In this issue of Circulation, Enriquez-Sarano and his associates1 at the Mayo Clinic have made a valuable contribution to our understanding of mitral valve repair in relation to mitral valve replacement. In an elegant and detailed multivariate analysis, they conclude that mitral valve repair is an independent predictor of improved operative long-term mortality, ejection fraction, and functional result. These differences were confirmed after statistical allowance for multiple variables, including year of surgery, preoperative ejection fraction, age, sex, preoperative New York Heart Association classification, presence of congestive heart failure, atrial fibrillation, renal status, blood pressure, and concomitant coronary artery bypass surgery.
While this conclusion had been assumed intuitively by cardiovascular physicians, particularly during the last decade as the techniques of valve repair improved, previous studies had not confirmed that mitral valve repair produced better outcomes than valve replacement.2 3 The authors of the present study offer a statistically powerful argument on the basis of the size and homogeneity of their patient population and of the uniform collection of patient data from preoperative ventricular function studies. Their findings dispel the suspicion that mitral valve repair produces better results solely because patients undergoing repair have better functional status before surgery. Indeed, their data do confirm the better preoperative functional status, ejection fraction, and other prognostic factors in the repair group. Furthermore, operative mortality in fact was dramatically higher in the replacement group (10.3% versus 2.6%), which reflected primarily the difference in preoperative status and confirmed that preoperative variables do result in a better outcome for mitral repair. Nonetheless, after statistically accounting for the bias favoring repair, the 10-year overall survival rate was better for valve repair (68%) than for replacement (52%) and approaches that expected in the normal population.
The patient population studied is necessarily a somewhat select group. Specifically, the total series of 409 patients was culled from the authors’ total of 654 mitral valve operations performed during the study period. Excluded were patients with previous valve surgery and those with regurgitation of ischemic or functional origin. However, patients undergoing concomitant coronary artery bypass were not excluded. The series, therefore, consists almost exclusively of patients with primary degenerative disease, the so-called “end-stage prolapse syndrome.” The valve lesion in this syndrome is particularly amenable anatomically to repair, whereas valve lesions secondary to endocarditis, ischemia, or rheumatic disease are less so. Therefore, the selection of patients with primary degenerative disease may account for some bias in favor of repair in the series. In addition, the percentage of St Jude or other disk prostheses used in the valve replacement group (26 of 214, 12%) is, I believe, much lower than at other institutions worldwide; conversely, the percentage of Starr-Edwards ball valves used (84 of 214, 39%) is much higher than at other institutions. What effect the choice of prosthesis has on the results of replacement can only be conjectured, but it may be significant. Of interest also is the higher percentage of concomitant coronary artery bypass procedures in the valve repair group (29%) than in the replacement group (20%) (P=.024). This finding suggests that coronary disease may have been the primary indication for surgery in the repair group, whereas valve disease may have been the more severe indication for surgery in the replacement group.
Reoperation rates were similar and significant for both the repair and replacement patients (10% at 5 years), but they indicate the predictability of repair using current surgical techniques. Repair thus appears particularly attractive when one considers that (1) the rates of thromboembolism and endocarditis were similar between groups, (2) only half as many repair as replacement patients required Coumadin (P=.0001), and (3) freedom from hemorrhage complications was significantly better in the repair group (P=.002). Also important is the observation that although ejection fraction decreased after either valve repair or replacement, repair was an independent predictor of a higher postoperative ejection fraction regardless of preoperative status. Ejection fraction remained stable with time in both groups.
In summary, we owe a debt of gratitude to the authors for clearly elucidating the role of mitral valve repair. From my own surgical perspective and experience with over 300 mitral valve repair procedures, it is apparent that there is an early learning curve in obtaining predictable results, certainly more so than in simple valve replacement surgery. In mitral repair procedures, the art of surgery is preeminent, and there is no substitute for experience. Certainly, the authors’ experience, as much as any other factor, accounts for their excellent results in this enlightening report.
The opinions expressed in this editorial are not necessarily those of the editors or of the American Heart Association.
- Received December 28, 1994.
- Accepted December 29, 1994.
- Copyright © 1995 by American Heart Association