however, we believe that the incidence of preoperative atrial fibrillation in our series is representative of patients having mitral valve repair in North America. In current practice, the population of patients referred for mitral valvulopasty is elderly, has a low prevalence of rheumatic heart disease, and frequently has coronary artery atherosclerosis either in association with degenerative mitral valve disease or as the cause of mitral valve insufficiency. These patient characteristics have also been reported in surgical specimens of mitral valves excised in the course of prosthetic replacement for mitral insufficiency.l1 Maintenance of sinus rhythm is an important goal in the treatment of patients with chronic MR, and the serious long-term effects of chronic atrial fibrillation are documented. In the Framingham Heart Study, total mortality and mortality from cardiovascular causes were almost twice as high among patients with atrial fibrillation as among those without it, and the rate of stroke was increased among patients who had atrial fibrillation, whether or not they had rheumatic heart disease. 12, 13 Even in the absence of clinically apparent stroke, patients with chronic atrial fibrillation have a high incidence of abnormal findings on computerized tomography of the brain, indicating prior silent cerebral infarction. 14 For patients with preoperative sinus rhythm having valve repair for mitral insufficiency, the chance of maintaining sinus rhythm late after operation is excellent (95%), and chronic anticoagulation with warfarin may not be necessary for most patients. Further, our study suggests that a similarly high percentage of patients with recent onset of atrial fibrillation before operation (less than 3 months) will achieve sinus rhythm during late follow-up. This finding may be useful in counseling patients in regard to timing of valve repair, especially those who are reluctant to consider operation because of little or no clinical disability. Decision-making in the treatment of patients with chronic atrial fibrillation and mitral valve insufficiency is more complex. In our experience, approximately 20% of these patients will achieve sinus rhythm after successful valve repair, but we could not identify preoperative variables that would accurately predict return to sinus rhythm late after operation. IS, 16 Left atrial size was greater in patients with atrial fibrillation compared with the size in those with preoperative sinus rhythm, and there was a trend toward smaller atrial size in patients with preoperative atrial fibrillation who converted to sinus rhythm after operation; however, this difference was small,and the considerable overlap in data limits its predictive value. In our experience, preoperative atrial fibrillation had little influence on early or late mortality after mitral valve repair, and there are several possible explanations for this finding. First, our patients were heterogeneous as regards diagnosis and extent of preoperative functional impairment. Clearly, factors such as advanced age, presence of coronary atherosclerosis, and left ventricular dysfunction were powerful predictors oflate death. These factors were not evenly distributed between patients with preoperative atrial fibrillation and those with preoperative sinus rhythm, but the older age in one group might be expect-

4 1 4 Chua et al. The Journal of Thoracic and February 1994 ed to offset the slightly higher prevalence of coronary artery disease in the other. Also, multivariate analysis of late death did not suggest that preoperative atrial fibrillation affected late survival. Perhaps the most important potential advantage of maintenance of sinus rhythm after mitral valve repair is reduction in risk of late thromboembolism and elimination of the need for systemic anticoagulation with warfarin. In our patients, late occurrence of thromboembolic events was similar in patients with preoperative atrial fibrillation compared with that in those with sinus rhythm. All patients received systemic anticoagulation for the first 6 weeks after operation, and this was continued in most patients with atrial arrhythmias. Thus the neutralization of atrial fibrillation as a risk factor for late thromboembolism was achieved by the added risk of late bleeding from chronic warfarin therapy. Although our review did not detail the incidence of late bleeding episodes, many studies from the literature indicate that this risk is 1.5% to 2.5% per patient-year. 17 Again, the present review included a variety of patients having mitral valve repair including patients with concomitant aortic valve operation and many with atherosclerotic vascular disease; these patients might be expected to have increased risk of late neurologic events, and including them in the calculations may have masked small differences. However, subgroup analysis of only the patients whose primary indication for valve repair was severe mitral insufficiency did not suggest any difference in late risk of thromboembolism between the two groups. Another weakness in our review that should be acknowledged is lack of information on the status of anticoagulation of patients at the time of late thromboembolism. It was not possible for us to determine late risk of thromboembolism according to intensity of anticoagulation. In conclusion, performance of mitral valve repair before or soon after the onset of atrial fibrillation maximizes the chance of late postoperative sinus rhythm and reduces the need for long-term anticoagulation with warfarin. For patients with chronic atrial fibrillation, however, the decision for concomitant procedures to control atrial fibrillation should be made with the realization that early and intermediate-term outcome of mitral valve repair in such patients is relatively good, and it follows that additional operation for supraventricular arrhythmias must have negligible morbidity and no adverse effect on overall mortality. Thus, for the present, it seems prudent to restrict concomitant procedures for atrial fibrillation to those patients with chronic arrhythmias who are young and would be expected to have a long exposure to the risks of anticoagulation, and, possibly to patients who have contraindications to long-term anticoagulation or who have symptoms from rapid supraventricular arrhythmias not well controlled medically. REFERENCES 1. Cohn LH. Surgery for mitral regurgitation. JAMA 1988; 260:2883-7. 2. Galloway AG, Colvin SB, Baumann FG, Harty S, Spencer FC. Current concepts of mitral valve reconstruction for mitral insufficiency. Circulation 1988;78: 1087-98. 3. Loop FD. Long-term results of mitral valve repair. Semin Thorac Cardiovasc Surg 1989;1:203-10. 4. Orszulak TA, Schaff HV, Danielson GK, et al. Mitral regurgitation due to ruptured chordae tendineae: early and late results of valve repair. J THORAC CARDIOVASC SURG 1985;89:491-8. 5. Brodell GK, Cosgrove D, Schiavone W, Underwood DA, Loop FD. Cardiac rhythm and conduction disturbances in patients undergoing mitral valve surgery. Cleve Clin J Med 1991;58:397-9. 6. Cox JL, Schuessler RB, Boineau JP. The surgical treatment of atrial fibrillation: summary of the current concepts of mechanisms of atrial flutter and atrial fibrillation. J THORAC CARDIOVASC SURG 1991;101:402-5.. 7. Cox JL, Canavan TE, Schuessler RB, et al. The surgical treatment of atrial fibrillation: intraoperative electrophysiological mapping and description of electrophysiological basis of atrial flutter and atrial fibrillation. J THORAC CAR DIOVASC SURG 1991;101:406-26. 8. Cox JO, Schuessler RB, D' Agostino HJ Jr, et al. The surgical treatment of atrial fibrillation: development of a definitive surgical procedure. J THORAC CARDIOV ASC SURG 1991;101:569-83. 9. Duran CG, Pomar JL, Revuelta JM, et al. Conservative operation for mitral insufficiency: critical analysis supported by postoperative hemodynamic studies of 72 patients. J THORAC CARDIOVASC SURG 1980;79:326-37. 10. Carpentier A. Cardiac valve surgery: the "French correction." J THORAC CARDIOVASC SURG 1983;86:323-37. 11. Olson LJ, Subramanian R, Ackermann DM, Orszulak TA, Edwards WD. Surgical pathology of the mitral valve: a study of 712 cases spanning 21 years. Mayo Clin Proc 1987;62:22-34. 12. Kannel WB, Abbott RD, Savage DD, McNamara PM. Epidemiologic features of chronic atrial fibrillation: the Framingham Study. N Engl J Med 1982;306:1018-22. 13. WolfPA, Dawber TR, Thomas HE Jr, Kannel WB. Epidemiologic assessment of chronic atrial fibrillation and risk of stroke: the Framingham study. Neurology 1978;28: 973-7. 14. Petersen P, Madsen EB, Brun B, Pedersen F, Gyldensted C, Boysen G. Silent cerebral infarction in chronic atrial fibrillation. Stroke 1987;18:1098-. 15. Fulgelman MY, Hasin Y, Katznelson N, Kriwisky M, Shefer A, Gotsman MS. Restoration and maintenance of sinus rhythm after mitral valve surgery for mitral stenosis. Am J CardioI1984;54:617-9.

The Journal of Thoracic and Volume 107, Number 2 Chua et al. 4 I 5 16. Sanfilippo AJ, Abascal YM, Sheehan M, et al. Atrial enlargement as a consequence of atrial fibrillation: a prospective echocardiographic study. Circulation 1990;82: 792-7. 17. Grunkemeier GL, Starr A, Rahimtoola SH. Prosthetic heart valve performance: long-term follow-up. Curr Probl Cardiol 1992;37:383-6. Discussion Dr. Yasunaru Kawasbima(Osaka, Japan). Dr. Chua, I am very much impressed with your fine analysis and agree with your conclusion. We have studied the patients in whom we have performed the maze procedure since 1992. These operations were done by Dr. Kosakai and his group in my hospital. There have been 43 patients and 35 of the cases were mitral valve-related. The duration of atrial fibrillation was, on an average, 8 years. Operation for mitral valve disease was done in all 35 patients. Repair was done in 17 and replacement in 18, and concomitant repair or replacement for other valves was also done in 19 patients. There was no operative and no late mortality. Sinus rhythm was restored in 31 patients and only two patients remained in atrial fibrillation. I think this result supports our program to continue the maze operation in combination with mitral valve operation. Dr. Lawrence I. Bonchek (Lancaster, Pa.). This is a nice study and I hope I will be forgiven if I refer to this as the Coxmaze procedure, which is something I have done in a recent publication. I think Dr. Cox should get all the credit for it. I was a little disappointed that the data did not include any information about the functional capacity of the patients. One of the major advantages of the Cox-maze procedure is to improve hemodynamics and left atrial transport function, and although it is not surprising that there was no obvious difference in mortality, there may well be a difference in the patient's functional capacity, not only in obvious ways but in subtle ways as well. Is there any information about the functional capacity in these patients? Dr. James L. Cox (St. Louis, Mo.). I would like to take the liberty to make a comment or two about this paper. I think this type of information is extremely important because most of the combined procedures that are being done around the world now are no more than anecdotal experiences with the possible exception of the group in Osaka and Jantene's group in Sao Paulo, Brazil. I know of about 50 patients, and now with these 35, maybe 60 patients who have undergone combined procedures, and I do not think solid statements can be made with a short follow-up period in 60 patients. However, our own practice has been to limit severely the number of the patients in whom we do concomitant procedures, maze and mitral valve repair, because we really do not know the answer to the question that I think may have been answered today. Our practice has been to ignore the atrial fibrillation if it has been present for less than a year preoperatively. The authors have suggested in their abstract and in the presentation that less than 3 months is appropriate. One of the things that bothers me a bit about the data is the fact that conclusions are made on follow-up information in terms of longevity and transient ischemic attacks and other events on the basis of the preoperative arrhythmia, not the postoperative arrhythmia, and I am not sure that it is absolutely fair to do that. I would like the authors' comments on whether or not they have analyzed the data on the basis of what the postoperative rhythm is. Finally, I would concur entirely with what Dr. Bonchek said about the functional status of the patients. I think that the end points of death, longevity, and thromboembolism are perfectly valid, but I think that quality of life might also be included. Dr. Cbua. In response to Dr. Bonchek's question of functional capacity, about 80% of our patients were in NYHA class I or II at follow-up; however, we did not stratify patients according to the rhythm, and this should be done. Dr. Cox questioned whether we should look at follow-up rhythm rather than preoperative rhythm. We did stratify patients by follow-up rhythm, and the results turned out to be the same.