Frequency of Acute Myocardial Infarction in Valve Repairs Versus Valve Replacement for Pure Mitral Regurgitation Timothy P. Obarski, DO, Floyd D. Loop, MD, Delos M. Cosgrove, MD, Bruce W. Lytle, MD, and William J. Stewart, MD
Three hundred thirty-five conse&ive isokted mitral valve operations for mitral regurgitation artery disinpatkntswithnosignificantcoromuy ease were reviewed over a 26-month period for the presmce of a perkperative acute myocardial infarction (AMI). Of 224 patknts undergoing mitral valve repair 12 (5.4%) had ektrocardiiaphii and cardiac enzyme evidence of perioperative AMI develop. Of ill patients undergoing mitral valve perkperative AMI develop replacementnonehad as determined by ekctmcardkgraphkandenzyme criteria (p = 0.01). All 12 infarctkns after valve repair involved ths Werior wall by ektracardioAlthough no graphk 0~ e&ocardiographiccriteria. patient had significant clinical difficulty in recovery, 7 of the 12 patients (sS%) were left with Q waves upon hospital discharge. The etklogy of the AMI is believed to be air emboll inkeduced at ths time of testing valve compekknce during kft ventkular intwffkth under presswe. Changes in surgical technique may reduce or eliminate this compliition. (Am J Car-did 1990;sS:887-890)
perative repair of the mitral valve has been performed since 1922l with variable success. Since the technique of Carpentier et al* was described in 1971 popularity of valve repair for mitral regurgitation has increased. Observational studies comparing mitral valve repair versus mitral valve replacement have noted significant clinical benefit in favor of the former including a lower operative mortality,3-6 reduced perioperative morbidity, improved long-term survival’ and significantly fewer episodes of thromboembolism. Leaving the mitral apparatus intact also improves overall left ventricular performance and consequently reduces the incidence of low cardiac output states.s9 A potential complication of any cardiac operation is perioperative acute myocardial infarction (AMI), the incidence of which has not been described in patients undergoing mitral valve repair. We reviewed the incidence of perioperative AM1 in a cohort of patients undergoing surgery for isolated primary mitral regurgitation.
0
METHODS Patient population:
The population consists of a series of 335 consecutive patients over a 26-month period (June 1, 1986 to July 31, 1988) who required mitral valve surgery for mitral regurgitation but did not require coronary artery bypass grafts or other cardiac surgical procedures. No patient had hemodynamically significant coronary atherosclerosis determined by coronary arteriography or exercise stress testing. Patients who had primary myocardial disease, pathology of other valves or previous mitral valve surgery were excluded. Two hundred twenty-four patients underwent mitral valve repair and 111 patients had mitral valve replacement with a prosthetic device during this period. Surgical technique: A median stemotomy was performed and bicaval and aortic cannulation were carried out. When equipment was available a prepump intraoperative Doppler echocardiography study was performed in all patients in whom repair was contemplated. After heparinization cardiopulmonary bypass was instituted, moderate hypothermia was induced, the aorta was cross-clamped and the heart was arrested with cold crystalloid cardioplegia. The left atrium was opened in the interatrial groove and exposure afterward From the Departments of Cardiology and Cardiothoracic Surgery, was achieved using an adjustable self-retaining retracCleveland Clinic Foundation, Cleveland, Ohio. Manuscript received tor. July 21, 1989; revised manuscript received and accepted December 7, The mitral valve was repaired using previously de1989. scribed techniques. lo Cardioplegia solution was infused Address for reprints: William J. Stewart, MD, Department of Carunder pressure into the ascending aorta. Competence diology, Desk F.15, Cleveland Clinic Foundation, One Clinic Center, 9500 Euclid Avenue, Cleveland, Ohio 44195. was tested by insufflation (injecting saline into the left THE AMERICAN
JOURNAL
OF CARDIOLOGY
APRIL
1. 1990
887
ventricle through the mitral orifice and pressurizing the left ventricle by infusing anterograde cardioplegia). Frequently, multiple tests of leaflet coaptation were necessary to ensure valve competence. When regurgitation had been eliminated the left atrium was closed after it was filled with blood and the aortic clamp was removed. The ascending aorta and apex of the left ventricle were aspirated to remove air. When the patient was normothermic and stable, cardiopulmonary bypass was discontinued. Follow-up studies: Intraoperative epicardial or transesophageal Doppler echocardiography was performed in 116 (52%) patients in this series, primarily to test for residual mitral regurgitation and to detect other complications. In 1986, 2.7% of the patients in our series underwent intraoperative echocardiograms compared to 62% in 1987 and 98% in 1988. Electrocardiograms were performed upon arrival into the cardiovascular postoperative unit, daily until discharge to a regular nursing floor and on the day before discharge from the hospital. Cardiac enzymes including total creatinine phosphokinase, creatinine phosphokinaseMB isoenzymes and total serum glutamicoxaloacetic transaminase were routinely drawn at the time of arrival into the unit and at 6 and 18 hours thereafter. Further determinations of enzyme values were at the discretion of the attending physicians but the peak values recorded were used in this study. Before discharge from the hospital a repeat 2dimensional Doppler echocardiogram for evaluation of left ventricular and mitral valve function was performed. Myocardial ischemia was defined electrocardiographically as ST-segment elevation 11.5 mm from the isoelectric baseline in >2 leads referable to a myocardial region. Serial electrocardiograms were analyzed to follow resolution of any ischemic changes, the development of pathologic Q waves (width LO.04 s, depth 225% of the R wave) or symmetrical T-wave inversion. Total creatinine phosphokinase levels > 1,000, absolute creatinine phosphokinase-MB > 133 and serum glutamic-oxaloacetic transaminase levels > 100 III/liter were considered the threshold levels suggesting myocardial damage. l l,i2 StatIsties: A comparison between enzyme level data was made with nonpaired Student’s t test. Comparisons between frequency of myocardial infarction by enzyme and electrocardiographic criteria were made with Fisher’s exact test. RESULTS Ektrocardiograms:
Of the 224 patients undergoing mitral valve repair 12 (5.4%) had evidence of AM1 on the postoperative electrocardiogram. All 12 of these patients initially had ST-segment elevation involving the inferior leads (II, III, aVF). Seven (58%) of the 12 patients (3.1% of 224) had new pathologic Q waves before discharge from the hospital; 2 of these patients had new Q waves on their initial postoperative electrocardiogram. Two of the 12 patients had symmetrical T-wave inversion develop, consistent with a non-Q-wave AMI. Three of the 12 patients resolved their ST-segment ele888
THE AMERICAN
JOURNAL
OF CARDIOLOGY
VOLUME
66
vation and upon hospital discharge had no electrocardiographic evidence of AMI. The extent of the initial ST-segment elevation had no bearing on whether the electrocardiogram would resolve back to baseline level or evolve to an infarct pattern. Of the 111 patients who had mitral valve replacement 2 (1.8%) had electrocardiographic evidence of perioperative myocardial ischemia. The location of the ischemia was in the inferior wall in both. Both patients with ischemic ST-segment changes had new pathologic Q waves evolve in the inferior leads, which persisted to the time of discharge. None of the 14 patients who had perioperative ischemia develop had clinical symptoms consistent with angina pectoris before surgery. Eleven of the 14 (79%) patients had coronary arteriography preoperatively and none had any significant coronary narrowing. The 3 remaining patients did not have preoperative catheterization, but they were under 50-years-old and had normal maximum exercise stress tests using electrocardiographic criteria alone (2 patients) or using electrocardiographic and thallium-201 imaging (1 patient). Cardiac enzymes: Of the 12 patients with electrocardiographic evidence of &hernia after valve repair, enzymatic evidence for myocardial infarction was found on the basis of total creatinine phosphokinase levels in 11 patients (92%), on the basis of creatinine phosphokinase-MB levels in 11 patients (92%) and on the basis of serum glutamic-oxaloacetic transaminase levels in 9 patients (75%). Patients who had new pathologic Q waves develop on their electrocardiogram had a trend toward lower maximum values for both creatinine phosphokinase (mean 1,594 III/liter) and serum glutamiooxaloacetic transaminase (mean 150 II-J/liter) than the patients whose electrocardiogram failed to develop pathologic Q waves (1,826 and 222 IU/liter) (difference not significant). Both mitral valve replacement patients with electrocardiographic ischemia had creatinine phosphokinase levels
Three hundred thirty-five conse&ive isokted mitral valve operations for mitral regurgitation artery disinpatkntswithnosignificantcoromuy ease were reviewed over a 26-month period for the presmce of a perkperative acute myocardial infarction (AMI). Of 224 patknts undergoing mitral valve repair 12 (5.4%) had ektrocardiiaphii and cardiac enzyme evidence of perioperative AMI develop. Of ill patients undergoing mitral valve perkperative AMI develop replacementnonehad as determined by ekctmcardkgraphkandenzyme criteria (p = 0.01). All 12 infarctkns after valve repair involved ths Werior wall by ektracardioAlthough no graphk 0~ e&ocardiographiccriteria. patient had significant clinical difficulty in recovery, 7 of the 12 patients (sS%) were left with Q waves upon hospital discharge. The etklogy of the AMI is believed to be air emboll inkeduced at ths time of testing valve compekknce during kft ventkular intwffkth under presswe. Changes in surgical technique may reduce or eliminate this compliition. (Am J Car-did 1990;sS:887-890)
perative repair of the mitral valve has been performed since 1922l with variable success. Since the technique of Carpentier et al* was described in 1971 popularity of valve repair for mitral regurgitation has increased. Observational studies comparing mitral valve repair versus mitral valve replacement have noted significant clinical benefit in favor of the former including a lower operative mortality,3-6 reduced perioperative morbidity, improved long-term survival’ and significantly fewer episodes of thromboembolism. Leaving the mitral apparatus intact also improves overall left ventricular performance and consequently reduces the incidence of low cardiac output states.s9 A potential complication of any cardiac operation is perioperative acute myocardial infarction (AMI), the incidence of which has not been described in patients undergoing mitral valve repair. We reviewed the incidence of perioperative AM1 in a cohort of patients undergoing surgery for isolated primary mitral regurgitation.
0
METHODS Patient population:
The population consists of a series of 335 consecutive patients over a 26-month period (June 1, 1986 to July 31, 1988) who required mitral valve surgery for mitral regurgitation but did not require coronary artery bypass grafts or other cardiac surgical procedures. No patient had hemodynamically significant coronary atherosclerosis determined by coronary arteriography or exercise stress testing. Patients who had primary myocardial disease, pathology of other valves or previous mitral valve surgery were excluded. Two hundred twenty-four patients underwent mitral valve repair and 111 patients had mitral valve replacement with a prosthetic device during this period. Surgical technique: A median stemotomy was performed and bicaval and aortic cannulation were carried out. When equipment was available a prepump intraoperative Doppler echocardiography study was performed in all patients in whom repair was contemplated. After heparinization cardiopulmonary bypass was instituted, moderate hypothermia was induced, the aorta was cross-clamped and the heart was arrested with cold crystalloid cardioplegia. The left atrium was opened in the interatrial groove and exposure afterward From the Departments of Cardiology and Cardiothoracic Surgery, was achieved using an adjustable self-retaining retracCleveland Clinic Foundation, Cleveland, Ohio. Manuscript received tor. July 21, 1989; revised manuscript received and accepted December 7, The mitral valve was repaired using previously de1989. scribed techniques. lo Cardioplegia solution was infused Address for reprints: William J. Stewart, MD, Department of Carunder pressure into the ascending aorta. Competence diology, Desk F.15, Cleveland Clinic Foundation, One Clinic Center, 9500 Euclid Avenue, Cleveland, Ohio 44195. was tested by insufflation (injecting saline into the left THE AMERICAN
JOURNAL
OF CARDIOLOGY
APRIL
1. 1990
887
ventricle through the mitral orifice and pressurizing the left ventricle by infusing anterograde cardioplegia). Frequently, multiple tests of leaflet coaptation were necessary to ensure valve competence. When regurgitation had been eliminated the left atrium was closed after it was filled with blood and the aortic clamp was removed. The ascending aorta and apex of the left ventricle were aspirated to remove air. When the patient was normothermic and stable, cardiopulmonary bypass was discontinued. Follow-up studies: Intraoperative epicardial or transesophageal Doppler echocardiography was performed in 116 (52%) patients in this series, primarily to test for residual mitral regurgitation and to detect other complications. In 1986, 2.7% of the patients in our series underwent intraoperative echocardiograms compared to 62% in 1987 and 98% in 1988. Electrocardiograms were performed upon arrival into the cardiovascular postoperative unit, daily until discharge to a regular nursing floor and on the day before discharge from the hospital. Cardiac enzymes including total creatinine phosphokinase, creatinine phosphokinaseMB isoenzymes and total serum glutamicoxaloacetic transaminase were routinely drawn at the time of arrival into the unit and at 6 and 18 hours thereafter. Further determinations of enzyme values were at the discretion of the attending physicians but the peak values recorded were used in this study. Before discharge from the hospital a repeat 2dimensional Doppler echocardiogram for evaluation of left ventricular and mitral valve function was performed. Myocardial ischemia was defined electrocardiographically as ST-segment elevation 11.5 mm from the isoelectric baseline in >2 leads referable to a myocardial region. Serial electrocardiograms were analyzed to follow resolution of any ischemic changes, the development of pathologic Q waves (width LO.04 s, depth 225% of the R wave) or symmetrical T-wave inversion. Total creatinine phosphokinase levels > 1,000, absolute creatinine phosphokinase-MB > 133 and serum glutamic-oxaloacetic transaminase levels > 100 III/liter were considered the threshold levels suggesting myocardial damage. l l,i2 StatIsties: A comparison between enzyme level data was made with nonpaired Student’s t test. Comparisons between frequency of myocardial infarction by enzyme and electrocardiographic criteria were made with Fisher’s exact test. RESULTS Ektrocardiograms:
Of the 224 patients undergoing mitral valve repair 12 (5.4%) had evidence of AM1 on the postoperative electrocardiogram. All 12 of these patients initially had ST-segment elevation involving the inferior leads (II, III, aVF). Seven (58%) of the 12 patients (3.1% of 224) had new pathologic Q waves before discharge from the hospital; 2 of these patients had new Q waves on their initial postoperative electrocardiogram. Two of the 12 patients had symmetrical T-wave inversion develop, consistent with a non-Q-wave AMI. Three of the 12 patients resolved their ST-segment ele888
THE AMERICAN
JOURNAL
OF CARDIOLOGY
VOLUME
66
vation and upon hospital discharge had no electrocardiographic evidence of AMI. The extent of the initial ST-segment elevation had no bearing on whether the electrocardiogram would resolve back to baseline level or evolve to an infarct pattern. Of the 111 patients who had mitral valve replacement 2 (1.8%) had electrocardiographic evidence of perioperative myocardial ischemia. The location of the ischemia was in the inferior wall in both. Both patients with ischemic ST-segment changes had new pathologic Q waves evolve in the inferior leads, which persisted to the time of discharge. None of the 14 patients who had perioperative ischemia develop had clinical symptoms consistent with angina pectoris before surgery. Eleven of the 14 (79%) patients had coronary arteriography preoperatively and none had any significant coronary narrowing. The 3 remaining patients did not have preoperative catheterization, but they were under 50-years-old and had normal maximum exercise stress tests using electrocardiographic criteria alone (2 patients) or using electrocardiographic and thallium-201 imaging (1 patient). Cardiac enzymes: Of the 12 patients with electrocardiographic evidence of &hernia after valve repair, enzymatic evidence for myocardial infarction was found on the basis of total creatinine phosphokinase levels in 11 patients (92%), on the basis of creatinine phosphokinase-MB levels in 11 patients (92%) and on the basis of serum glutamic-oxaloacetic transaminase levels in 9 patients (75%). Patients who had new pathologic Q waves develop on their electrocardiogram had a trend toward lower maximum values for both creatinine phosphokinase (mean 1,594 III/liter) and serum glutamiooxaloacetic transaminase (mean 150 II-J/liter) than the patients whose electrocardiogram failed to develop pathologic Q waves (1,826 and 222 IU/liter) (difference not significant). Both mitral valve replacement patients with electrocardiographic ischemia had creatinine phosphokinase levels
