Reversal of Advanced Left Ventricular Dysfunction Following Aortic Valve Replacement for Aortic Stenosis

Transcription

1 Reversal of Advanced Left Ventricular Dysfunction Following Aortic Valve Replacement for Aortic Stenosis Robert P. Croke, M.D., Roque Pifarre, M.D., Henry Sullivan, M.D., Rolf Gunnar, M.D., and Henry Loeb, M.D. ABSTRACT A series of 12 consecutive patients who underwent aortic valve replacement (AVR) for aortic stenosis complicated by severe left ventricular dysfunction was reviewed. Ventricular dysfunction was reflected by pulmonary congestion, edema, renal and hepatic dysfunction, and by severely depressed ejection fractions (mean, 13%; range=o-z(wo). Aortic valve replacement was accompanied by mitral commissurotomy in 1 patient and aortocoronary bypass in 5. Three of 5 patients with greater than 50% coronary obstruction died without reversal of heart failure, and 1 of the 5 died after a stroke. The 1 survivor of this group has done well. All 7 patients with minimal or no coronary disease survived operation and are now in New York Heart Association Class I or 11. Postoperative catheterization (2 to 12 months) in 6 patients showed improved cardiac index and filling pressures. Left ventricular diastolic volume fell from 159 to 82 ml/m2, and ejection fraction rose from 13 to 45%. We conclude that left ventricular dysfunction owing to aortic stenosis alone is reversible and that AVR results in great clinical improvement. When coronary disease is present, survival may be accompanied by great improvement but the operative mortality is much higher. In several studies of the mortality and results of cardiac valve surgery [7, 12, 151, preoperative left ventricular function has been emphasized as an important prognostic factor, with a depressed ejection fraction signaling a higher operative mortality and less satisfactory results in survivors. Since it might be supposed that ventricular function could deteriorate to such an From the Divisions of Cardiothoracic Surgery and Cardiology, Departments of Surgery and Medicine, Loyola University School of Medicine, Maywood, and the Veterans Administration Hospital, Hines, IL. Presented at the Thirteenth Annual Meeting of The Society of Thoracic Surgeons, Jan 24-26, 1977, San Francisco, CA. Address reprint requests to Dr. Croke, Hines Veterans Administration Hospital, Hines, IL extent that operative mortality would be prohibitively high, or the results universally disappointing, we have undertaken a review of our experience with replacement of stenotic aortic valves in the setting of advanced left ventricular dysfunction. Methods The records of all patients who underwent aortic valve replacement (AVR) for aortic stenosis at Loyola University Hospital and the Hines Veterans Administration Hospital from July 1, 1974, through June 30, 1976, were reviewed. The 12 patients, each of whom had markedly depressed left ventricular ejection fractions of 20% or less were selected for detailed analysis. All had undergone preoperative right and left cardiac catheterization with standard techniques. Preoperative cardiac output was measured by the Fick or indicator-dilution technique or both. Left ventriculography was performed by injection directly into the left ventricle in the right anterior oblique projection. Ventricular volumes were measured by the method of Kasser and Kennedy [14]. All patients had preoperative selective coronary arteriography by the Sones [19] or Judkins 1131 technique. In all cases AVR was done with the Bjork-Shiley tilting-disc prosthesis. Moderate systemic hypothermia (26 C) was used, and profound local hypothennia (2" to 4 C) was established by continuous bathing of endocardia1 and epicardial surfaces. In the 5 patients who underwent aortocoronary bypass (1 vessel in 3 patients, 2 vessels in 2 patients), the proximal anastomoses were completed before cardiopulmonary bypass was begun and the distal anastomoses before the aortic valve was replaced. Aortic crossclamping was done proximal to the graft insertion sites so as to allow coronary perfusion. Both coronary orifices were cannulated and perfused 38

2 39 Croke et al: Left Ventricular Dysfunction after Aortic Valve Replacement directly from the arterial line (26 C) in all 12 patients. Ventricular fibrillation occurred during induction of hypothermia or by electrical impulse. Mitral commissurotomy was performed in 1 patient. In another the intraortic balloon pump (IABP) was inserted before induction of anesthesia; in 3 patients IABP was required during or shortly after operation. Postoperative follow-up (at 10 to 27 months for survivors) was conducted by one of the authors in all cases. At postoperative right heart catheterization (6 patients), cardiac output was measured by thermodilution [9] and left ventricular function was analyzed using the levophase of a pulmonary cinearteriogram. During follow-up the Bruce treadmill method [5] was used to test exercise tolerance in 7 patients. Patient Population The patients were divided into two groups. Group 1 included 5 patients whose preoperative angiograms showed at least one major coronary obstruction (> 50%). Each had aortocoronary bypass in addition to AVR. Group 2 comprised 7 patients: 5 with normal coronary arteriograms and 2 with lesions of less than 50%. No Group 2 patient received a bypass graft. Since the two groups were quite similar in most other preoperative characteristics (Table l), the 12 patients are described together. Mean age of the 11 men and 1 woman was 59 years. Preoperatively, 7 were in New York Heart Association Class IV and 5 were in Class III. Dyspnea was a prominent symptom, and all patients showed marked cardiomegaly and pulmonary congestion by physical and radiographic examination. One patient had overt pulmonary edema; none was in shock. Most had signs of right ventricular failure, which included jugular venous congestion, pedal edema, and hepatomegaly. Two patients had Cheyne- Stokes respiratory patterns. Nine had azotemia (BUN 7 25 mg/loo ml) and 6 had elevation of serum bilirubin (51.5 mg1100 ml). In several patients the murmur and pulse contour typical of aortic stenosis were not apparent, probably because of the very depressed stroke output. In these patients a diagnosis of aortic stenosis was not seriously entertained at initial evaluation. Detection of aortic valve calcification or thickening by fluoroscopy (11 of 12) or echocardiography (9 of 9) was a useful indicator of aortic valve disease. Electrocardiographic abnormalities included atrial fibrillation (3 patients), left bundle-branch block (2 patients), and inferior myocardial infarction (1 patient). Hemodynamic and angiographic data obtained preoperatively are summarized in Table 2. Car- diac index was reduced (< 2.2 L/min/m2) in 10 of the 12 patients. Left ventricular end-diastolic volume index (LVEDVI) was increased in 11 of the 12 patients (average, 189 ml/m2). Although LVEDVI tended to be higher in Group 1 patients (258 ml/m2 versus 141 ml/m2 in Group 2), this difference was not statistically significant. Ejection fractions were similarly low in both groups. None exceeded 20%, and the average was 13%. In keeping with very depressed stroke volumes, Table 1. Preoperative Clinical Characteristics of 12 Patients with Aortic Stenosis and Severe Heart Failure Mean (Range)a Characteristic Group 1 Group 2 Age (Yr) Duration of symptoms (mo) Functional Class I11 (no. of patients) Functional Class IV (no. of patients) Bilirubin (mg1100 ml) 61 (52-68) 30 (7-36) ( ) 57 (45-68) 29 (2-96) ( ) Creatinine (mg1100 ml) 1.5 ( ) 1.5 ( ) BUN (mg1100 ml) 30 (15-51) 31 (24-50) The differences between Groups 1 and 2 are not statistically significant by the unpaired t-test. BUN = blood urea nitrogen.

3 40 The Annals of Thoracic Surgery Vol 24 No 1 July 1977 Table 2. Preoperative Hemodynamic Characteristics of 12 Patients with Aortic Stenosis and Severe Heart Failure Mean (Range)a Characteristic Group 1 Group 2 Cardiac index (L/min/m2) 1.9 ( ) 1.8 ( ) LVEDVI (mum') 258 ( ) 141 (82-257) LV ejection fraction (%) 14 (12-16) 13 (-0-20) Peak-to-peak gradient (mm Hg) 44 (25-75) 65 (32-92) Mean gradient (mm Hg) 42 (16-55) 48 (29-70) Aortic orifice index (cmz/m2) 0.3 ( ) 0.3 ( ) *The differences between Groups 1 and 2 are not statistically significant by the unpaired t-test. LVEDVI = left ventricular end-diastolic volume index; LV = left ventricular. the transaortic gradients were not high. The mean peak-to-peak gradient was 56 mm Hg, and 6 patients had peak-to-peak gradients of 50 mm Hg or less. The calculated valve orifice areas were very small (average, 0.3 cdld; range, cd/d). Additional diagnoses of some importance included hypertension in 2 patients and mitral stenosis in l. Among the 5 Group 1 patients with significant coronary artery disease, 2 had left main coronary occlusion of 70 to 80% in addition to obstructions elsewhere. The other 3 had single, double, and triple vessel disease, respectively. Three of these Group 1 patients also had areas of discrete ventricular akinesia in addition to diffuse hypokinesia. Eight of the 12 patients had mild to moderate mitral regurgitation, which was attributed to left ventricular dysfunction. None required mitral valve replacement. Aortic regurgitation, when present, was never more than trivial. Results Four deaths have occurred, all among the patients with coronary disease. Three of these appear to have been attributable to inadequate reversal of left ventricular dysfunction. One man could not be weaned from cardiopulmonary bypass despite the aid of the intraaortic balloon pump; he died in the operating room. Another man, who had left main coronary artery disease, was able to leave the operating room with the IABP in place but remained hemodynamically unstable with low cardiac output syndrome; he died after 23 days. The only woman in the group died 43 days postoperatively with Escherichia coli pneumonia, but she had never thrived nor had she been rid of radiographic evidence of pulmonary edema. The fourth patient did well initially, with disappearance of pulmonary congestion and gratifying progress in ambulation. This course, however, was interrupted by a devastating stroke which led to the patient's death 66 days postoperatively. The 8 surviving patients, including all 7 in Group 2, improved quickly and dramatically with respect to heart failure and denied limitation of activity 10 to 32 months postoperatively. Objective testing of exercise tolerance by treadmill, using the Bruce protocol, has been performed in 7 of these survivors. The average duration of exercise was 6.9 minutes (range, 4-9 min). Functional aerobic capacity, an age- and sex-adjusted expression of exercise tolerance [5], ranged from 50 to 105% and averaged 83% of normal. Postoperative (2 to 12 months) right heart catheterization has been performed in 6 survivors including the Group 1 patient (Table 3). Cardiac index had returned to normal in all 6, and mean filling pressures for the right and left ventricles were in or near the normal range. The most remarkable changes took place in left ventricular volumes and ejection fractions. Postoperatively, left ventricular size at end-diastole was only half the preoperative value, and ejection fraction had increased from the preoperative average of 13% to apostoperative average of 45 '/o.

4 41 Croke et al: Left Ventricular Dysfunction after Aortic Valve Replacement Table 3. Hernodynamic Values in 6 Survivors before and after Aortic Valve Replacement Mean Value Preoperative Postoperative Cardiac index (L/min/m2) 1.9 ( ) 3.3 ( ) Right ventricular filling pressure 11 (4-30) 4 (1-7) (RVEDP) (mm Hg) Left Ventricular filling pressure 29 (14-40) 10 (5-14) (LVEDP) (mm Hg) LVEDVI (ml/m*) 159 ( ) 82 (46-132) Ejection fraction ('3'0) 13 (20-20) 45 (34-75) adifferences between preoperative and postoperative values are significant at p < by the paired t-test. RVEDP = right ventricular end-diastolic pressure; LVEDP = left ventricular end-diastolic pressure; LVEDVI = left ventricular end-diastolic volume index. Comment Several studies [7,10,12,15] of the risk determinants in cardiac operations have emphasized the association of left ventricular dysfunction with higher operative mortality and with continued severe disability in survivors. These observations have been particularly clear in coronary bypass procedures, in which advanced left ventricular dysfunction, indicated by such factors as history of heart failure, elevated enddiastolic volume or pressure (or both), or depressed ejection fraction, has emerged as a major contraindication to surgical intervention in the absence of a correctable mechanical lesion. For valve replacements, a similar relationship between ventricular dysfunction and surgisal outcome has been suggested. Thus Tsuchioka and associates [21], studying the influence of left ventricular function on the postoperative course of 32 patients who had received prosthetic valves, found that preoperative depression of stroke work index in relation to end-diastolic pressure marked patients who were unlikely to survive operation. Two studies [4, 121 have identified radiographic cardiomegaly (cardiothoracic ratio > 0.50) as a predictor of high operative mortality. The ejection fraction, which offers several theoretical and empirical advantages [7] over the cardiac index and end-diastolic filling pressure, was found to be more sensitive than either or both of these in indicating an elevated likelihood of operative death or unsatisfactory results of valve operation as well as myocardial revascularization [7]. De- spite these observations, the role of left ventricular function in the selection of patients for valve procedures has not been as clear as in coronary bypass operations. This is probably because of the heterogeneity of valve lesions, because the stenotic or regurgitant valve constitutes a correctable impediment to ventricular function and because a few patients have experienced gratifying operative results despite advanced circulatory deterioration in the form of pulmonary edema or shock [ll, 16,201. Operations combining aortic valve replacement with aortocoronary bypass grafts have been the subject of a number of reports [l-3,6,8,17,181, most of which show a higher operative mortality in patients who received the combined operation than in those who underwent either procedure individually. There has been the usual tendency for operative risk to fall in the more recent series. Although detailed data regarding left ventricular function are not available, at least two reports [6, 181 mention advanced heart failure as a prominent feature in patients who died during or shortly after combined operations. Virtually all patients who survive the operation are said to experience gratifying improvement in symptoms. The present series is of value in examining the role of surgical intervention in a specific group of patients with aortic stenosis accompanied by advanced left ventricular failure. It is a consecutive series over a recent 2-year period and includes coronary arteriography, preoperative and postoperative quantitative left ventriculography, and confirmation of postopera-

5 42 The Annals of Thoracic Surgery Vol 24 No 1 July 1977 tive functional status by objective techniques. Of the 12 patients operated on, all,8 long-term survivors show remarkable improvement in ventricular performance. Each of the 4 deaths occurred in patients with significant coronary disease and 3 of these were attributed to persisting ventricular dysfunction. It seems that reversibility of left ventricular dysfunction associated with aortic stenosis may be dependent upon the cause. When aortic stenosis alone is responsible, left ventricular dysfunction, no matter how severe, will improve following successful relief of aortic obstruction. On the other hand, left ventricular dysfunction that results from major coronary artery disease is unlikely to be reversed following aortic valve replacement and may in fact worsen as a result of complications incurred during operation. In such patients a high operative mortality should be expected, with improvement in survivors unpredictable. In summary, patients with isolated aortic stenosis can be expected to experience marked reversal of left ventricular dysfunction and concomitant improvement in functional capacity even when preoperative heart failure is severe. Obstructive coronary disease raises the operative risk by introducing an element of irreversibility with respect to ventricular function. Management of the latter patients would be aided greatly if there were an accurate means of assessing the reversible component of ventricular dysfunction due to outflow tract obstruction versus the irreversible ventricular damage resulting from coexistent coronary artery disease. At present we know of no clear way of making this distinction. References Barnhorst DA, Giuliani ER, Pluth JR, et al: Open heart surgery in patients more than 65 years old. Ann Thorac Surg 18:81, 1974 Beiger TJ, Karp RB, Kouchoukos NT: Valve replacement and myocardial revascularization: results of combined operation in 59 patients. Circulation 51,52:Suppl 1:126, 1975 Berndt TB, Hancock EW, Shumway NE, et al: Aortic valve replacement with and without coronary by-pass surgery. Circulation 50:967, 1974 Braun LO, Kincaid OW, McGoon DW: Prognosis of aortic valve replacement in relation to preoperative heart size. J Thorac Cardiovasc Surg 65:381, 1973 Bruce RA, Kusumi F, Hosmer D: Maximal 0 intake and nomographic assessment of functional aerobic impairment in cardiovascular disease. Am Heart J 85:546, 1973 Callard GM, Flege JB, Todd JC: Combined valvular and coronary artery surgery. Ann Thorac Surg 22:338, 1976 Cohn PF, Gorlin R, Cohn LH, et al: Left ventricular ejection fraction as a prognostic guide in surgical treatment of coronary and valvular heart dis- ease. Am J Cardiol34:136, 1974 Cooley DA, Dawson JT, Hallman GL, et al: Aor- tocoronary saphenous vein bypass: results in 1,492 patients, with particular reference to patients with complicating features. Ann Thorac Surg 16:380, 1973 Ganz W, Donoso R, Marcus HS, et al: A new technique for the measurement of cardiac output by thermodilution in man. Am J Cardiol 27:392, 1971 Hammermeister KE, Kennedy JW: Predictors of surgical mortality in patients undergoing direct myocardial revascularization. Circulation 49,5O:Suppl 2:112, Hutter AM, DeSanctis RW, Nathan MJ, et al: Aortic valve surgery as an emergency procedure. Circulation 41:623, Isom OW, Dembrow JM, Glassman E, et al: Factors influencing long term survival after isolated aortic valve replacement. Circulation 49,50: Suppl 2:154, Judkins MP: Percutaneous transfemoral selective coronary arteriography. Radiol Clin North Am 6:467, Kasser IS, Kennedy JW: Measurement of left ven- tricular volume in man by single plane cineangiocardiography. Invest Radiol 4:83, Mundth ED, Austen WG: Surgical measures for coronary heart disease. N Engl J Med 293:13, 75, 124, Najafi H, Dye WS, Javid H, et al: Emergency open heart surgery for acquired heart disease. Dis Chest 55:456, Pupello DF, Blank RH, Bessone LN, et al: Local deep hypothermia for combined valvular and coronary heart disease. Ann Thorac Surg 21:508, Rossiter SJ, Hultgren HN, Kosek JC, et al: Isch- emic myocardial injury with aortic valve replacement and coronary by-pass. Arch Surg 109:652, 1974 Sones FM: Cine coronary arteriography, The Heart. Second edition. Edited by JW Hurst, BR Logue. New York, McGraw-Hill, 1970, p 377 Stinson EB, Shumway NE: Emergency heart valve replacement. Calif Med 109:441, 1968

6 43 Croke et al: Left Ventricular Dysfunction after Aortic Valve Replacement 21. Tsuchioka H, Fukuke I, Iyomasa Y: Effects of left ventricular function in patients with acquired cardiac valvular disease on postoperative course of prosthetic valvular replacement. J Cardiovasc Surg (Torino) 15:331, 1974 Discussion DR. WILLARD M. DAGGETT: (Boston, MA): We share the authors enthusiasm for replacing the stenotic aortic valve in the presence of severe biventricular failure even under emergency conditions. I think their results emphasize how serious the combination of valve disease, coronary artery disease, and a bad left ventricle can be. More specifically, the afterloaded heart, as represented by aortic stenosis, is totally unlike the preloaded heart (as indicated by coronary artery disease, aortic insufficiency, or mitral regurgitation) in terms of the potential for surgical reversibility of ventricular dysfunction. Certainly no patient with pure aortic stenosis should be rejected for operation purely on the basis of a low ejection fraction. In the present study intraaortic balloon pumping was used preoperatively in 1 patient, and I wonder whether more patients might have benefited by this maneuver, which Bolooki and his co-workers have recently advocated. I have three brief questions. First, do you think, Dr. Croke, that coronary perfusion in the fibrillating heart, as a form of intraoperative myocardial protection, played any role in the poorer results observed in those patients with coronary artery disease? Second, Patients 1, 4, and 5 had threevessel coronary artery disease and only one or two grafts. Could incomplete revascularization possibly have contributed to poorer results in these patients? Finally, was there any correlation between valve size of the inserted prosthesis and the clinical results? DR. CROKE: I would like to thank Dr. Daggett for his kind words. I am a cardiologist and am not entirely familiar with the ins and outs of operative technique. We at Hines like the balloon pump very much and have used it a great deal for our surgical patients in general. In this series there was 1 patient in whom the balloon pump could not be inserted because of peripheral vascular disease. Both femoral arteries were approached in the patient with coronary disease who died after a stroke. There were 2 other patients, both with coronary disease, who could not have the balloon pump inserted because of peripheral vascular disease. While I think that establishment of circulatory support, especially during the induction of anesthesia when the hemodynamics are so unstable, offers theoretical advantages, patients with coronary disease unfortunately are also inclined to have peripheral vascular disease. Perhaps another method of inserting the balloon pump or of establishing femorofemoral bypass under local anesthesia before induction of general anesthesia would be more appropriate. The incomplete revascularization question is, of course, a good one. There is a time limit on how much can be accomplished when one is taking on an aortic valve replacement and a series of bypasses, and I believe the choice of the operating surgeon was to bypass the vessels most likely to benefit the ventricular function. There was at least 1 patient whose inferior wall was known to be infarcted, and there was no point in bypassing that right coronary vessel. Our study includes only 12 patients. Valve sizes were in the middle range of the Bjork-Shiley group, from 22 to 26 mm, and I could not discern any difference in survival with respect to prosthetic valve size. DR. ROQUE PIFARRE: Ventricular fibrillation will occur when the heart is arrested with cold saline. We believe perfusing the coronary arteries will protect the myocardium, as evidenced by the fact that in many of these patients, when we are warming up the ventricle, the myocardium will start contracting on its own, needing no defibrillation. Many of these patients are being operated on by residents, who are, in many cases, necessarily a little slower than the attending physicians. Therefore, I think any protection that can be afforded to the heart, especially at the beginning until perfusion can be instituted, is indicated. We believe in total revascularization. More bypasses were not performed because we thought that it was not possible to do them.