Early and Late Results of Combined Mitral-Aortic Valve Surgery

Results There were 57 men and 56 women, median age 59 (18–84) years. The 30-day mortality was 9% ( n = 10). This cohort contained a number of high risk patients, 38% were classified as New York Heart Association class IV, 33.5% had at least moderate ventricular impairment, 20% were redo procedures and 17% urgent procedures. Survival estimates at 5 and 10 years were 85% (0.76–0.90) and 65% (0.49–0.77), respectively. Multivariate pre-operative predictors of death included renal dysfunction (creatinine >200 μmol/L) and hypertension. Rheumatic aetiology was associated with improved survival. Conclusion This study shows acceptable short and long-term survival in patients undergoing combined aortic-mitral valve surgical procedures at a single centre. Renal impairment and hypertension were associated with a poorer long-term prognosis and rheumatic aetiology was associated with improved survival. Age, LVEF and NYHA class were not associated with a worse outcome. This may affect future decision making in light of an aging population. Keywords Combined valvular surgery Prosthetic valve Bioprosthetic valves Renal failure Rheumatic heart disease Introduction Advances in cardiac surgical techniques over the past two decades have seen a wider range of treatment options for valvular heart disease, which include valvular repair, valvuloplasty and/or valve replacement. Whilst double valve replacement (DVR) comprises approximately 2% of open-heart surgery cases, it has been regarded as a higher risk operation than other procedures such as single valve replacement. 1 Factors associated with adverse outcome in the past have included older age, 2,3 New York Heart Association (NYHA) class III or IV, 2–4 low left ventricular ejection fraction, 5 and concomitant coronary artery bypass surgery (CABG). 6 In the past studies, double valve replacement has been associated with peri-operative morbidity and mortality of 3–15%. 6–9 The purpose of this study was to determine the 30-day mortality rates, and all cause mortality during the follow up period (1989 to May 2004) at a medium sized single centre. Furthermore, we analysed the data set in an attempt to determine factors which may be associated with a poor outcome. Patients and Methods Patient Population Between January 1993 and January 2003, 113 patients, 57 men and 56 women, median age 59 years (range 18–84 years) were identified by the hospital's Intensive Care and Cardiac surgical and perfusion unit databases as undergoing double valve replacement. Patients having concomitant bypass grafting were included. The campus includes both public and private open heart surgical units (OHSU). The same group of eight surgeons, anaesthetists and cardiac perfusionists worked in both public and private units, with open heart surgery commencing at the private campus in 1994. The clinical characteristics of the patients are shown ( Table 1 ). This cohort included a number of high risk patients. Patients were classified as high risk if they had one or more traditional factors associated with poor outcomes. These included New York Heart Association class (NYHA) IV, moderate or severe impairment of left ventricular function, redo and urgent procedures. Thus 38% were classified as NYHA IV, 33.5% had at least moderate left ventricular systolic impairment, 20% were redo procedures and 17% urgent procedures. 2–4 In total 46 of the 113 patients (41%) were deemed high risk. The predominant cause of valvular disease was rheumatic (64%). Other aetiologies included degenerative (27%) and infective (9%). To determine the left ventricular ejection fraction (LVEF), the pre-operative notes were assessed. Pre-operative echocardiography or cine-ventriculography assessment of LVEF was determined in 103 patients. The cohort was then classified by the following schema: I. I Normal left ventricular function (EF > 50%), II. II Mildly impaired systolic function (EF 40–50%), III. III Moderately impaired systolic function (EF 26–39%), IV. IV Severely impaired systolic function (EF ≤ 25%). Examination of the history and surgical notes was made in order to assess the type of valvular patho-physiology (stenotic or regurgitant, see Table 1 ). Pre-operative creatinine was determined and renal impairment was specified as a pre-operative creatinine of greater than 200 μmol/L (normal range 70–110 μmol/L), 14% of patients were classified as having renal impairment. Post-operative renal failure was defined as a creatinine rise to greater than 200 μmol/L (normal range 70–110 μmol/L). Atrial fibrillation was present pre-operatively in 43 (41%) cases. Surgical Procedures All operations were performed through a median sternotomy. Cardiopulmonary bypass was instituted via aortic and bicaval cannulation at 32 °C. Up until 2000, myocardial protection was afforded according to surgeon's preferences, using either intermittent cold crystalloid cardioplegia or intermittent antegrade and retrograde room temperature blood cardioplegia. The latter was used exclusively after 2000. The mitral valve was exposed either via a lateral left atriotomy posterior to the inter-atrial groove or a superior septal (extended trans-septal) approach, according to the surgeon's preference. The mitral valve was seated using interrupted pledgeted sutures with preservation of at least the posterior leaflet sub-valvar structures. Following this, the atrial incisions were closed and the aortic valve implanted through a curvilinear aortotomy using interrupted pledgeted sutures. When there was associated coronary artery bypass grafting, the distal anastomoses were performed before valve implantation. Mean cross clamp time and cardiopulmonary bypass time were 129 ± 30 min and 158 ± 36 min, respectively. Intra-operative transoesophageal echocardiography was used in all cases to confirm pre-operative pathology and document valvular function post-implantation. Ninety-four patients (84%) received two mechanical bileaflet valves, and the remainder received tissue valves in either mitral or aortic with only one patient receiving tissue valves in both locations. Mechanical bileaflet valves used were either St Jude or Carbomedics. One patient underwent triple valve replacement (aortic, mitral and tricuspid) and five patients underwent concomitant coronary artery bypass grafting. Follow-up Initially 30-day follow up was documented over a three-month period by obtaining and scrutinising patient records. Long-term follow up was obtained by telephone contact of the patient, family members or their General Practitioner. Survival was assessed to determine all cause mortality. Hospital cardiac, pathology and radiology databases were assessed if patient contact could not be achieved. Where no phone contact was made, the last date of International normalised ratio (INR) was substituted as last known survival time. Overall, there was 100% 30-day and 94% long-term follow up. Median follow up time was 5.4 years with a maximum of 12.9 years, giving a combined follow up of 618 patient years. Post-operative Anticoagulation All patients were commenced on 5000 units calcium heparin bd subcutaneously 24 hours after the operation, and coumarin treatment was commenced 24 hours after the operation to a target International normalised ratio of 3–3.5. Definitions Complications were defined in accordance with the published guidelines for reporting valve-related morbidity and mortality after cardiac valvular operation. 10 Statistical Analysis Data are expressed as mean (±S.D.) or median (range) for continuous variables and as number or percentage for categorical variables. Thirty-day mortality was modelled using logistic regression (Hosmer and Lemeshow). Univariate predictors were divided into pre-operative (age, gender, left ventricular function, New York Heart Association classification, hypertension, diabetes, creatinine (μmol/L), atrial fibrillation, COPD, aortic stenosis, aortic regurgitation, mixed aortic valve disease, mitral stenosis, mitral regurgitation, mixed mitral valve disease, rheumatic origin, endocarditic origin and re-operation), intra-operative (CABG, inotrope use, cardio-pulmonary bypass time, cross-clamp time) and post-operative (inotrope use, IABP, low cardiac output, arrest, tamponade, reoperation for bleeding/tamponade, readmission to ICU, pneumonia, renal failure, stroke, permanent pacemaker, infection and a composite variable of all complications); those predictors achieving p < 0.20 were included in subsequent multiple logistic regression using stepwise (forward and backward) selection. Model suitability was assessed for discrimination (sensitivity and specificity, area under ROC curve) and calibration (Hosmer-Lemeshow chi-squared statistic). Long-term survival was estimated using the Kaplan-Meier technique. Cox regression analysis was used for multivariate analysis of survival using the same predictors as described above and with stepwise (forward and backward) selection. Two models were analysed. Model 1 assessed long-term pre-operative factors associated with poor long-term outcomes. Model 2 assessed pre-operative, intra-operative and post-operative factors associated with poor outcomes. Model suitability was assessed with published techniques. 11 Analyses were performed with Stata Version 8 (Stata Corp. College Station, TX). Confidence intervals are presented at the 95% level. Statistical significance was set at p < 0.05. Results The median ICU stay was two days (range 1–28 days), and median hospital stay was nine days (range 6–85). 30-day Survival Ten patients died within the first 30 days of operation, yielding a 9% 30-day mortality rate. Univariate pre-operative, intra-operative and post-operative indicators of early post-operative mortality are demonstrated in Table 2 . Neither age, sex, LVEF, nor NYHA class predicted early mortality. Multivariate modelling was used to determine the best predictors of death from pre, intra and post-operative factors. The pre-operative model demonstrated pre-operative renal failure, the intra-operative model demonstrated inotropes commenced in the OR and the post-operative model demonstrated cardiac arrest and post-operative renal failure to predict a poor outcome. Overall, pre-operative renal failure and cardiac arrest were the best predictors of 30-day post-operative mortality ( Table 3 ). Morbidity was significant in this group. Thirty percent of patients required inotropic support within the intensive care unit, and 32% required the transfusion of blood products. Low cardiac output necessitated intra-aortic balloon counter-pulsation in four patients. Ten patients suffered a cardiac arrest, and eight patients required re-operation for cardiac tamponade. Atrial fibrillation was noted in 41 patients in the post-operative period, and 11 patients required permanent pacing. Twenty-one patients had documented pneumonia, nine patients suffered a cerebrovascular accident and four patients developed lactic acidosis of which three required laparotomy. Thirty-eight patients (33%) suffered at least one significant complication or death as a result of double valve replacement. Of the 10 patients who died within 30 days of their operation, seven died of pump failure, four of whom had redo procedures and three were morbidly obese. Of the remaining three patients, two died of unexplained lactic acidosis. No cause was found at laparotomy. The final patient died of pump failure and unexplained lactic acidosis. Of note, in the 10 patients who died, renal failure requiring renal replacement therapy was present in three cases, and pre-operative renal failure (creatinine >200 μmol/L) was present in a further two cases. The latter two of these patients required haemofiltration in the post-operative period. In total, 18 patients developed post-operative renal failure. Late Deaths The five-year survival rate was 85% (0.76–0.90) and 10-year survival rate 65% (0.49–0.77). 75% of patients were alive at nine years ( Fig. 1 ). One patient died suddenly at day 35 after an uneventful post-operative course. Univariate predictors of death are presented (see Table 4 ). Two multivariate analysis models were constructed to assess the late outcome. The first model only included pre-operative variables. Pre-operative predictors of death included renal failure and hypertension, with rheumatic aetiology conferring a survival advantage. The second model used pre-operative/intra-operative and post-operative variables and showed that rheumatic aetiology again associated with improved survival, but this did not become significant until after 4.5 years follow up ( Tables 5 and 6 ). Discussion The 30-day mortality for double valve replacement in this single centre study was 9%. Pump failure/low cardiac output syndrome was the predominant cause in his group. Early morbidity was significant, with around one-third of patients suffering a significant complication. The cohort studied contained 46 patients, who, based upon prior studies 5,9,12,13 were at higher risk of significant morbidity and mortality as evidenced by the LV grade, NYHA class, age, percentage of redo procedures as well as urgent procedures. Neither LVEF, age nor NYHA class were predictors of long-term mortality in this study. Previous studies have demonstrated LVEF as a predictor of long-term mortality, 5–13 but Remadi et al. 9 also found it was not significant. We believe that inclusion bias might explain this, as patients with the greatest impairment of LV function are unlikely to undergo surgery due to the prohibitive mortality. Secondly, improvements in myocardial protection with the use of retrograde room temperature cardioplegia 14,15 have been shown to reduce early morbidity. This coupled with advances in medical care of patients with LV dysfunction, such as the routine use of ACE inhibitors 16 and more recently beta blockade 17 would be expected to alter the natural history of the disease, improving survival. In this study, three of the patients who died early had renal impairment, one with a failed renal transplant. A further two patients had a baseline creatinine greater than 200 μmol/L. There is a paucity of data in the literature regarding patients with pre-existing renal impairment and DVR surgery, most likely due to a reluctance to operate on these patients. A recent study of survival for dialysis patients after CABGS showed that five-year survival was 55.8% for patients with renal failure, and if diabetes and peripheral vascular disease were present, 42.2%. 18 Although that study did not look at patients undergoing cardiac valve surgery alone, it suggests that patients with chronic renal impairment prior to cardiac surgery have a higher post-operative mortality. Gillinov et al. 13 also demonstrated that pre-existing renal disease was associated with an excess early mortality. Recent studies have established a graded relationship between reduced glomerular filtration rate and cardiovascular mortality. 19 Whilst the possible mechanisms are beyond the scope of this article, it appears that excess cardiovascular mortality from underlying atherosclerosis is the predominant cause of death, independent of age, sex and other co-morbidities. Multivariate analysis demonstrated that the presence of pre-operative hypertension was associated with an increased risk of death in the long term. Hypertension is clearly established as a risk factor for coronary artery disease and cerebrovascular disease, 20 and our data, with a collective follow-up of greater than 600 patient years supports this. Unfortunately, we did not have data regarding the use, and adequacy, of blood pressure control during patient follow up to assess whether patients with DVR respond differently to antihypertensive therapy. This is a potentially modifiable risk factor however and we do not believe in the current era of antihypertensive strategies that this should effect decisions in regard to DVR. One patient in this study died suddenly. This has been recognised post DVR, 21 and appears not predictable. The patient had severe aortic regurgitation and a dilated left ventricle, and presumably died from arrhythmia. The patient in the current study had an uneventful operative course, dying 35 days postoperatively. Rheumatic heart disease was associated with improved survival after four years when assessed with multivariate analysis but with a small increase in mortality in the early post-operative period using univariate analysis. This finding is in agreement with Gillinov et al., 13 who demonstrated that rheumatic aetiology conferred a survival advantage in both repaired and replaced mitral valves in patients undergoing concomitant aortic replacement. Possible explanations include preserved left ventricular function in patients with mitral stenosis and the difference in the natural history of rheumatic valvular disease compared to degenerative or ischaemic valvular disease. A large number of patients in the current study received mechanical valves (86%). A previous study has shown that the linearised incidence of haemorrhage associated with anticoagulation was 3.2%/pt year for DVR. 22 As a result, there has been a shift towards aortic replacement with mitral repair or placement of a bioprosthetic valve in the mitral valve, where a lower INR can be used (compared to mitral valve replacement). Gillinov et al. 13 demonstrated that the mitral valve repair has a survival advantage over replacement, but this does not obviate the need for anticoagulation in cases where a mechanical aortic valve is implanted. This survival advantage persists regardless of aetiology, age, LVEF, NYHA class. Importantly, the durability of repaired mitral valves exceeded the durability of bioprosthetic valves, hence it would seem appropriate to repair the mitral valve when feasible. The freedom from mitral valve replacement is lower in patients with a rheumatic aetiology who have undergone mitral valve repair, presumably due to progression from rheumatic disease, but the advantages of repair outweighed the risk of re-operation. Limitations This study suffers some limitations. The retrospective nature of the study can introduce selection bias. Furthermore, follow-up was not complete with 94% of patient data available for long-term survival modelling. Finally, the relatively small number of events (26 deaths of 113 patients) results in wide confidence intervals which could reduce the validity of the findings. Despite these caveats, this study's strength is that it represents a medium sized cardiac surgical unit's results which are more likely to be in keeping with current local practice and outcomes. Conclusion Within the limits of a retrospective study, the data presented is a modern day cohort of patients undergoing DVR in a medium sized single centre. The results are in keeping with published literature, and analysis of late mortality demonstrates a high risk group consisting of pre-existing hypertension and renal impairment/failure. Age, NYHA status and LVEF were not predictive of early or long-term mortality, which needs to be factored into the decision to repair/replace valves in an increasing aging population. Acknowledgements Dr Kim Connelly is supported by a Postgraduate research award from the National Heart Foundation of Australia PC 02M 0875, a Pfizer Cardiovascular Research grant, an NHMRC Neil Hamilton Fairley scholarship (ID 447712) and a TACTICS grant (Canada). References 1 T. Caus P. Rouviere F. Collart A. Mouly-Bandini J.R. Monties T. Mesana Late results of double-valve replacement with biologic or mechanical prostheses Ann Thorac Surg 71 2001 S261 S264 2 T. Caus J.N. Albertini Y. Chi F. Collart J.R. Monties T. Mesana Multiple valve replacement increases the risk of reoperation for structural degeneration of bioprostheses Ann Thorac Surg 8 1999 376 383 3 A.C. Fiore M.T. Swartz T.G. Sharp K.A. Kesler H.B. Barner K.S. Naunheim G.L. Grunkemeier D.A. Moroney G.C. Kaiser Double-valve replacement with Medtronic-Hall or St. Jude valve Ann Thorac Surg 59 1995 1113 1118 [discussion 19] 4 U. Bortolotti A. Milano A. D’Alfonso C. Piccin G. Mecozzi P. Magagna A. Fabbri A. Mazzucco Evaluation of valve-related complications in patients with Sorin Bicarbon prosthesis: a seven-year experience J Heart Valve Dis 10 2001 795 801 5 J. Turina T. Stark B. Seifert M. Turina Predictors of the long-term outcome after combined aortic and mitral valve surgery Circulation 100 1999 II48 II53 6 C.W. LaSalle J.F. Csicsko M.J. Mirro Double cardiac valve replacement: a community hospital experience Indiana Med 86 1993 422 426 7 A.C. Fiore H.B. Barner M.T. Swartz L.R. McBride A.J. Labovitz K.J. Vaca J. St Vrain G.L. Grunkemeier G.C. Kaiser Mitral valve replacement: randomized trial of St. Jude and Medtronic Hall prostheses Ann Thorac Surg 66 1998 707 712 discussion 12–3 8 F. Santini G. Casali F. Viscardi A. Favaro G.B. Luciani S. Pentiricci M. Lusini A. Rossi A. Mazzucco The CarboMedics prosthetic heart valve: experience with 1,084 implants J Heart Valve Dis 11 2002 121 126 [discussion 27] 9 J.P. Remadi O. Baron C. Tribouilloy J.C. Roussel O. Al-Habasch P. Despins J.L. Michaud D. Duveau Bivalvular mechanical mitral-aortic valve replacement in 254 patients: long-term results—a 22-year follow-up Ann Thorac Surg 76 2003 487 492 10 L.H. Edmunds Jr. R.E. Clark L.H. Cohn G.L. Grunkemeier D.C. Miller R.D. Weisel Ad Hoc Liaison Committee for Standardizing Definitions of Prosthetic Heart Valve Morbidity of The American Association for Thoracic Surgery and The Society of Thoracic Surgeons Guidelines for reporting morbidity and mortality after cardiac valvular operations J Thorac Cardiovasc Surg 112 1996 708 711 11 M. Cleves W. Gould R. Gutierrez An introduction to survival analysis using stata 2002 Stata Press College Station, TX 12 X.M. Mueller H.T. Tevaearai F. Stumpe A.P. Fischer M. Hurni P. Ruchat L.K. von Segesser Long-term results of mitral-aortic valve operations J Thorac Cardiovasc Surg 115 1998 1298 1309 13 A.M. Gillinov E.H. Blackstone D.M. Cosgrove 3rd J. White P. Kerr A. Marullo P.M. McCarthy B.W. Lytle Mitral valve repair with aortic valve replacement is superior to double valve replacement J Thorac Cardiovasc Surg 125 2003 1372 1387 14 Y. Bar-El Z. Adler A. Kophit V. Kertzman S. Sawaed A. Ross O. Cohen S. Milo Myocardial protection in operations requiring more than 2 h of aortic cross-clamping Eur J Cardiothorac Surg 15 1999 271 275 15 P. Raskin R. Guthrie J. Flack R. Reeves R. Saini The long-term antihypertensive activity and tolerability of irbesartan with hydrochlorothiazide J Hum Hypertens 13 1999 683 687 16 L. Bistritz H. Amad P. Tandon J. Man J.C. Mullen F.A. McAlister Use of angiotensin-converting enzyme inhibitors after coronary artery bypass grafting in patients with diabetes mellitus Am J Cardiol 92 2003 77 79 17 M. Packer M.B. Fowler E.B. Roecker A.J. Coats H.A. Katus H. Krum P. Mohacsi J.L. Rouleau M. Tendera C. Staiger T.L. Holcslaw I. Amann-Zalan D.L. DeMets Effect of carvedilol on the morbidity of patients with severe chronic heart failure: results of the carvedilol prospective randomized cumulative survival (COPERNICUS) study Circulation 106 2002 2194 2199 18 L.J. Dacey J.Y. Liu J.H. Braxton R.M. Weintraub J. DeSimone D.C. Charlesworth S.J. Lahey C.S. Ross F. Hernandez Jr. B.J. Leavitt G.T. O’Connor Long-term survival of dialysis patients after coronary bypass grafting Ann Thorac Surg 74 2002 458 462 [discussion 62–3] 19 A.S. Go G.M. Chertow D. Fan C.E. McCulloch C.Y. Hsu Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization N Engl J Med 351 2004 1296 1305 20 S. MacMahon Blood pressure and the risk of cardiovascular disease N Engl J Med 342 2000 50 52 21 F. Armenti L.W. Stephenson L.H. Edmunds Jr. Simultaneous implantation of St. Jude Medical aortic and mitral prostheses J Thorac Cardiovasc Surg 94 1987 733 739 22 S.S. Khan A. Trento M. DeRobertis R.M. Kass M. Sandhu L.S. Czer C. Blanche S. Raissi G.P. Fontana W. Cheng A. Chaux J.M. Matloff Twenty-year comparison of tissue and mechanical valve replacement J Thorac Cardiovasc Surg 122 2001 257 269