Aortic Valve Replacement for Aortic Stenosis With Severe Left Ventricular Dysfunction
Background Aortic valve replacement for aortic stenosis (AS) carries an increased risk in the presence of left ventricular (LV) systolic dysfunction. Few data are available on the outcome of such patients.
Methods and Results Between 1985 and 1992, 154 consecutive patients (107 men and 47 women) with LV systolic dysfunction (ejection fraction [EF] ≤35%) underwent aortic valve replacement for AS. The mean preoperative characteristics included EF, 27±6%; aortic valve mean gradient, 44±18 mm Hg; aortic valve area, 0.6±0.2 cm2; and cardiac output, 4.1±1.5 L/min. Simultaneous coronary artery bypass graft surgery was performed in 78 patients (51%). Perioperative (30-day) mortality was 9% (14 of 154 patients). Fifty patients died during follow-up. Coronary artery disease (P=.002) and a reduced preoperative cardiac output (P=.03) were significantly related to reduced overall survival rate by multivariate analysis. Postoperative improvement occurred in most patients; 88% were New York Heart Association class III or IV before surgery versus 7% after surgery. Postoperative EF was assessed in 76% of survivors; 76% of these demonstrated improvement. By multivariate analysis, change in EF was inversely related to coronary disease (P=.002) and preoperative aortic valve area (P=.03).
Conclusions Despite LV dysfunction, the risk of aortic valve replacement for AS was acceptable and related to coronary artery disease and mean aortic gradient, and long-term survival was related to coronary disease and cardiac output. Improvement in symptoms and EF occurred in most patients.
Severe aortic stenosis carries a dismal prognosis when associated with congestive heart failure, with an average life expectancy of <2 years if not operated on.1 Aortic valve replacement is the only effective treatment, but the operative risk2 increases in the presence of LV systolic dysfunction.3 4 5 6 LV dysfunction may be related to critical narrowing of the aortic valve (“afterload mismatch”) without permanent myocardial damage, which often improves after successful aortic valve replacement or because of a superimposed myocardial process such as fibrosis. In the latter situation, the LV dysfunction may not improve after aortic valve replacement.
Few data are available on the clinical outcome of patients with severe aortic stenosis with severely decreased LV systolic function who undergo aortic valve replacement. Predictors of postoperative survival and global systolic function after aortic valve replacement have not been identified in a large group of patients. Therefore, with this study, we sought to test the hypothesis that aortic valve replacement in patients with aortic stenosis and LV systolic dysfunction can be performed at an acceptable operative and long-term risk.
From the Mayo Clinic medical and surgical databases, we identified all patients who had aortic valve replacement for isolated native aortic stenosis in the presence of severe LV systolic dysfunction (EF ≤35%) between 1985 and 1992. CAD was not an exclusion factor. Preoperative EF was assessed by echocardiography in 134 patients, by biplane left ventriculography in 19 patients, and by radionuclide angiography in 1 patient. Patients were excluded if they had concomitant valvular operations other than aortic valve replacement, a previous aortic valve replacement, or predominant aortic regurgitation or were <18 years old. The medical records of 154 consecutive patients (107 men and 47 women) who fulfilled the entry criteria for the study were reviewed, including preoperative clinical data, 2D Doppler echocardiographic results, cardiac catheterization hemodynamics and coronary artery anatomy, and operative data (Table 1⇓).
Comprehensive 2D and Doppler echocardiographic assessment was performed at the Mayo Clinic in 141 patients (92%) ≤30 days before aortic valve replacement. 2D echocardiograms were recorded in multiple tomographic planes, as previously described.7 The EF was calculated by 2D echocardiography with a modification of the method of Quiñones et al8 9 in 43 cases. When these measurements were inadequate, a visual estimate of EF was made10 ; this was performed in 91 patients. The M-mode measurements were guided by 2D echocardiography. The LV diameters, EF, and wall thicknesses were measured.11 End-diastolic diameter could be measured in 89 patients.
Aortic valve hemodynamics were assessed by Doppler echocardiographic examination.12 The LV outflow tract area was calculated from the diameter of the outflow tract (area=diameter2×0.785), assuming a circular geometry. The velocity of the LV outflow tract was obtained by pulsed-wave Doppler echocardiography from the apical long-axis view, and the maximal instantaneous (aortic valve) gradient was calculated from the peak aortic Doppler velocity by the modified Bernoulli equation (pressure gradient=4×velocity2). With on-line software, mean aortic pressure gradient and time velocity integral of the aortic and LV outflow tract flow velocities were measured. Three to five cardiac cycles were measured, and the values were averaged. Aortic valve area was calculated with the continuity equation: AVA=(LVOT area×LVOT TVI)/aortic TVI, where AVA is aortic valve area, LVOT is LV outflow tract, and TVI is time-velocity integral.
Of the 154 patients, 105 (68%) had preoperative hemodynamic assessment by cardiac catheterization. Hemodynamic data were recorded with fluid-filled manometers. LV and systemic arterial pressures were recorded simultaneously. Cardiac output and index were measured at the time of cardiac catheterization, and aortic valve area was calculated from the Gorlin equation.13 Mean gradient was measured in all but 20 patients, and in 13 of the 20 patients, the peak-to-peak gradient was used. Forty-nine patients (32%) had aortic valve replacement based solely on Doppler hemodynamics and without hemodynamic cardiac catheterization. Cardiac output was measured in 148 patients at our institution: cardiac output was calculated by echocardiography14 in 38 patients, by the Fick method with oxygen consumption in 73, by the thermodilution technique in 35, and by the dye technique in 2.
Selective coronary angiography was performed before surgery in 149 patients (97%) (mean age, 73±10 years). Three patients were operated on emergently without coronary angiography, and 1 other patient did not undergo angiography because of young age (32 years) at the time of the operation. Coronary artery stenosis was defined as a luminal diameter narrowing ≥70% in one of the major epicardial coronary arteries or ≥50% luminal diameter narrowing in the left main coronary artery.
When more than one method was used to define preoperative EF, mean pressure gradient, or aortic valve area, the echocardiographic Doppler method was used if performed up to 30 days before aortic valve replacement.
Preoperative aortic valve hemodynamics were assessed by Doppler echocardiography in 133 patients, and aortic valve area was calculated by hemodynamic cardiac catheterization in 105 patients. Cardiac output was measured at the time of cardiac catheterization in 101 patients.
All surgical records were reviewed to determine the type and size of aortic valve prosthesis and whether coronary artery bypass graft surgery or aortic root enlargement was performed concomitantly with aortic valve replacement. The surgical procedures performed, valve types used, and cross-clamp and cardiopulmonary bypass times are outlined in Table 1⇑.
Categorical variables were compared between groups by the χ2 test for independence. Continuous variables were compared by the two-sample t test. Group data were summarized by mean and SD or by frequency percentages. The relationship, univariate and multivariate, of potential risk factors with operative mortality (ie, death within 30 days of operation) was assessed by logistic regression. Overall survival was estimated by the Kaplan-Meier method, and the predictors were analyzed by Cox proportional-hazards model. The relationship of preoperative variables to postoperative EF was assessed by simple and multiple linear regression.
The 30-day mortality was 9% (14 of the 154 patients). Univariate analysis of risk factors for operative mortality identified two significant preoperative variables associated with perioperative mortality. These included a decreased preoperative mean gradient (P=.009) and a history of prior myocardial infarction (P=.03). The average preoperative mean gradient in survivors was 45±18 mm Hg, compared with 35±18 mm Hg in patients who died perioperatively. Multivariate analysis identified significant CAD (two-vessel disease or greater or left main CAD) as the sole independent predictor of 30-day mortality (P=.01). Patients with significant CAD were 4.6 times more at risk for 30-day mortality (95% CI, 1.4 to 15.6). After adjustment for this variable, no other factors were significant.
In addition to the 14 patients who died within 30 days after surgery, 36 others died during a median overall follow-up period of 1.2 years (up to 8.2 years). Two patients were lost to follow-up. Of the 36 late deaths, 11 (31%) were from noncardiac causes. Fig 1⇓ shows the Kaplan-Meier survival curve of this patient population. The survival of patients undergoing aortic valve replacement for aortic stenosis with reduced LV function in the absence of significant CAD was similar to the expected survival in the overall population (5-year survival, 69% [95% CI, 58% to 82%] versus 77%, respectively). By univariate analysis, the presence of significant CAD (P=.03), previous myocardial infarction (P=.02), and a decreased preoperative cardiac output (P=.005) were associated with reduced overall survival. Only the presence of significant CAD (P=.03) and a lower preoperative cardiac output (P=.02) remained significantly related to a reduced overall survival by multivariate analysis. The mean cardiac output for late survivors was 4.4±1.7 L/min compared with 3.5±1.0 L/min for patients who died late after surgery. The overall 5-year survival was 58% (95% CI, 48% to 70%; SEM, 6%), but 5-year survival in patients without significant CAD was much better than that for patients with significant CAD (69% and 39% [95% CI, 24% to 63%; SEM, 10%], respectively; P=.02).
The year and urgency of operation and prosthesis size did not significantly affect survival in this series. Preoperative mean aortic gradient, which was an important prognostic variable for 30-day mortality, was not significantly related to overall survival.
Symptomatic improvement was noted in most of the patients who survived aortic valve replacement. A total of 106 patients had functional status noted both before and after surgery. Of these, 89% were severely symptomatic (NYHA class III or IV) before surgery and only 7% were severely symptomatic after surgery. With respect to the NYHA functional classification, the condition of 66% of the patients had improved by two or more classes and the condition of 88% had improved by at least one class at follow-up (25 patients improved three classes; 45 patients improved two classes; 23 patients improved one class; 10 patients had no change; and in 3 patients, functional class worsened after surgery).
EF was assessed with echocardiography in 76% of 30-day survivors (107 of 140 patients) at a mean interval of 14 months after surgery without knowledge of this study. Of the patients in whom LVEF was assessed after surgery, 76% showed a positive change in the EF. The mean change was an increase of 12±14 EF units (P<.001) (mean preoperative EF, 27±6%; mean postoperative EF, 39±14%) (Fig 2⇓). A change in EF from before to after surgery was positively associated with a higher preoperative peak aortic systolic velocity (P=.008) and a higher preoperative mean aortic valve gradient (P=.02). Also, preoperative to postoperative change in EF was inversely related to the extent of CAD (P=.002), male sex (P=.006), and aortic valve area (P=.009) by univariate analysis. By multivariate analysis, change in EF was related to less CAD (P=.002) and smaller aortic valve area (P=.03).
LV dysfunction is a major prognostic indicator of the outcome of patients undergoing aortic valve replacement for aortic stenosis15 16 ; however, the outcome of patients with LV dysfunction who undergo aortic valve replacement is poorly characterized. Thus, we undertook to further stratify risk in this high-risk group of patients.
The paucity of available data on the outcome of aortic valve replacement in patients with severe aortic stenosis and decreased LV systolic function led us to review 154 such patients in an attempt to determine perioperative mortality, long-term survival, and predictors of outcome. By multivariate analysis, we found that 30-day mortality was related only to the presence of significant CAD and that late mortality was also related to the presence of significant CAD in addition to preoperative cardiac output. All patients had reduced LV function; therefore, further EF analysis alone was not related to survival. Improved postoperative EF was related to a lesser extent of CAD and to a smaller aortic valve area. These findings are similar to those in previous series; however, there are no reports on surgical or late outcome in a large group of patients with aortic stenosis and substantial LV dysfunction. Many of the reported series have included patients undergoing aortic valve replacement for both aortic regurgitation and aortic stenosis.2 15 16 17 18 Furthermore, the reported surgical series are largely older series, and surgical techniques, particularly regarding myocardial preservation, have advanced since that time.
In severe aortic stenosis, the left ventricle compensates for the chronic pressure overload by hypertrophy in an attempt to normalize wall stress. Initially, EF and cardiac output are maintained. When wall stress exceeds the compensating mechanism, LV function declines. Thus, when LV dysfunction is due to “afterload mismatch,”4 as seen in severe aortic stenosis, aortic valve replacement will result in the improvement of EF and symptoms.
In the present study, substantial CAD was associated with increased surgical mortality. Of the 14 patients who died perioperatively, 11 had CAD and all had simultaneous coronary artery bypass graft surgery. Only 1 of these 14 patients had previously had coronary artery bypass surgery. The extent of CAD and preoperative cardiac output were important determinants of long-term survival by both univariate and multivariate analyses. The mean preoperative cardiac output in late survivors was 4.4±1.7 L/min, compared with 3.5±1.0 L/min in patients who died late after surgery. The impact of CAD can be partially explained by the fact that 28 of 59 patients with two-vessel or greater CAD had a history of a previous myocardial infarction, which had an impact on cardiac output and EF. Still, cardiac output was an important independent predictor of overall survival.
The type or size of prosthesis was not related to operative or late survival, as previously demonstrated by Morris et al,16 and may be related to the relatively large mean size (23±2 mm) of prosthesis used in this population. In addition, 14% of patients had simultaneous aortic root enlargement to allow placement of a larger prosthesis.
Effect of Aortic Valve Replacement on Postoperative EF
Aortic valve replacement for aortic stenosis decreases ventricular afterload19 ; subsequent changes include adaptation and remodeling, with regression of hypertrophy and LV mass.5 20 EF, therefore, would be expected to improve after aortic valve replacement in patients with reduced preoperative EF.5 6 19 20 Those who do not improve probably have fixed myocardial damage. Previous studies have shown that decreased preoperative EF, previous myocardial infarction, and low preoperative aortic valve gradient are associated with reduced postoperative EF.21
An improvement in postoperative LVEF of 12 EF units (Fig 2⇑) was noted in our study. The relationship of important CAD to lack of improvement in EF after aortic valve replacement has been described.2 16 It is related to increased risk of perioperative myocardial infarction and an increased incidence of irreversibly damaged myocardium or scar before aortic valve replacement.
Sex differences in LV adaptation to aortic stenosis have been described recently.22 23 24 Morris et al16 similarly found that there is a substantial sex-associated difference in regression of LV adaptation to chronic pressure and volume overload. The factors accounting for these sex differences have not been characterized. The sex difference in our series (improved postoperative EF in women) was an isolated sex-related finding, ie, EF and mean gradient were not significantly different between men and women before surgery.
A higher preoperative peak aortic velocity and mean gradient and a smaller preoperative aortic valve area were associated with change in EF. These factors relate more to cardiac output and to the capability of the myocardium to generate a substantial pressure gradient across the stenotic aortic valve. Therefore, they can be expected to predict improved LV systolic function.
Follow-up. The postoperative EF was determined for 76% of surgical survivors (107 of 140 patients). This should be considered in the interpretation of the results, since performing echocardiography may have been a proxy for better outcome. Comparison of baseline characteristics between survivors with and without determination of postoperative EF indicated no differences in proportion of female patients, preoperative NYHA functional class, age, or percentage of patients with CAD in this assessment.
EF analysis. Echocardiographic estimates of LVEF may be a cause of concern; the use of echocardiography in this clinical setting, however, is standard clinical practice. Previous studies from our institution and others have documented acceptable correlations to angiography25 and have confirmed reproducibility.26 27 28
Lack of control group. All the patients included in this study underwent aortic valve replacement; thus, the clinical dilemma differentiating LV systolic dysfunction due to aortic stenosis from moderate aortic stenosis with a coexistent cardiomyopathic process cannot be addressed.
Cardiac output determination. It is well known that cardiac output determination may be less reliable when measured by the thermodilution technique, particularly in low-output states. In our series, however, only 15 of 149 patients in whom cardiac output was measured (10%) had measurement by thermodilution, so that this technique should not have adversely affected our data.
In summary, despite decreased LVEF, advanced age, and increased incidence of CAD in our patient population, surgical mortality was acceptable and related to cardiac output and mean aortic gradient. The early and intermediate follow-up results are acceptable in comparison with those of age- and sex-matched controls and are related to the presence of CAD and cardiac output status. Marked improvement in symptoms and LVEF occurred in most patients (>70%). Improved postoperative EF was influenced considerably by a lesser extent of CAD, higher preoperative mean gradient and peak aortic valve velocity, a smaller aortic valve area, and female sex.
Selected Abbreviations and Acronyms
|CAD||=||coronary artery disease|
|NYHA||=||New York Heart Association|
Reprint requests to Dr Heidi M. Connolly, Division of Cardiovascular Diseases and Internal Medicine, Mayo Clinic, 200 First St SW, Rochester, MN 55905.
- Received September 9, 1996.
- Revision received December 2, 1996.
- Accepted December 14, 1996.
- Copyright © 1997 by American Heart Association
Ross J Jr, Braunwald E. Aortic stenosis. Circulation. 1968;38(suppl V):V-61-V-67.
Carabello BA, Green LH, Grossman W, Cohn LH, Koster JK, Collins JJ Jr. Hemodynamic determinants of prognosis of aortic valve replacement in critical aortic stenosis and advanced congestive heart failure. Circulation. 1980;62:42-48.
Kennedy JW, Doces J, Stewart DK. Left ventricular function before and following aortic valve replacement. Circulation. 1977;56:944-950.
Smith N, McAnulty JH, Rahimtoola SH. Severe aortic stenosis with impaired left ventricular function and clinical heart failure: results of valve replacement. Circulation. 1978;58:255-264.
Quiñones MA, Waggoner AD, Reduto LA, Nelson JG, Young JB, Winters WL Jr, Ribeiro LG, Miller RR. A new, simplified and accurate method for determining ejection fraction with two-dimensional echocardiography. Circulation. 1981;64:744-753.
Lewis JF, Kuo LC, Nelson JG, Limacher MC, Quiñones MA. Pulsed Doppler echocardiographic determination of stroke volume and cardiac output: clinical validation of two new methods using the apical window. Circulation. 1984;70:425-431.
Hwang MH, Hammermeister KE, Oprian C, Henderson W, Bousvaros G, Wong M, Miller DC, Folland E, Sethi G. Preoperative identification of patients likely to have left ventricular dysfunction after aortic valve replacement: participants in the Veterans Administration Cooperative Study on Valvular Heart Disease. Circulation. 1989;80(suppl I):I-65-I-76.
Villari B, Hess OM, Campbell SE, Krayenbuehl HP. Sex-dependent response of the left ventricle to aortic valve stenosis: influence of the collagen network. Circulation. 1992;86(suppl I):I-538. Abstract.
Carroll JD, Carroll EP, Feldman T, Ward DM, Lang RM, McGaughey D, Karp RB. Sex-associated differences in left ventricular function in aortic stenosis of the elderly. Circulation. 1992;86:1099-1107.
Aurigemma GP, Silver KH, McLaughlin M, Orsinelli D, Sweeney AM, Gaasch WH. Gender influences the pattern of left ventricular hypertrophy in elderly patients with aortic stenosis. Circulation. 1992;86(suppl I):I-538. Abstract.
Enriquez-Sarano M, Tajik AJ, Schaff HV, Orszulak TA, Bailey KR, Frye RL. Echocardiographic prediction of survival after surgical correction of organic mitral regurgitation. Circulation. 1994;90:830-837.