Ischemic Versus Degenerative Mitral Regurgitation: Does Etiology Affect Survival?

Transcription

1 Ischemic Versus Degenerative Mitral Regurgitation: Does Etiology Affect Survival? A. Marc Gillinov, MD, Eugene H. Blackstone, MD, Jeevanantham Rajeswaran, MS, Maurice Mawad, MD, Patrick M. McCarthy, MD, Joseph F. Sabik III, MD, Takahiro Shiota, MD, Bruce W. Lytle, MD, and Delos M. Cosgrove, MD Departments of Thoracic and Cardiovascular Surgery, Biostatistics and Epidemiology, and Cardiovascular Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio Background. Ischemic mitral regurgitation (MR) is associated with poor survival and degenerative MR with excellent survival. We hypothesized that in some patients with degenerative MR requiring concomitant coronary artery bypass grafting (CABG), ischemic disease would dominate prognosis, resulting in survival as poor as in patients with ischemic MR. Thus, we (1) determined survival impact of etiology (degenerative vs ischemic) after combined mitral valve repair and CABG and (2) explored survival differences within etiology groups. Methods. From 1985 to 2003, 710 patients underwent mitral valve repair for degenerative MR and concomitant CABG (two diseases); 400 patients had mitral annuloplasty and CABG for functional ischemic MR (one disease). Patients were propensity-matched on demography, symptoms, comorbidities, coronary artery disease, and left ventricular function. Survival was compared between matched groups and within groups. Results. Compared with patients with degenerative MR, those with ischemic MR had more extensive coronary artery disease, worse ventricular function, more comorbidities, and more symptoms (p < 0.05). Unadjusted 5-year survivals were 64% and 82% for patients with ischemic and degenerative MR, respectively. However, 123 ischemic and degenerative MR matched pairs had equivalently poor 5-year survival (p > 0.9), 66% and 65%, respectively. Among patients with degenerative MR, survival varied widely, depending largely on ischemic burden and extent of left ventricular dysfunction. Conclusions. The large survival discrepancy between patients with ischemic and degenerative MR is attributable to differences in patient profile, particularly extent of ischemic disease and left ventricular dysfunction. Thus, ischemic and degenerative MR patients with equivalent characteristics have equivalently poor survival. (Ann Thorac Surg 2005;80:811 9) 2005 by The Society of Thoracic Surgeons Operations that include both mitral valve surgery and coronary artery bypass grafting (CABG) have been associated with high hospital mortality and limited longterm survival [1 9]. However, most studies of these procedures contain heterogeneous patient groups, making it difficult to isolate the impact of mitral valve disease etiology on survival when mitral valve dysfunction and coronary artery disease coexist [4, 6, 9]. In general, ischemic mitral regurgitation (MR) is associated with poor survival and degenerative MR with excellent survival [1 4, 10 12]. We hypothesized that in some patients with degenerative MR requiring concomitant CABG, ischemic disease would dominate prognosis, resulting in survival as poor as in patients with ischemic MR. Thus, we (1) isolated the survival impact of etiology (degenerative vs ischemic) after combined mitral valve repair and Accepted for publication March 16, Presented at the Forty-first Annual Meeting of The Society of Thoracic Surgeons, Tampa, FL, Jan 24 26, Address reprint requests to Dr Gillinov, The Cleveland Clinic Foundation, 9500 Euclid Avenue / Desk F24, Cleveland, OH 44195; gillinom@ccf.org. CABG and (2) explored survival differences within etiology groups. Patients and Methods Definitions Mitral regurgitation was judged to be of ischemic etiology when valve leaflets and chordae appeared normal and MR was caused by consequences of a myocardial infarction; all patients with ischemic MR had at least one previous myocardial infarction, based on standard criteria extracted from review of clinical information, echocardiograms, and direct surgical inspection [1, 2, 10, 11, 13]. Those with normal-appearing papillary muscles, chordae, and leaflets were classified as having functional ischemic MR and were included in this study. Their leaflets failed to coapt, and echocardiograms frequently demonstrated restricted leaflet motion (Carpentier type IIIb leaflet motion) [11 13]. Mitral regurgitation was judged to be of degenerative etiology if microscopic pathologic examination of resected leaflet material revealed myxomatous degeneration, although we recognize that myxoid material might 2005 by The Society of Thoracic Surgeons /05/$30.00 Published by Elsevier Inc doi: /j.athoracsur

2 812 GILLINOV ET AL Ann Thorac Surg ISCHEMIC AND DEGENERATIVE MITRAL REGURGITATION 2005;80:811 9 Table 1. Patient Characteristics According to Etiology of Mitral Valve Dysfunction Characteristic Degenerative (n 710) Ischemic (n 400) n a No. (%) n a No. (%) p Demography Male 538 (76) 220 (55) Age (years) b Cardiac comorbidity NYHA functional class I 127 (18) 18 (4.5) II 383 (54) 151 (38) III 140 (20) 130 (32) IV 58 (8.2) 100 (25) Left ventricular dysfunction None 371 (57) 16 (4.4) Mild 145 (22) 42 (12) Moderate 93 (14) 150 (42) Severe 38 (5.9) 153 (42) Previous MI (32) (100) Mitral regurgitation (6.1) 55 (14) (28) 203 (51) (66) 142 (36) Atrial fibrillation/flutter c (18) (6.8) Number of diseased coronary systems d (5.2) 2 (0.5) (32) 37 (9.2) (28) 79 (20) (34) 282 (70) Noncardiac comorbidity History of smoking (49) (65) Peripheral vascular disease (29) (48) Hypertension (61) (71) Treated diabetes (13) (37) History of renal disease (3.1) (10) Creatinine (mg dl 1 ) e , 1.1, , 1.1, Hematocrit (%) b a Available data; b mean standard deviation; c on preoperative electrocardiogram; d 50% stenosis; 0 disease is generally left main disease or bypass of vessels with disease 50%; e Median and 15th and 85th percentiles. MI myocardial infarction; NYHA New York Heart Association. represent a final common pathway for possible diverse causes. When pathology reports were unavailable, valves were classified as degenerative if on echocardiography or direct surgical inspection they had typical findings of leaflet prolapse caused by chordal elongation or rupture and thick, floppy leaflets [12]. Finally, those rare patients with annular dilatation and thick billowing leaflets, but without frank prolapse, were judged to have degenerative disease. Study Group Using these definitions, patients undergoing mitral valve repair from 1985 to 2003 who could potentially be included in this study were identified by query of the Cardiovascular Information Registry. This registry contains detailed demographic, clinical, pathologic, operative, and outcome variables on all patients undergoing cardiac surgery at The Cleveland Clinic Foundation, abstracted from clinical records concurrent with patient care. The registry is approved for use in research by the Institutional Review Board. Because definitions of etiology of valve dysfunction have evolved, and because of the need to clearly separate degenerative disease from ischemic disease, search criteria were intentionally broad to cast a wide net that we believed would capture all instances of mitral valve repair in the settings of degenerative disease with CABG and functional ischemic MR. Medical records of the 2,000 patients so identified were reviewed to verify etiology, netting 1,110 patients; 710 had mitral valve repair for degenerative MR with concomitant CABG (two diseases), and 400 had mitral annuloplasty for functional ischemic MR and CABG (one disease). All had at least 2 (moderate) MR.

3 Ann Thorac Surg GILLINOV ET AL ;80:811 9 ISCHEMIC AND DEGENERATIVE MITRAL REGURGITATION Table 2. Patient Characteristics According to Etiology of Mitral Valve Dysfunction: Propensity-Matched Pairs Degenerative (n 123) Ischemic (n 123) Characteristic n a No. (%) n a No. (%) p Demography Male 73 (59) 70 (57) 0.7 Age (years) b Cardiac comorbidity NYHA functional class I 12 (9.8) 8 (6.5) II 48 (39.) 52 (42) III 38 (31) 33 (27 IV 25 (20) 30 (24) Left ventricular dysfunction None 7 (6) 12 (11) Mild 29 (27) 23 (20) Moderate 45 (41) 46 (41) Severe 28 (27) 31 (28) Previous MI (69) (100) Mitral regurgitation (10) 13 (10) 3 59 (48) 61 (50) 4 51 (41) 49 (40) Atrial fibrillation/flutter c (13) (11) 0.7 Number of diseased coronary systems d (0.82) 0 (0) 1 12 (9.8) 24 (20) 2 33 (27) 22 (18) 3 76 (62) 77 (63) Noncardiac comorbidity History of smoking (62) (58) 0.5 Peripheral vascular disease (46) (45) 0.9 Hypertension (71) (68) 0.6 Treated diabetes (31) (30) 0.8 History of renal disease (7.3) (9.9) 0.5 Creatinine (mg dl 1 ) e , 1.1, , 1.2, Hematocrit (%) b a Data available; b mean standard deviation; c on preoperative electrocardiogram; d 50% stenosis; 0 disease is generally left main disease or bypass of vessels with disease 50%; e Median, 15th and 85th percentiles. MI myocardial infarction; NYHA New York Heart Association. Patients undergoing concomitant tricuspid valve repair were included, but those undergoing aortic valve or Maze procedures or reoperations were not. Mean patient age of the entire group was 68 9 years. Patients with ischemic MR had more extensive coronary artery disease, more severe left ventricular dysfunction, higher New York Heart Association (NYHA) functional class, and more severe noncardiac comorbidity (Table 1). Patients with degenerative mitral valve disease had more severe MR. Surgical Details Mitral valve repair was performed using standard techniques [12 14]. For patients with functional ischemic MR, repair consisted of annuloplasty alone. Annuloplasty techniques included posterior bovine pericardial band, classic Carpentier ring (Edwards Lifesciences, LLC, Irvine, CA) and, in recent years, an undersized Cosgrove-Edwards Annuloplasty System (Edwards Lifesciences, LLC, Irvine, CA). For patients with degenerative mitral valve disease, repair techniques included annuloplasty, posterior leaflet resection, chordal transfer, chordal shortening, and commissure closure. Intraoperative echocardiography was used routinely, and patients left the operating room with MR graded as mild (1 ) or less. Follow-Up Patients were followed systematically at two-year intervals by means of a mailed questionnaire, telephone interview, or examination at the Cleveland Clinic. Data regarding vital status were augmented using the Social

4 814 GILLINOV ET AL Ann Thorac Surg ISCHEMIC AND DEGENERATIVE MITRAL REGURGITATION 2005;80:811 9 Table 3. Patient Characteristics: Ischemic Group, Matched Versus Unmatched Characteristic Unmatched (n 277) Matched (n 123) n a No. (%) n a No. (%) p Demography Male 150 (54) 70 (57) 0.7 Age (years) b Cardiac comorbidity NYHA functional class I 10 (3.6) 8 (6.5) II 99 (36) 52 (42) III 97 (35) 33 (27) IV 70 (25) 30 (24) Left ventricular dysfunction None 4 (1.6) 12 (11) Mild 19 (7.7) 23 (20) Moderate 104 (42) 46 (41) Severe 22 (49) 31 (28) Previous MI (100) (100) c Mitral regurgitation (15) 13 (10) (51) 61 (50) 4 93 (34) 49 (40) Atrial fibrillation/flutter d (4.7) (11) 0.01 Number of diseased coronary systems e (0.72) 0 (0) 1 13 (4.7) 24 (20) 2 57 (20) 22 (18) (74) 77 (63) Noncardiac comorbidity History of smoking (68) (58) 0.05 Peripheral vascular disease (49) (45) 0.4 Hypertension (72) (68) 0.4 Treated diabetes (40) (30) 0.05 History of renal disease (10) (9.9) 0.9 Creatinine (mg dl 1 ) f , 1.1, , 1.2, Hematocrit (%) b a Data available; b Mean standard deviation; c by definition of ischemic group; d on preoperative electrocardiogram; e 50% stenosis, 0 disease is generally left main disease or bypass of vessels with disease 50%; f median, 15th and 85th percentiles. MI myocardial infarction; NYHA New York Heart Association. Security Death Index. Median follow-up durations for degenerative and ischemic groups were 4.1 years and 3.1 years, respectively. This yielded 3,766 patient-years of follow-up for the degenerative group and 1,480 years for the ischemic group. Among propensity- matched patients with ischemic or degenerative MR (123 in each group), median follow-up was 2.6 years in the degenerative group and 3.2 years in the ischemic group, yielding totals of 484 and 443 patient-years of follow-up, respectively. Survival estimates were considered reliable to at least 9 years. Data Analysis UNADJUSTED SURVIVAL. Unadjusted survival estimates between and within etiology groups were obtained using the nonparametric Kaplan-Meier method and a parametric method that resolved phases of instantaneous risk (hazard) [15]. (For additional details, see clevelandclinic.org/heartcenter/hazard.) PROPENSITY MATCHING AND COMPARATIVES SURVIVAL. Propensity matching was used to assess the effect of etiology of mitral valve dysfunction on survival [16, 17]. Logistic regression analysis was used to identify demographic, clinical, comorbid, and coronary disease factors associated with ischemic versus degenerative etiology (Appendix). To these were added nonsignificant variables for each of the above classes of factors (saturated model), 61 variables in all. Patients with ischemic and degenerative etiologies were then matched on propensity score for variables used in the analysis, yielding 123 propensitymatched pairs (Table 2). Survival was compared between these matched ischemic MR and degenerative groups,

5 Ann Thorac Surg GILLINOV ET AL ;80:811 9 ISCHEMIC AND DEGENERATIVE MITRAL REGURGITATION Table 4. Patient Characteristics: Degenerative Group, Matched Versus Unmatched Unmatched (n 587) Matched (n 123) Characteristic n a No. (%) n a No. (%) p Demography Male 465 (79) 73 (59) Age (years) b Cardiac comorbidity NYHA functional class I 115 (20) 12 (9.8) II 335 (57) 48 (39) III 102 (17) 38 (31) IV 33 (5.6) 25 (20) Left ventricular dysfunction None 364 (68) 7 (6.4) Mild 116 (22) 29 (27) Moderate 48 (8.9) 45 (41) Severe 10 (1.9) 28 (26) Previous MI (24) (69) Mitral regurgitation (5.1) 13 (11) (24) 59 (48) (71) 51 (41) Atrial fibrillation/flutter c (20) (13) 0.08 Number of diseased coronary systems d (6.2) 1 (0.82) (37) 12 (9.8) (28) 33 (27) (29) 76 (62) Noncardiac comorbidity History of smoking (46) (62) Peripheral vascular disease (25) (46) Hypertension (59) (71) 0.02 Treated diabetes (8.8) (31) History of renal disease (2.2) (7.3) Creatinine (mg dl 1 ) e , 1.1, , 1.1, Hematocrit (%) b a Data available; b mean standard deviation; 2te-para c on preoperative electrocardiogram; d 50% stenosis, 0 disease is generally left main disease or bypass of vessels with disease 50%; e Median, 15th and 85th percentiles. MI myocardial infarction; NYHA New York Heart Association. and also between matched versus unmatched ischemic MR patients, and matched versus unmatched degenerative disease groups (Tables 3 and 4). Presentation Continuous variables are presented as mean standard deviation or equivalently as 15th, 50th (median), and 85th percentiles when the distribution of values was skewed. Mortality and survival estimates are accompanied by asymmetric 68% confidence limits (CL), comparable to 1 standard error. Results Unadjusted Survival Hospital mortality was 3.1% (22/710, CL 2.4% and 3.9%) in patients with degenerative mitral disease and 5.0% (20/400, CL 3.9% and 6.4%) in those with ischemic MR. Unadjusted 5-year survival was 82% in the degenerative group versus 64% in the ischemic MR group (p [log-rank] ; Fig 1). Propensity Matching and Comparative Survival Among the 123 propensity-matched pairs, 5-year survival was similar; 65% in the degenerative group versus 66% in the ischemic MR group (p 0.9; Fig 2). Survival Within Etiology Groups Unmatched ischemic MR patients, despite somewhat worse disease, experienced survival similar to matched ischemic MR cases (p [log-rank] 0.7; Table 3, Fig 3). Compared with matched patients, unmatched degenerative disease patients had less extensive ischemic disease (p ),

6 816 GILLINOV ET AL Ann Thorac Surg ISCHEMIC AND DEGENERATIVE MITRAL REGURGITATION 2005;80:811 9 ventricular dysfunction. Thus, ischemic and degenerative MR patients with equivalent characteristics have equivalently poor survival. Ischemic Mitral Regurgitation The poor long-term survival [1, 2, 10, 18, 19] seen in patients with coronary artery disease and ischemic MR is associated with a graded effect: the greater the degree of MR, the lower the survival [20, 21]. Recognizing the deleterious impact of ischemic MR, surgeons have studied this entity intensely over the last decade [1 9, 11, 13, 18, 19]. With improved understanding of its pathogenesis and advances in surgical and perioperative management, hospital mortality has been reduced from 10% or greater to 2% 7% [1, 11, 18]. The most common type of ischemic MR is functional ischemic MR. In this condition, myocardial infarction and subsequent remodeling cause changes in ventricular and Fig 1. Unadjusted survival after mitral valve repair and coronary artery bypass grafting in patients with ischemic mitral regurgitation or degenerative mitral valve and coronary artery diseases. (A) Survival. Each symbol represents a death positioned on the vertical axis by the nonparametric Kaplan-Meier estimator, vertical lines are asymmetric 68% confidence limits (equivalent to 1 standard error), and numbers in parentheses are patients still alive and traced. Solid lines are parametric estimates enclosed within 68% confidence limits. Ischemic mitral regurgitation is represented by squares and degenerative mitral regurgitation by circles. (B) Instantaneous risk of death (hazard function). Dashed lines enclose 68% confidence limits. better left ventricular function (p ), and more severe MR (p ). These patients had substantially better survival than matched degenerative disease patients; 85% versus 65% at 5 years (p [log-rank] ; Table 4, Fig 4). Comment Key Findings Patients with functional ischemic MR have one disease process, and long-term survival is generally poor. In contrast, patients with degenerative MR undergoing mitral valve repair and CABG have two distinct diseases, and the severity of the coronary artery disease and extent of secondary left ventricular dysfunction have a major, but variable, impact on survival. The large unadjusted survival discrepancy between patients with ischemic and degenerative MR is attributable to differences in patient profile, particularly extents of ischemic disease and left Fig 2. Survival after mitral valve repair and coronary artery bypass grafting in propensity-matched patients with ischemic and degenerative mitral regurgitation (123 patients in each group). (A) Survival. Each symbol represents a death positioned on the vertical axis by the nonparametric Kaplan-Meier estimator, vertical lines are asymmetric 68% confidence limits (equivalent to 1 standard error), and numbers in parentheses are patients still alive and traced. Solid lines are parametric estimates enclosed within 68% confidence limits. Ischemic mitral regurgitation is represented by squares and degenerative mitral regurgitation by circles. (B) Instantaneous risk of death (hazard function). Dashed lines enclose 68% confidence limits.

7 Ann Thorac Surg GILLINOV ET AL 2005;80:811 9 ISCHEMIC AND DEGENERATIVE MITRAL REGURGITATION 817 annular geometry that prevent leaflet coaptation, resulting in a jet of MR that is usually directed centrally or posteriorly [1, 10, 22]. It is generally accepted that mitral valve repair is superior to replacement for this entity, with repair usually consisting of annuloplasty alone [1, 2, 10, 13]. Although mitral valve repair is technically successful in the majority of these patients, long-term survival is limited [1 4]. The most important determinants of mortality include advanced age, reduced left ventricular function, and higher NYHA functional class [1 10]. Patients with functional ischemic MR represent a relatively homogeneous group with respect to these factors, which explains the equivalent survival curves of matched and unmatched ischemic MR patients (see Fig 3). These similarities in patient presentation and outcome in ischemic MR distinguish this entity from other etiologies of MR. Fig 3. Survival after mitral valve repair and coronary artery bypass grafting in matched and unmatched patients with ischemic mitral regurgitation. Each symbol represents a death positioned on the vertical axis by the nonparametric Kaplan-Meier estimator, vertical lines are asymmetric 68% confidence limits (equivalent to 1 standard error), and numbers in parentheses are patients still alive and traced. Solid lines are parametric estimates enclosed within 68% confidence limits. Ischemic mitral regurgitation is represented by squares and degenerative mitral regurgitation by circles. Fig 4. Survival after mitral valve repair and coronary artery bypass grafting in matched and unmatched patients with degenerative mitral regurgitation. Each symbol represents a death positioned on the vertical axis by the nonparametric Kaplan-Meier estimator, vertical lines are asymmetric 68% confidence limits (equivalent to 1 standard error), and numbers in parentheses are patients still alive and traced. Solid lines are parametric estimates enclosed within 68% confidence limits. Ischemic mitral regurgitation is represented by squares and degenerative mitral regurgitation by circles. Degenerative Mitral Valve Disease In contrast to patients with ischemic MR, those with coexisting degenerative mitral valve disease and coronary artery disease have two diseases; our findings suggest that in such patients, variable severity of coronary artery disease and, secondly, left ventricular dysfunction, have a major impact on long-term survival. There is a wide spectrum of cardiac comorbidity in patients presenting for surgical correction of degenerative mitral valve disease. Those with isolated degenerative mitral valve disease present for surgical repair at a mean age of years; most have normal left ventricular function, and 84% are in NYHA functional class I or II [12]. Hospital mortality is less than 1%, and 5-year and 10-year survivals are greater than 90% and 80%, respectively [12, 23, 24]. In contrast, patients with degenerative mitral valve disease and coexisting coronary artery disease have more variable clinical presentations and outcomes, which explain the different survival curves between matched and unmatched patients (see Fig 4). As in ischemic MR, long-term survival is modulated by patient age, extent of left ventricular dysfunction, and NYHA functional class [14]. Impact of Etiology on Survival There are few data comparing outcomes in these two groups of patients with both mitral valve dysfunction and coronary artery disease. In a review from the Mayo Clinic, Dahlberg and colleagues [3] studied 302 patients undergoing mitral valve repair or replacement and CABG; approximately half had ischemic MR and half had degenerative mitral valve disease. As in our study, patients with ischemic disease tended to have more left ventricular dysfunction, more extensive coronary artery disease, and higher NYHA functional class. Unadjusted 10-year survivals were 33% in the ischemic group and 52% in the degenerative group. Advanced age, reduced left ventricular function, and extent of coronary artery disease influenced survival; etiology of mitral valve dysfunction was not an independent risk factor. Like us, they concluded that survival after mitral valve surgery and CABG is determined primarily by extent of coronary artery disease and left ventricular dysfunction rather than by etiology of the mitral valve disease. The sequence of myocardial infarction and remodeling associated with ischemic MR has a greater impact on survival than does the degree of MR. Others [4, 5, 25] have also noted important preoperative differences between patients with ischemic MR and those with combined degenerative MR and coronary artery disease. Glower and colleagues [25] found that etiology of mitral dysfunction was not an independent predictor of survival; rather, advanced age and comorbidities determined long-term survival. Similarly, Bouchard and colleagues [5] noted that in patients undergo-

8 818 GILLINOV ET AL Ann Thorac Surg ISCHEMIC AND DEGENERATIVE MITRAL REGURGITATION 2005;80:811 9 ing mitral valve replacement and CABG, comorbidities, not etiology of valve dysfunction, affected late survival. Limitations There were important preoperative differences between patients with ischemic and degenerative MR (Table 1). By using propensity matching, we attempted to adjust for survival differences in patient characteristics (Table 2); however, it is possible that other, unrecorded variables also influenced outcomes. In this report, we have not addressed the issue of durability of mitral valve repair. This topic has been studied extensively and reported in previous publications [1 3, 12, 13, 23, 24, 26, 27]. We cannot address in a formal fashion the relationship between progression of postoperative MR and survival. Others have suggested that MR recurrence jeopardizes survival [3]; however, to our knowledge, statistical methodology to properly test this hypothesis does not exist because both recurrent MR and death are outcomes of surgery. Examination of their interdependence and their relationship to surgical technique requires developing novel statistical methodology. Clinical Inferences Patients with functional ischemic MR represent a relatively homogeneous group with limited postoperative survival. In contrast, patients with degenerative mitral valve disease and coronary artery disease have two distinct diseases; in these patients, extent of ischemic disease and resulting left ventricular dysfunction are critical determinants of long-term survival. Thus, the sequence of myocardial infarction and subsequent remodeling has a greater impact upon survival than does the degree of MR. This suggests that strategies to prevent or reverse the remodeling sequence associated with the development of ischemic MR may be important components of future efforts to manage this condition. References 1. Gillinov AM, Wierup PN, Blackstone EH, et al. Is repair preferable to replacement for ischemic mitral regurgitation? J Thorac Cardiovasc Surg 2001;122: Grossi EA, Goldberg JD, LaPietra A, et al. Ischemic mitral valve reconstruction and replacement: comparison of longterm survival and complications. J Thorac Cardiovasc Surg 2001;122: Dahlberg PS, Orszulak TA, Mullany CJ, Daly RC, Enriquez- Sarano M, Schaff HV. Late outcome of mitral valve surgery for patients with coronary artery disease. Ann Thorac Surg 2003;76: ; discussion Seipelt RG, Schoendube FA, Vazquez-Jimenez JF, Doerge H, Voss M, Messmer BJ. Combined mitral valve and coronary artery surgery: ischemic versus non-ischemic mitral valve disease. Eur J Cardiothorac Surg 2001;20: Bouchard D, Pellerin M, Carrier M, et al. Results following valve replacement for ischemic mitral regurgitation. Can J Cardiol 2001;17: Szecsi J, Herijgers P, Sergeant P, Daenen W, Scheys I, Flameng W. Mitral valve surgery combined with coronary bypass grafting: multivariate analysis of factors predicting early and late results. J Heart Valve Dis 1994;3: Kirklin JK, Naftel DC, Blackstone EH, Kirklin JW, Brown RC. Risk factors for mortality after primary combined valvular and coronary artery surgery. Circulation 1989;79:I Akins CW, Hilgenberg AD, Buckley MJ, et al. Mitral valve reconstruction versus replacement for degenerative or ischemic mitral regurgitation. Ann Thorac Surg 1994;58:668 75; discussion Thourani VH, Weintraub WS, Guyton RA, et al. Outcomes and long-term survival for patients undergoing mitral valve repair versus replacement: effect of age and concomitant coronary artery bypass grafting. Circulation 2003;108: Miller DC. Ischemic mitral regurgitation redux to repair or to replace? J Thorac Cardiovasc Surg 2001;122: Adams DH, Filsoufi F, Aklog L. Surgical treatment of the ischemic mitral valve. J Heart Valve Dis 2002;11(suppl 1):S Gillinov AM, Cosgrove DM, Blackstone EH, et al. Durability of mitral valve repair for degenerative disease. J Thorac Cardiovasc Surg 1998;116: McGee EC, Gillinov AM, Blackstone EH, et al. Recurrent mitral regurgitation after annuloplasty for functional ischemic mitral regurgitation. J Thorac Cardiovasc Surg 2004; 128: Gillinov AM, Faber C, Houghtaling PL, et al. Repair versus replacement for degenerative mitral valve disease with coexisting ischemic heart disease. J Thorac Cardiovasc Surg 2003;125: Blackstone EH, Naftel DC, Turner ME Jr. The decomposition of time-varying hazard into phases, each incorporating a separate stream of concomitant information. J Am Stat Assoc 1986;81: Rosenbaum PR, Rubin DB. Reducing bias in observational studies using subclassification on the propensity score. J Am Stat Assoc 1984;79: Blackstone EH. Comparing apples and oranges. J Thorac Cardiovasc Surg 2002;123: Cohn LH. Mitral valve repair for ischemic mitral regurgitation. Adv Cardiol 2002;39: Cohn LH, Rizzo RJ, Adams DH, et al. The effect of pathophysiology on the surgical treatment of ischemic mitral regurgitation: operative and late risks of repair versus replacement. Eur J Cardiothorac Surg 1995;9: Grigioni F, Enriquez-Sarano M, Zehr KJ, Bailey KR, Tajik AJ. Ischemic mitral regurgitation: long-term outcome and prognostic implications with quantitative Doppler assessment. Circulation 2001;103: Ellis SG, Whitlow PL, Raymond RE, Schneider JP. Impact of mitral regurgitation on long-term survival after percutaneous coronary intervention. Am J Cardiol 2002;89: Kwan J, Shiota T, Agler DA, et al. Geometric differences of the mitral apparatus between ischemic and dilated cardiomyopathy with significant mitral regurgitation: real-time three-dimensional echocardiography study. Circulation 2003;107: Braunberger E, Deloche A, Berrebi A, et al. Very long-term results (more than 20 years) of valve repair with Carpentier s techniques in nonrheumatic mitral valve insufficiency. Circulation 2001;104: Mohty D, Orszulak TA, Schaff HV, Avierinos JF, Tajik JA, Enriquez-Sarano M. Very long-term survival and durability of mitral valve repair for mitral valve prolapse. Circulation 2001;104: Glower DD, Tuttle RH, Shaw LK, Orozco RE, Rankin JS. Patient survival characteristics after routine mitral valve repair for ischemic mitral regurgitation. J Thorac Cardiovasc Surg 2005;129: Tahta SA, Oury JH, Maxwell JM, Hiro SP, Duran CM. Outcome after mitral valve repair for functional ischemic mitral regurgitation. J Heart Valve Dis 2002;11:11 8; discussion Hung J, Papakostas L, Tahta SA, et al. Mechanism of recurrent ischemic mitral regurgitation after annuloplasty: continued LV remodeling as a moving target. Circulation 2004; 110:II85 90.

9 Ann Thorac Surg GILLINOV ET AL 2005;80:811 9 ISCHEMIC AND DEGENERATIVE MITRAL REGURGITATION Appendix Variables Used in Analysis 819 Preoperative Demographic: Age (years), gender, weight (kg), height (cm), body surface area (m 2 ), body mass index (kg m -2 ) Symptoms: New York Heart Association functional class (I-IV) Ventricular dysfunction: Degree of left ventricular dysfunction (1 none, 2 mild, 3 moderate, 4 severe), graded wall motion abnormality (0 6): inferior/ diaphragmatic, anterior, apical, lateral, basal/infraatrial, septal/atrioseptal (0 trivial/none, 1 mild hypokinesis/dyskinesis, 2 mild to moderate, 3 moderate, 4 moderate to severe, 5 severe to akinesis, 6 aneurysm present) Mitral valve Mitral valve regurgitation (2 moderate, 3 moderate to severe, 4 severe) pathophysiology: Cardiac comorbidity: Atrial fibrillation, ventricular arrhythmia Noncardiac comorbidity: Treated diabetes, history of peripheral vascular disease, history of smoking, carotid disease, peripheral vascular disease, creatinine (mg dl 1 ), blood urea nitrogen (mg dl 1 ), hematocrit (%) Coronary stenosis: Left main trunk, left anterior descending (proximal, mid, distal, first and second diagonal, major septal perforator, maximum in system), left circumflex (ramus, mid, distal, high lateral, lateral, posterior lateral, maximum in system), right coronary artery (proximal, mid, distal, marginal, posterior descending, posterior lateral, maximum in system) Experience: Date of operation (years since 1/1/1985) DISCUSSION DR STEVEN F. BOLLING (Ann Arbor, MI): Clearly it has long been known that those patients with ischemic mitral regurgitation have a different disease than those who have coronary artery disease and concomitant degenerative mitral disease: one has a disease of the valve and one has a disease of the ventricle. Did you do follow-up on these patients looking at recurrent or residual mitral regurgitation, both short term and long term? As we know, those patients with ischemic mitral disease do extremely poorly when they have recurrent mitral regurgitation. And I know that this study spans a long period of time. Were these treated the same way by you? Were the rings undersized? Did you use a complete ring? We know that in your very nice follow-up studies the pericardial band patients at your institution fare relatively poorly. This is a long and beautiful study. Perhaps you could comment on those questions. DR GILLINOV: We have performed follow-up on these patients, although it is not presented as a component of this paper. As you know, in the patients with degenerative mitral valve disease, the long-term freedom from reoperation and recurrent mitral regurgitation is excellent, more than 90% at 10 years, and in recent studies, similar figures at 20 years. In patients with functional ischemic mitral regurgitation, the results are not as good. Up to a third of the patients with ischemic MR (mitral regurgitation) treated in our series with various types of annuloplasty, developed mitral regurgitation that was 2 or greater over the first postoperative year. You raise the question, can we relate recurrent mitral regurgitation to survival, and the answer is that we cannot. Others have attempted to show a relationship between these two outcomes, but in truth the statistical methodology to demonstrate that one outcome has an impact on another outcome namely, that mitral regurgitation reduces survival that methodology doesn t exist. On the other hand, it does make clinical sense to suppose that the more recurrent mitral regurgitation a patient has, the worse the clinical outcome. However, testing this hypothesis is challenging. DR LISHAN AKLOG (New York, NY): I enjoyed your presentation. It was, as usual, a very elegant statistical analysis. I wouldn t quibble with the data or the analysis of the data but I am having a hard time understanding the message specifically. That is because the number of patients that you were able to match each group was relatively small. By my math, about a third of the ischemic patients and about a fifth of the degenerative patients were ultimately matched. That left a large number of unmatched patients in both groups. So what is the real message for us out there in treating these two diseases given that your analysis was based on fairly select subgroups? You said that the unmatched and matched patients in the ischemic group were fairly similar. If they were fairly similar, why was such a small fraction ultimately able to be entered in the propensity matching analysis? And who are these patients in the degenerative group that do poorly? What are their characteristics? DR GILLINOV: The message is that patients who have both degenerative mitral valve disease and coronary artery disease span a wide spectrum of clinical presentations; just because somebody has coronary disease and mitral valve dysfunction does not mean that they are going to have a poor outcome. Amongst patients with ischemic mitral regurgitation, the outcomes are generally pretty poor. But if you go to operate on a patient who needs bypass grafting and a mitral valve procedure, that patient, if he has degenerative disease, may do very well. So I think that is one of the primary messages. The reason we could not match every degenerative patient is that we did not have enough ischemic patients who had favorable clinical features, meaning good left ventricular function, less coronary artery disease, and fewer comorbidities.