Risks of Mitral Valve Replacement and

Transcription

1 Risks of Mitral Valve Replacement and Mitral Valve Replacement with Coronary Artery Bypass James A. Magovern, M.D., John L. Pennock, M.D., David B. Campbell, M.D., William S. Pierce, M.D., and John A. Waldhausen, M.D. ABSTRACT One hundred thirty consecutive patients who underwent mitral valve replacement (MVR) or MVR with coronary artery bypass grafting (CABG) using cold crystalloid cardioplegic solution were analyzed to determine operative mortality and risk factors. Twenty-eight patients had mitral stenosis (MS), 37 had mitral regurgitation (MR), 37 had mixed MS and MR, 23 had MR with coronary artery disease (CAD), and 5 had MS with CAD. Preoperative pulmonary artery pressure, pulmonary capillary wedge pressure, and cardiac index were not different among groups, but patients with MR and CAD had a significantly higher left ventricular end-diastolic pressure (LVEDP) and a significantly lower ejection fraction than other groups. Mortality was 7.% for patients with MS, 5.4% for MR, 8.% for mixed MS and MR, 0 for MS with CAD, and 2.7% for MR and CAD. Overall mortality was 9.2%. Eleven patients had emergency operations for cardiogenic shock with a mortality of 45%. Nineteen additional patients in New York Heart Association (NYHA) Functional Class IV had MVR or MVR plus CABG with a mortality of 26%. Sixteen patients required intraaortic balloon pump assistance, and 9 survived. Four patients with MR and CAD required the left ventricular assist device, and 3 survived. Excluding patients who had emergency operations, overall mortality was 5.8%. Excluding patients who had emergency operations and patients in NYHA Functional Class IV, overall mortality was 2%. Factors associated with death were cardiogenic shock, NYHA Class IV, LVEDP greater than 5 mm Hg (6% mortality), and age greater than 60 years (5% mortality). Symptomatic mitral valve disease results in progressive deterioration of cardiac function, with 0-year survival of less than 40% for patients in New York Heart Association (NYHA) Functional Class with mitral stenosis (MS), mitral regurgitation (MR), or mixed MS and MR [l, 2. Mitral valve replacement (MVR) has greatly improved the prognosis for these patients, but it involves From the Division of Cardiothoracic Surgery, Department of Surgery, College of Medicine, The Pennsylvania State University and The Milton S. Hershey Medical Center, Hershey, PA. Accepted for publication July 0, 984 Address reprint requests to Dr. Pennock, Department of Surgery, College of Medicine, The Pennsylvania State University, Hershey, PA risks of operative death and complications related to the prosthetic valve [3]. Mortality for MVR has fallen from approximately 5% in the late 960s [4] to less than 0% in the late 970s and early 980s [5, 6. Recent reports, however, discuss patients seen from 970 to 980, many of whom were operated on without cardioplegic solution for myocardial protection and without sophisticated intensive care unit monitoring postoperatively. In addition, patients with combined mitral valve dysfunction and coronary insufficiency now undergo combined MVR and coronary artery bypass grafting (CABG), whereas in the early 970s some patients with both ischemic and valvular disease underwent valve replacement without coronary revascularization [5, 7. To determine the current operative mortality for MVR and for MVR plus CABG, we have reviewed our experience with these operations from 977 to 982. Because the risks for individual patients vary with the severity of cardiac dysfunction, we have analyzed preoperative and intraoperative factors to establish the mortality for specific groups of patients. Late follow-up studies are also presented. Material and Methods The last 30 consecutive patients having MVR or MVR with CABG were reviewed. Patients having emergency as well as elective operations were included. Patients with associated aortic, tricuspid, or pulmonary valve disease were excluded. During and after operation, all patients were monitored with either a Swan-Ganz catheter or a left atrial line, and postoperatively all were observed in the intensive care unit for at least 48 hours. Multidose cold crystalloid cardioplegic solution, infused through the aortic root and monitored with a myocardial temperature probe, was used for myocardial protection in each patient [8]. Cardiopulmonary bypass using a roller pump, bubble oxygenator, and moderate systemic hypothermia (25 C) was routinely employed. Operative data, including cardiopulmonary bypass time, aortic cross-clamp time, and type of valve used, were recorded for each patient. Postoperative cardiac enzymes were measured in 85% of the patients. Cardiac catheterization was performed on all patients except. Ninety-eight percent of the patients had both right and left heart catheterization, and 97% of those older than 35 years had coronary arteriography. Coronary artery bypass was done for all angiographically severe stenoses. Grouped according to the preoperative diagnosis, 28 patients had MS, 37 had MR, and 37 had mixed MS and 346

2 347 Magovern et al: Risks of Mitral Valve Replacement Table. Preoperative Hemodynamic Data According to Diagnosis Variable Normal MS MR MS + MR MR + CAD MS + CADb LVEDP (mm Hg) 2 i 3 3 * 5 7 i 9 t " i EF (%) c 50 2 " CI (L/min/m2) 2.5 f i i PAP (mm Hg) ? 3 32? 2 36 f 2 28? PCWP (mm Hg) 2 i 3 25 i 7 22 i 23 i 6 9 f 24 i 0 "Statistical significance: p < bpatients in this group not included in statistical analysis because of small numbers. LVEDP = left ventricular end-diastolic pressure; EF = ejection fraction, calculated by planimehy from the ventriculogram; C = cardiac index; K P = mean pulmonary artery pressure; PCWP = pulmonary capillary wedge pressure; MS = mitral stenosis; MR = mitral regurgitation; CAD = coronary artery disease. MR. Twenty-eight patients with coronary artery disease (CAD) underwent MVR; in this group, 5 had MS with CAD and 23 had MR with CAD. Four patients had acute MR secondary to acute myocardial infarction, but the, remainder had chronic MR associated with coronary insufficiency. In most patients it was not possible to determine whether the regurgitation was induced by or coincidental with the myocardial ischemia. The preoperative hemodynamic data, grouped according to preoperative diagnosis, are presented in Table. The clinical data are as follows. There were 86 women (66%) and 44 men (34%). The average age was 55 years. Thirty-two patients (25%) were in NYHA Class, 68 patients (52%) were in Class, and 30 patients (23%) were in Class IV. Eleven patients had emergency operations for cardiogenic shock secondary to acute myocardial infarction (4 patients), acute bacterial endocarditis (4 patients), primary rupture of the chordae tendineae (2 patients), and a thrombosed prosthetic valve ( patient). All of these patients were considered to be in NYHA Class IV. Seventy-seven patients (59%) were in normal sinus rhythm, 5 patients (39%) were in atrial fibrillation, had a pacemaker, and had a junctional rhythm. Mechanical valves were used in 62 patients (48%); 42 patients (32%) received Bjork-Shiley and 20 (5%), Starr-Edwards valves. Sixty-eight patients (52%) had tissue valves inserted; 60 (46%) received Hancock valves; 6 (5%), Carpentier-Edwards valves; and 2 (2%), Ionescu-Shiley prostheses. Anticoagulation therapy with crystalline warfarin sodium (Coumadin) was started after the chest tubes were removed, usually on the second or third postoperative day. All patients with mechanical valves and all patients who were in atrial fibrillation received anticoagulants indefinitely. A few patients with tissue valves who were in normal sinus rhythm were taken off Coumadin after 6 months if no embolic complications had occurred. When difficulty in weaning from cardiopulmonary bypass was encountered, volume infusion and the administration of catecholamines were tried. If these measures failed, counterpulsation with an intraaortic balloon pump was started. If the patient still could not be weaned from cardiopulmonary bypass, a left ventricular assist device was employed. The technique and indications have been described previously [9]. Patients were weaned from the ventricular assist pump by briefly discontinuing pumping and measuring the patient's cardiac output by thermodilution. As the patient's heart recovered function, the pump rate was decreased to allow the native heart to eject. The pump was removed in the operating room after adequate hemodynamics were demonstrated several times with the pump turned off for a few minutes. Six patients had an intraaortic balloon pump inserted preoperatively because of cardiogenic shock in an attempt to stabilize cardiac function while preparing for emergency operation. An additional 0 patients could not be weaned from cardiopulmonary bypass and had an intraaortic balloon pump placed postoperatively. No patient having an elective operation had an intraaortic balloon pump placed preoperatively. Three patients who had persistent low cardiac output despite the administration of catecholamines and intraaortic balloon pumping were placed on a left ventricular assist pump. A fourth patient, in whom left ventricular failure developed several hours after the operation, was returned to the operating room and placed on the assist pump. A summary of data for patients requiring the left ventricular assist pump is presented in Table 2. Operative mortality was defined as death within thirty days of operation. One-month follow-up was available for all patients, and long-term follow-up was available for 97%. Each patient was contacted by a physician and specifically questioned regarding activity, current medications, and subsequent cardiac catheterization or operation, as well as the incidence of complications such as cerebrovascular accident (CVA), peripheral embolism, or hemorrhage. Chi-square tests were used to determine whether preoperative catheterization data, cardiopulmonary bypass times, aortic cross-clamp times, age, functional status, and postoperative cardiac enzyme levels were associated with death. An analysis of variance, followed by a Student-Newman-Keuls test, was used to determine whether differences existed between groups that were organized by preoperative diagnosis. Actuarial analysis was used to determine late mortality [lo].

3 348 The Annals of Thoracic Surgery Vol 39 No 4 April 985 Table 2. Clinical and Hemodynamic Data for Patients Requirina a Left Ventricular Assist Pumv Preop. Age LVEDP No. of Days Survival Postop. No. (yr) (mm Hg) Preop. Diagnosis Opera tion Assisted (mo) NYHA Class Acute MI, MR, cardiogenic shock MVR, CABG x > 24 I MR, CAD MVR, CABG x > 0 I MR, CAD, CHF MVR, CABG x >6 I MR, CAD, CHF MVR, CABG x Died... LVEDI' = left ventricular end-diastolic pressure; MI = myocardial infarction; MR = mitral regurgitation; CAD = coronary artery disease; CHF = congestive heart failure; MVR = mitral valve replacement; CABG = coronary artery bypass grafting; NYHA = New York Heart Association. Early Results Early Findings Overall early mortality was 9.2%. The causes of early death are listed in Table 3. The presence of cardiogenic shock resulted in an operative mortality of 45%, whereas the mortality for hemodynamically stable patients was 5.8%. Preoperative functional status also was related strongly to early mortality, with an operative death rate of 2% for patients in Classes I and I and 26% for patients in Class IV. Patients older than 60 years had an incidence of postoperative death nearly four times higher than did those less than 60 years old. Clinical features associated with high mortality are depicted in Figure. In the analysis of hemodynamic factors, only elevated left ventricular end-diastolic pressure (LVEDP) was found to correlate significantly with operative death. When LVEDP was less than 5 mm Hg, mortality was Table 3. Causes of Operative and Late Deaths Cause EARLY DEATHS Low cardiac output Sudden death Renal failure Cerebrovascular accident Disruption of atrioventricular groove Bacterial endocarditis Total LATE DEATHS No Sudden death 5 Bacterial endocarditis 2 Cardiac failure Bleeding esophageal varices Ruptured spleen secondary to trauma Thrombosed prosthetic valve Gastrointestinal bleeding Unknown Total 3 2.8%, but when LVEDP was greater than 5 mm Hg, mortality rose to 5.8%. Mortality was greater with a depressed cardiac index and ejection fraction, and with an increased pulmonary capillary wedge pressure and mean pulmonary artery pressure. However, these trends were not statistically significant. Patients with aortic cross-clamp times of more than 00 minutes had a higher mortality, but the elevation of myocardial enzymes of creatine phosphokinase (CPK-MB) to more than three times the normal level had no association with operative death. Mortality for the group with MR and CABG was significantly higher than that for other groups. The mortality by preoperative diagnosis is illustrated in Table 4. The number of patients with MS and CAD was quite small; therefore, the data are presented but are not included in the statistical analysis. The group with MR and CABG had a significantly higher LVEDP and crossclamp time and a significantly lower ejection fraction than the other groups. A total of 6 patients could not be weaned from cardiopulmonary bypass in the postoperative period. Six of these patients were in cardiogenic shock preoperatively, and 3 of them ultimately died of extensive myocardial infarction. Two patients recovered after placement of an intraaortic balloon pump, and patient, who initially experienced failure with intraaortic balloon pumping, recovered after placement of a left ventricular assist Pump. Ten patients could not be weaned from cardiopulmonary bypass after elective operation, and 6 patients in this group survived. Four of them survived with the help of an intraaortic balloon pump, and the other 2 survived after placement of a left ventricular assist device. One patient died of massive hemorrhage shortly after placement of the left ventricular assist pump. One patient died of right ventricular failure following MVR for mixed MS and MR. Normal cardiac function was restored in 2 patients with the aid of an intraaortic balloon pump. However, l subsequently died of a CVA, and the other died suddenly just prior to discharge from the hospital. All of the patients who required a left ventricular assist pump had a preoperative diagnosis of MR and CAD.

4 349 Magovern et al: Risks of Mitral Valve Replacement I00 MORTALITY 50 5 I* 459, TOTAL CARDIO- ELECTIVE AORTIC MR + LVEDP > AGE > GENIC NYHA CROSS CABG 5mmHg 60years SHOCK CLASSIV CLAMP > 00 miri Fig I. Factors ussoc ted with increased early mortality. (NYHA = New York Heart Association; MR = initral regurgitation; CABG = coronary artery Iypass graft; LVEDP = left ventricular elid-diastolic pressure; * = p < 0.05.) Late Findings Follow-up information was available for 97% of the patients who survived the operation. The period of observation ranged from 6 months to six years, with a mean follow-up time of 29 months. Thirteen late deaths occurred, but 8% of all patients operated on were still alive at the end of the study period. Actuarial 5-year survival was 75%. The causes of late death are listed in Table 3. Eight late CVAs occurred for a rate of 2.9% per patient-year. Two of these events were secondary to bacterial endocarditis, but the causes of the remaining CVAs could not be determined. Patients with mechanical valves did not have a higher incidence of CVA than those with tissue valves. Similarly, patients with atrial fibrillation were not at higher risk of CVA when compared with those in sinus rhythm. Four patients had prosthetic valve endocarditis during the study period, and 2 of them died. Six patients required reoperation: 2 for bacterial endocarditis, 2 for perivalvular leak, for valve thrombosis, and for tissue valve stenosis secondary to calcification. Seven patients had late postoperative hemorrhage. There were two cerebral hemorrhages, which resolved Table 4. Mortality by Preoperative Diagnosis Diagnosis No. Percent Mitral stenosis 2/28 7. Mitral regurgitation Mitral stenosis and 3/37 8. regurgitation Mitral regurgitation and " coronary artery bypass Mitral stenosis and coro nary artery bypass "Significantly greater than other groups ( p < 0.0). without neurologica. deficit, and one episode of hematuria, which also resolved. Three patients died of hemorrhage while taking Coumadin. One patient died at home of massive upper gastrointestinal bleeding, but no postmortem examination was done. One patient died of a bleeding peptic ulcer that complicated an existing prosthetic valve endocarditis, and patient died of a ruptured spleen. One patient died of bleeding esophageal varices secondary to hepatic cirrhosis, which had been present preoperatively. However, she was not taking Coumadin at the time of her death. Significant improvement in functional status occurred in most patients regardless of preoperative status. At a mean follow-up of 29 months, all patients who survived emergency operation were still alive and in NYHA Functional Class I or. Combining patients in Class IV who had elective and emergency operations, 8 of 20 patients who survived operation are alive: 2 in Class I, 5 in Class, and in Class. Similar results have been obtained for patients who were in Functional Classes I and preoperatively; these results are summarized in Figure 2. Long-term survival was not statistically different among patients grouped by preoperative diagnosis. Despite a higher operative mortality, late mortality for patients with MR and CABG was not higher than for the other groups. Comment We have reviewed our last 30 consecutive patients having either MVR or MVR plus CABG to determine the current operative mortality for different groups and the factors that predict increased operative risks. In addition, we have calculated long-term results at a mean follow-up of 29 months. Some investigators have cited an operative mortality for MVR of less than 9% [5, 6. These series contain only patients undergoing elective operations dating from 970; hence, many of the operations were done without cardioplegic solution and without the sophisticated monitoring now employed. In earlier reports, there was no uniform policy regarding combined CABG and MVR, and some patients with combined valvular and ischemic disease did not have revascularization at the time of valve replacement.

5 350 The Annals of Thoracic Surgery Vol 39 No 4 April 985 IV /: I II IV NYHA FUNCTIONAL CLASS I II I II pre-op Fig 2. Late functional resultsfor patients suwiving the operation. (NYHA = New York Heart Association; * = worse, dead, or lost to follow-up.) Our study presents a series of patients with MVR and MVR plus CABG who had both elective and emergency operations in which a uniform treatment plan was used. All patients received cold crystalloid cardioplegic solution for myocardial protection, and all were monitored in a modern intensive care unit [8]. Nearly all patients older than 35 years had coronary arteriography, and all major coronary lesions were repaired with CABG. We have found that operative risk is strongly influenced by several preoperative factors. Patients in cardiogenic shock have the highest operative mortality. Nearly one-half of our patients undergoing emergency operation were in shock secondary to myocardial infarction with acute mitral regurgitation. A high mortality is to be expected with this group. Daggett and associates [ll] reported greatly improved results by treating ventricular septa defects secondary to myocardial infarction with early operation. For patients requiring urgent MVR and CABG, similar results can be expected with more expeditious diagnosis and operation. Previous reports have emphasized that the preoperative functional status of the patient is an important determinant of operative mortality [5, 7. We have confirmed this observation. Of the 2 operative deaths in this series, only 2 occurred in patients who were not in shock or NYHA Functional Class IV preoperatively. The implications for early operation are evident, because waiting until a patient is severely disabled will greatly increase the operative risk. We have also confirmed reports of increased mortality in older patients; most likely, this reflects more severe cardiac disease in this group [5. Elevated LVEDP is a good indicator of marked cardiac dysfunction and, in this study, provided the best hemodynamic predictor of operative risk. Mortality increased post-op directly with rising LVEDP. Depressed ejection fractions have been used to predict results in patients with CAD. In patients with MR, however, the ventricle usually empties well because of the incompetent valve, and the ejection fraction is not as useful a variable unless antegrade ejection is calculated. In many of our patients antegrade ejection was not calculated, and therefore we did not analyze this variable in relation to operative death. Our data showed a trend toward increased mortality with elevated pulmonary artery pressure and elevated pulmonary capillary wedge pressure, but the differences were not statistically significant. Perhaps with larger numbers of patients these trends would be substantiated. Patients with MR and CAD had a mortality three times higher than other groups. Within this group, women had a higher mortality than men (33% versus 4%; p < 0.05); however, when divided by sex and diagnosis, the numbers in each group were small, and it was not possible to tell if sex was an independent predictor of outcome. The higher mortality in patients with MR and CAD can in part be explained by poorer preoperative cardiac function in this group of patients than in other groups, as evidenced by higher LVEDP. In addition, these patients had a substantially longer aortic cross-clamp time. It is difficult, however, to determine what effect this had on operative mortality, given present methods of myocardial protection and current techniques for temporary ventricular support, which have enabled complex operations to be performed with good results. In addition, there is an inherent bias in correlating aortic cross-clamp times with operative mortality, because the most severe cardiac problems often take the longest time to repair. It is difficult to separate the effect of the cross-clamp time from the underlying disease. Nonetheless, some surgeons believe that aortic cross-clamp times greater than 00 minutes can result in depressed myocardial function [2]. Our patients with MVR and CABG had a mean cross-clamp time of minutes, nearly twice the time

6 35 Magovern et al: Risks of Mitral Valve Replacement for isolated MVR, which might have put them at a higher risk for cardiac failure. A more important factor contributing to high mortality in this group relates to hemodynamic changes following valve replacement. With MR the left ventricle becomes dilated and loses contractility, but because of the incompetent valve, ventricular afterload is reduced by ejection into the atrium. After valve replacement, this mechanism is eliminated and the ventricle faces increasing afterload. With the addition of ischemic ventricular dysfunction and the longer cross-clamp time necessary for operation, this acute increase in ventricular afterload can precipitate ventricular failure despite adequate myocardial protection and coronary revascularization. In this regard, it is important to note that perioperative myocardial infarction, as evidenced by a threefold elevation of CPK-MB, had no association with operative death. Thus, patients who could not be weaned from cardiopulmonary bypass had global failure in contractility rather than a discrete infarction. Experimental studies have shown that a brief period of ischemia to the left ventricle results in a depression of myocardial adenosine triphosphate levels, which persists for several days before returning to normal. Reimer and colleagues [3] induced myocardial ischemia by occluding coronary blood flow for 5 minutes and studied myocardial metabolism in the reperfusion period. They demonstrated that depression of adenosine triphosphate and adenine nucleotides persisted for several days in the reperfused ventricle. Histological examination at 72 hours revealed no myocardial necrosis but did show ultrastructural abnormalities. Hence, an ischemic injury not severe enough to produce myocardial infarction can depress ventricular function for several days after the restoration of coronary blood flow. Ventricular failure after cardiac operations remains a serious complication. When it follows an elective cardiac surgical procedure with adequate myocardial protection and appropriate repair of the cardiac defect, recovery of function should be possible. It appears from our study that the majority of ventricular failures after elective cardiac operations are reversible. We had 0 patients who could not be weaned from cardiopulmonary bypass after elective operation. Four patients recovered completely with intraaortic balloon pumping and are now in NYHA Class I or. Three patients had persistent ventricular failure despite balloon pumping; 2 of the 3 recovered after the insertion of a left ventricular assist pump, but the third died of a coagulopathy shortly after the insertion of the device. Three additional patients eventually died, but only died of cardiac failure. This was a woman with MS and MR who died of right ventricular failure secondary to persistent pulmonary hypertension. The other 2 patients recovered normal cardiac output, but died of a CVA and the other of sudden cardiac arrest. Thus, with appropriate support, myocardial contractility will usually return, but mortality in these patients remains high secondary to arrhythmias or failure of other organ systems. The left ventricular assist device was successful in 3 of 4 patients. Currently all survivors are at home and in Functional Class I or. It appears that this technique offers great promise for patients who cannot be weaned from cardiopulmonary bypass but who have had appropriate repair of the cardiac defect, and who do not have extensive transmural myocardial infarction. Ventricular failure after preoperative cardiogenic shock remains a difficult problem with high mortality. During preparation for emergency operation, an intraaortic balloon pump was inserted in 6 patients. Two patients with acute MR secondary to ruptured chordae tendineae, but without infarction, recovered and are now in NYHA Class I. Four patients had acute MR secondary to myocardial infarction, and 3 died in the operating room, despite balloon pumping and catecholamines. None of these patients was a candidate for a ventricular assist device because of advanced age and extensive myocardial necrosis. One patient with MR resulting from infarction could not be weaned from cardiopulmonary bypass with an intraaortic balloon pump, but recovered with a left ventricular assist pump and is now in NYHA Class I. In this small group of patients, preoperative balloon pumping did not improve survival in those with myocardial necrosis but did appear to be helpful in those without infarction. The role of the ventricular assist pump for support of patients operated on for cardiogenic shock is still under evaluation. We were successful in one instance, but the widespread success of this technique will depend on methods of identifying patients with enough viable myocardium to sustain cardiac function after assist pumping is stopped. The problems of late mortality and morbidity remain troublesome. Two of the 3 late deaths were due to unrelated causes, but the remainder could be attributed to either the underlying cardiac disease or the prosthetic valve. The most common cause of death was sudden cardiac arrest, but we could find no predictor of this event. Preoperative functional class, LVEDP, age, and preoperative cardiogenic shock could not be correlated with long-term survival. These results are in contrast to reports that have shown lower survival for patients in Functional Class IV [2]. Similarly, we found no significant differences in survival related to the type of mitral dysfunction, whereas several groups have reported poor 5-year survival for patients with MR [2, 5. We are hesitant, however, to draw firm conclusions from these data because our follow-up was only 29 months. Perhaps differences will become apparent with time. We have found good functional improvement in all groups of patients, regardless of preoperative status. Our rates of CVA, hemorrhage, and endocarditis were comparable to those in other reports in the literature [4]. Neither the presence of sinus rhythm nor the use of tissue valves afforded any protection from stroke, and all patients taking Coumadin risked hemorrhage. The development of endocarditis is an unusual and unpredictable catastrophe. Thus, all patients are at risk for late

7 352 The Annals of Thoracic Surgery Vol 39 No 4 April 985 complications following valve replacement regardless of preoperative status or postoperative results. References. Rapaport E: Natural history of aortic and mitral valve disease. Am J Cardiol 35:22, Oleson K: The natural history of 27 patients with mitral stenosis under medical treatment. Br Heart J 24:349, Fowler NO, Van Der Bel-Kahn JM: Indications for surgical replacement of the mitral valve. Am J Cardiol44:48, Kirklin JW, Pacific0 AD: Surgery for acquired valvular heart disease. N Engl J Med 288:33, Chaffin JS, Daggett WM: Mitral valve replacement: a nineyear follow-up of risks and survival. Ann Thorac Surg 2732, Lepley D Jr, Flemma RJ, Mullen DC, et al: Long-term follow-up of the Bj(irk-Shiley prosthetic valve used in the mitral position. Ann Thorac Surg 30:64, Miller DC, Stinson EB, Rossiter SJ, et al: Impact of simultaneous myocardial revascularization on operative risk, functional result and survival following mitral valve replacement. Surgery 84:848, Tyers GFO, Manley NJ, Williams GH, et al: Preliminary clinical experience with isotonic hypothermic potassium-induced arrest. J Thorac Cardiovasc Surg 74:674, Pierce WS, Parr GVS, Myers JL, et al: Ventricular assist pumping in patients with cardiogenic shock after cardiac surgery. N Engl J Med 305:60, Grunkemeier GL, Starr A: Actuarial analysis of surgical results: rationale and method. Ann Thorac Surg 24:404, 977. Daggett WM, Buckley MJ, Akins CW, et al: Improved results of surgical management of post-infarction ventricular septa defect. Ann Surg 96:269, Stiles QR, Kirklin JW: Myocardial preservation symposium. J Thorac Cardiovasc Surg 82:870, Reimer KA, Hill MC, Jennings RB: Prolonged depletion of ATP and of the adenine nucleotide pool due to delayed resynthesis of adenine nucleotides following reversible myocardial ischemic injury in dogs. J Mol Cell Cardiol 3229, Angell WV, Angell JD, Kosek JC: Twelve-year experience with glutaraldehyde preserved porcine xenografts. J Thorac Cardiovasc Surg 83:493, 982 Notice from the Southern Thoracic Surgical Association The Thirty-second Annual Meeting of the Southern Thoracic Surgical Association will be held at the Boca Raton Hotel and Club, Boca Raton, FL, November 7-9, 985. There will be a $00 registration fee for nonmember physicians except for guest speakers, authors and coauthors on the program, and residents. There will be a $25 registration fee for attendees of the Postgraduate Course on Thursday, November 7, 985. Members wishing to participate in the Scientific Program should submit abstracts in triplicate by May 5, 985, to Harvey W. Bender, Jr., M.D., Secretary- Treasurer, Southern Thoracic Surgical Association, 2 2st Ave S, Suite 338, Nashville, TN Abstracts must be submitted on the Southern Thoracic Surgical Association abstract submission form. These forms may be obtained from the Secretary-Treasurer s office or in this issue (page A-2) of The Annals of Thoracic Surgery. All slides used during the presentation must be 35 mm. Manuscripts of accepted papers should be submitted to The Annals of Thoracic Surgery prior to the 985 meeting or to the Secretary-Treasurer at the opening of the Scientific Session. Applications for membership should be completed by July, 985, and forwarded to Harry K. Daugherty, M.D., 960 Randolph Rd, Charlotte, NC Harvey W. Bender, Jr., M.D. Secretary-Treasurer Southern Thoracic Surgical Association 2 2st Ave S, Suite 338 Nashville, TN 3722