Can the results of contemporary aortic valve replacement be improved?

J

THORAC CARDIOVASC SURG

92:37-46, 1986

Can the results of contemporary aortic valve replacement be improved? Although the results of contemporary aortic valve replacement are excellent, cardiac surgeons mest identify the factors that predictpostoperative morbidity and mortalityto develop alternative strategiesfor high-risk patients. Two hundred seventy-seven consecutive patients Undergoing isolated aortic valve replacement between 1982 and 1984 were evaluated. Thirty-seven c6nical and 13 preoperative hemodynamic variables were analyzed by univariate and multivariate statistics to determine the risk factors for postoperativ~ morbidity and mortality. The operative mortality was 3 %, the incidence of a postoperative lowoutput syndrome was 12 %, and the incidence of a perioperative myocardial infarction was 5 %• A multivariate: logistic regression analysis found that age was the onlyindependent predictorof mortality. Three factors independently predicted postoperative lowoutput syndrome: age, the presence of coronary artery disease, and the peak systolic gradient in patients with aortic stenosis. Patients with aortic stenosis had a higher incidence of postoperative ventricular dysfunction (17%) than those with mixed valvular disease (9%) or aortic regurgitation (5% ~ Perioperative myocardial infarction was predicted by the extent of coronary artery disease. the incidence of perioperative myocardial infarction was higher in patients with triple-vessel coronary artery disease (13%) and those with left main stenosis (18 %) than in patientswithsingle- or double-vessel disease (4%) or those without coronaryartery disease (4 %~ Because of the higherrisk of aortic valve replacement in older patients, the risk-benefit ratio of the operation must be carefully assessed in the elderly. Improved methods of myocardial protection may reduce the risks for patients with aortic stenosis and symptomatic triple-vessel coronary artery disease.

George T. Christakis, M.D., Richard D. Weisel, M.D., Stephen E. Fremes, M.D., Kevin H. Teoh, M.D., J. Patrick Skalenda, Cathy P. Tong, M.R.L., Jeri Y. Azuma, Leonard Schwartz, M.D., Lynda L. Mickleborough, M.D., Hugh E. Scully, M.D., Bernard S. Goldman, M.D., and Ronald J. Baird, M.D.,

Toronto, Ontario, Canada

h e results of aortic valve replacement have improved dramatically during the past 15 years. Operative mortality was 10% to 33% between 1969 and 1975. 1-6 The mortality for contemporary aortic valve replacement is From the Divisions of Cardiovascular Surgery and Cardiology, the Toronto General Hospital and the University of Toronto, Toronto, Ontario. Canada. Supported by the Heart and Stroke Foundation of Ontario, the Canadian Heart Foundation, and the Medical Research Council of Canada. Received for publication June 24, 1985. Accepted for publication Aug. 29, 1985. Address for reprints: Richard D. Weisel, M.D., Cardiovascular Surgery, Toronto General Hospital, 200 Elizabeth St., Eaton North 13-224, Toronto. Ontario, Canada M5G 2C4.

less than 5%.7,8 The improved surgical results reflect a variety of changes in surgical practice. Noninvasive investigations have improved the timing of surgical referral."!' The use of coronary angiography permits coronary bypass grafting in patients with severe or symptomatic coronary artery disease. Multidose, cold potassium cardioplegia has improved myocardial preservation and the advent of vein graft infusions has permitted more uniform cardioplegic protection." Routine perioperative hemodynamic monitoring and improved anesthetic techniques have also contributed to the decrease in operative mortality." Although the current perioperative risks are low, improvement may be possible, The risk factors for contemporary aortic valve replacement must be identified to reduce perioperative 37

3 8 Christakis et al.

morbidity and mortality. Late survival may be limited in patients who had a perioperative myocardial infarction or who had low output syndrome postoperatively.' Identifying the risk factors for postoperative morbidity may permit strategies to improve the long-term results of aortic valve replacement. This study was designed to evaluate the factors that contribute to contemporary operative morbidity and mortality. Methods Patient population. Two hundred seventy-seven consecutive patients who underwent aortic valve replacement between January, 1982, and June, 1984, were evaluated prospectively. Patients who had concomitant coronary bypass operations were included. Seventythree additional patients who had combined aortic and mitral valve operations and 19 patients who had concomitant aortic valve replacement and ascending aortic repair were excluded from the analysis. All patients underwent preoperative cardiac catheterization including coronary angiography. Thirty-seven clinical and 13 hemodynamic variables were analyzed by univariate and multivariate statistics to assess their influence on postoperative outcome. Clinical presentation. The clinical presentation at the time of hospital admission was recorded prospectively. The New York Heart Association (NYHA) functional class, the major incapacitating symptom, the urgency of the operative procedure, the presence of unstable angina, the preoperative rhythm, and a history of preoperative thromboembolic events or previous valvular procedures were recorded. Sixty patients were mildly symptomatic (NYHA Class lor 11), 168 were in NYHA Class III, and 49 were in NYHA Class IV. Urgent operation (within 48 hours of cardiac catheterization) was performed in 36 patients: four for acute infective endocarditis, six for cardiogenic shock, 10 for unstable angina, four for prosthetic valve dysfunction, and 12 for persisting symptoms at rest because of severe aortic stenosis. Thirty-seven of the 109 patients with coronary artery disease had prolonged episodes of rest pain and/or ischemic electrocardiographic changes with pain during the month before the operation (unstable angina). Forty patients had stable angina and 32 were asymptomatic with coincidental coronary artery disease. A history of persistent atrial fibrillation was found in 36 patients and 241 were in normal sinus rhythm at the time of hospital admission. Thirty-two patients had preoperative thromboembolic events; 23 were reversible and nine resulted in residual deficits. Seventeen patients had had previous aortic valvular operations and two patients had had previous coronary bypass grafting.

The Journal of Thoracic and Cardiovascular Surgery

Cardiac catheterization data. Cardiac catheterization permitted an accurate diagnosis of the aortic valve lesion and an estimate of ventricular function. The left ventricle was entered by either a transseptal or a transvalvular route. The peak systolic aortic valve gradient was calculated either by simultaneous pressure measurements or by the pullback technique (after transvalvular catheterization). Pressure measurements included left ventricular end-diastolic, right atrial, pulmonary arterial, and pulmonary capillary wedge. Cardiac index was calculated by the thermodilution or the Fick methods. Both pulmonary vascular resistance and systemic vascular resistance were calculated by standard formulas. Ejection fraction was estimated from the contrast ventriculogram. The severity of aortic insufficiency was estimated by the extent of left ventricular opacification after aortic root injection (minimal = 1+, mild = 2+, moderate = 3+, severe = 4+). The severity of aortic leaflet and annular calcification was graded on a scale of 1 to 4 (none = grade 1, mild = grade 2, moderate = grade 3, severe = grade 4). One hundred nineteen patients had predominant aortic stenosis with a peak systolicgradient greater than 40 mm Hg, an aortic valve area less than 1 cm-, and no more than minimal (l +) aortic regurgitation. Fifty-six patients had predominant aortic insufficiency with 3+ or greater opacification of the left ventricle after aortic root injection and a pulse pressure equal to or greater than 60% of the systolic aortic pressure. One hundred two patients had mixed aortic valve disease with characteristics of both aortic stenosis and aortic insufficiency at catheterization and at operation. Coronary angiography revealed a greater than 50% (diameter) narrowing of one of three major coronary arteries (left anterior descending, circumflex, and right coronary artery) in 109 patients. Single-vessel disease was found in 44 patients and two-vessel disease was found in 29 patients. Thirty-six had triple-vessel disease, 11 in association with left main coronary artery stenosis and 25 without left main stenosis. Etiology of valve disease. The etiology of the aortic valve disease was determined from the clinical history, cardiac catheterization information, the operative assessment, and the gross and microscopic pathologic evaluation. Bicuspid aortic valves were believed to be congenital in origin in 51 patients. Eight patients had clinical, microbiologic, and pathologic finding consistent with acute or remote infective endocarditis. Thirteen patients had operative and pathologic findings of redundant, myxomatous aortic valves. Idiopathic calcified aortic valve disease was present in 65 patients, and 15 patients underwent reoperation for prosthetic valve

Volume 92 Number 1

Aortic valve replacement 3 9

July. 1986

dysfunction. Rheumatic thickening and fusion of valve leaflets were confirmed by microscopic findings in 125 patients. The high incidence of rheumatic aortic valve disease may reflect the large population of recent immigrants from southern Europe and the Middle East in Toronto. Operative techniques. Surgical techniques were similar among the five operating surgeons. Anesthesia was induced and maintained with fentanyl (75 ~g/kg), and muscle relaxation was achieved with pancuronium (100 ~g/kg). Patients were ventilated with 100% oxygen, and anesthesia was supplemented with isoflurane as required. Moderate hemodilution (hematocrit 20% to 25%) and systemic hypothermia (25° C) were maintained during cardiopulmonary bypass. Intermittent, multidose, cold crystalloid potassium cardioplegia was employed to maintain the myocardial temperature below 15° C. Concomitant coronary artery bypass grafting was performed with saphenous veins. Distal anastomoses were constructed first to permit vein graft infusions during aortic valve replacement. The selection of the prosthesis was based on the patient's age, history of previous thromboembolism or bleeding disorder, and on the preferences of the patient, cardiologist, and surgeon. Mechanical prostheses were inserted in 42 patients (36 Bjork-Shiley and six Medtronic Hall) and tissuevalves in 235 patients (120 Ionescu-Shiley pericardial, 80 Hancock pericardial, and 35 CarpentierEdwards porcine). All valves were inserted with interrupted pledget-supported sutures. Outcome. The postoperative outcome was assessed prospectively by reviewing the chart and interviewing each patient at the time of hospital discharge. Preoperative and postoperative electrocardiograms were reviewed by a cardiologist who was not involved in the care of the patient. An operative death was recorded if the patient died during the hospital stay. Postoperative low output syndrome was recorded when inotropic medication (for greater than 30 minutes) or the intra-aortic balloon pump were required to treat hypotension (systolic blood pressure less than 90 rom Hg) and hypoperfusion despite an adequate preload. A postoperative myocardialinfarction was reported with the appearance of a new Qwave. Patients who had a left bundle branch block or inadequate R-wave progression had a perioperative myocardial infarction if the creatine kinase MB isoenzyme was elevated (greater than 50 IV It or 8% of the total creatine kinase). A postoperative stroke occurred when a persistent neurologic deficit was present at the time of discharge or death. An incisional infection occurred when prolonged hospitalization was required

because of antibiotic therapy in patients in whom pathogenic organisms were recovered from the wound. Statistical analysis. Statistics were performed using the SAS statistical programs. * The risk factors for each of the postoperative outcomes were assessed by chi square, Fisher's exact, or Mann-Whitney V tests for discrete variables. Continuous variables were tested by t tests or an analysis of variance and differences were specified by Duncan's multiple range test when the F ratio was significant (p < 0.05). Correlations were performed using the General Linear Models procedure. * Stepwise logistic regression was performed with the BMDP program'[ utilizing all discrete and continuous variables. Step selections were based on maximumlikelihood ratios. A p value of less than 0.15 was required for entry, and a p value greater than 0.10 was required for removal. An improvement chi square was calculated for each variable entered. Variables were maintained in the multivariate model if the improvement chi square was statistically significant (p < 0.05). Illustration of the logistic curves was based on the predicted probability for operative mortality and postoperative ventricular dysfunction using the variables found to. be significant. Results Ten patients died, for an overall mortality of 3%. The low output syndrome developed in 34 patients (12%), and a perioperative myocardial infarction occurred in 14 (5%). Ten patients had a persistent neurologic deficit (4%). Twenty-two patients required reoperation for postoperative bleeding (8%). Eleven patients had evidence of a postoperative wound infection (4%). Predictors of mortality. All clinical and hemodynamic variables were included in a stepwise, logistic regression model that selected age to be the only independent predictor of mortality (with an improvement chi square of 7.4, p < 0.0001). Fig. 1 demonstrates the predicted probability of mortality based on age. Patients between 82 and 85 years had a predicted mortality between 15 and 22%. Age was a significant predictor of outcome. The patients who died were older (70 ± 12 years) than the survivors (58 ± 14 years, p < 0.01). No patient less than 50 years old died. Older patients may have had more severe heart disease. Age correlated positively with pulmonary vascular resistance (r = 0.66, p = 0.006) and systemic vascular resistance (r = 0.64, p = 0.0007), and age correlated negatively with cardiac index (r = 0.89, *Statistical Analysis Systems Institute Inc., Cary, N. C. tBMDP Statistical Software, Los Angeles, Calif.

The Journal of

40

Christakis et al.

Thoracic and Cardiovascular Surgery

I

I

I I

I

15

~

~0

::2


"1

I

20

lO

5 /

//

/

/

/

/

/

/

/

/

/

/

/

/

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I I I

60

I I

:; Q. :;

0 ;:

20

15

0 ...J

#

10

5

----//////

Low Risk

40

50 Age

years

Fig. 1. Age was the most important independent predictor of mortality (%) by a logistic regression analysis. The predicted probability of mortality and the 70% confidence limits are depicted.

p = 0.00(1). Although older patients had a higher incidence of coronary artery disease, age had a greater influence on mortality than coronary artery disease. Patients between 50 and 65 years of age had a lower mortality (0.8%) than patients older than 65 years (5%, p < 0.01), although the incidence of coronary artery disease was higher in patients 50 to 65 years old. Among patients older than 50 years, the operative mortality for those with coronary artery disease (5/104, 5%) was similar to that for patients without coronary artery disease (5/116, 4%). The mortality was also similar for patients of any age who had coronary artery disease (4.5%) compared to those without coronary artery disease (1.7%, P = 0.10). However, patients who had anginal chest pain had a higher mortality (6.5%) than those who had coronary artery disease but no anginal symptoms (1.5%) and those who had no coronary disease (1.7%). Coronary bypass grafting did not influence the perioperative mortality, because those who underwent coronary bypass operations had a mortality (4/94, 4%) similar to that of patients who had coronary artery disease but did not undergo coronary bypass operations (0/15,0%). Coronary bypass grafting was not performed in 15 patients with asymptomatic single-vessel coronary artery disease. Aortic valve lesions also did not influence operative mortality, because patients with aortic stenosis had a similar mortality (4.2%) to that of

Age Gradient ·f CAD

20 35

56 50

I I .42 39 Logit Score

I 32

81 50

56 50

56 100

I

17 High Risk

87

106

Fig. 2. Coronary artery disease (CAD +). peak systolic aortic valve gradient (Gradient. mm Hg) and age (yr) were (in decreasing order of significance) the most important independent risk factors for postoperative low output syndrome (%). Combinations of these three factors and their predicted contribution to low output syndrome are illustrated. The predicted incidence of low output syndrome in hypothetical low-risk and high-risk patients is indicated. The dashed lines refer to the 70% confidence limits of the prediction. The logit score is x-axis the predicted log odds of low output.

patients with aortic insufficiency (2.5%) or mixed aortic valve disease (1%). Postoperative ventricular dysfunction. Multivariant analysis identified the following significant risk factors for postoperative ventricular function (in decreasing order of significance): coronary artery disease (with an improvement chi square of 13.6, p < 0.0001), systolic aortic valve gradient (with an improvement chi square of 10.3, p < 0.(01), and age (with an improvement chi square of 4.0, p < 0.04). Fig. 2 illustrates the contribution of these factors to the predicted probability of postoperative low output syndrome. The figure illustrates the risk for hypothetical high-risk and low-risk patients. The incidence of postoperative low output syndrome was higher in patients with coronary disease (22%) than in those without coronary artery disease (6%, p < 0.001). Coronary bypass grafting did not influence postoperative low output syndrome, because those who underwent coronary bypass operations (23%) had an incidence similar to that of patients who had coronary artery disease but did not undergo coronary bypass operations (5%).

Aortic valve replacement

41

20

15 c 0

~

'" E (ij

~

10

'"o0

>-

~

~

5

1 and 2

3

NoLM

3

LM

Number of Diseased Vessels

Fig. 3. Both coronary artery disease (CAD) and age contributed to the incidence (%) of postoperative low output syndrome. Patients older than 65 years had a higher incidence of low output syndrome with or without coronary artery disease. The numbers within the bars refer to the number of patients in each subgroup.

Age was also a predictor of postoperative low output syndrome. Patients with this complications were older (66 ± 11 years) than those without this complication (57 ± 14 years, p < 0.01). Fig. 3 illustrates the combined effect of age and coronary artery disease on the incidence of postoperative low output syndrome. Older patients (> 65 years) had a higher incidence of low output syndrome with or without coronary artery dis-

ease. Patients with aortic stenosis had a higher incidence of postoperative low output syndrome (20/119, 16.8%) than patients with mixed aortic valve disease (8/102, 7.8%) for aortic insufficiency (3/56, 5.3%). The low risk among patients with aortic regurgitation undergoing valve replacement may be due to our current recommendation that operation should be performed with the advent of symptoms or when noninvasive studies demonstrateeven mild ventricular dysfunction in asymptomatic patients. The current recommendation has reduced the number of patients with aortic regurgitation and severe ventricular dysfunction. The peak systolic aortic valve gradient predicted the incidence' of low output syndrome (Table I). When patients with coronary artery disease were excluded from this analysis, systolic aortic gradient was higher in those who had a low output syndrome (93 ± 25 mm Hg) than those who did not

Fig. 4. The extent of coronary artery disease was the only independent predictor of perioperative myocardial infarction (%). Patients with single- or double-vessel coronary artery disease had an incidence similar to that of patients without coronary artery disease. Patients with triple-vessel disease with and without those with left main (LM) stenosis had a significantly higher incidence of perioperative myocardial infarction. The numbers within the bars refer to the number of patients in each subgroup.

(62 ± 30 mm Hg, p = 0.04). The systolic aortic valve gradient was found to correlate positively with left ventricular end-diastolic pressure (r = 0.80, P = 0.01). Patients with moderate or severe aortic stenosis (systolic gradient> 60 mm Hg) had a higher incidence of postoperative low output syndrome (20%) than those with a preoperative gradient less than 60 mm Hg (7%, P = 0.01).

Predictors of postoperative myocardial infarction. Perioperative myocardial infarction was predicted by the extent of coronary artery disease (Fig. 4) by a stepwise logistic regression analysis (with an improvement chi square of 7.6, p < 0.01). The incidence was lowest in those without coronary artery disease and highest in those with triple-vessel disease and left main stenosis. Patients with coronary artery disease had a 7% incidence of perioperative myocardial infarction compared to a 3% incidence in those without coronary artery disease (p = 0.09). Patients with coronary artery disease had a lower cardiac index (2.1 ± 0.6 Lzmin/m') than patients without coronary artery disease (2.5 ± 0.8 L/min/m 2, p = 0.005). Coronary bypass grafting did not influence the incidence of perioperative myocardial

The Journal of

42

Christakis et al.

Thoracic and Cardiovascular Surgery

Table I. The contribution of hemodynamic variables to the incidence of mortality, low output syndrome (LOS), and myocardial infarction (MI) Mortality

LVEDP (mm Hg) PCWP (mm Hg) EF (%) Systolic gradient (mm Hg) CI (Ljminjm') SVR (units) PVR (units) MPAP (mm Hg) RAP (mm Hg)

20 16 55 66 2.2 22 2.9 28 9

± ± ± ± ± ± ± ± ±

6 5 17 25 0.4 6 1.1 12 2

Survival

21 15 64 61 2.3 23 2.3 23 7

± ± ± ± ± ± ± ± ±

9 7 12 30 0.8

10 1.6 10 3

LOS

19 14 56 77 2.1 28 2.5 22 6

± ± ± ± ± ± ± ± ±

7 5 15 27 0.6 10 1.3 9 3

No LOS

21 16 56 60 2.3 25 2.3 23 7

± ± ± ± ± ± ± ± ±

9 7 15 34' 0.8

10 1.6 10 3

M!

20 14 55 84 2.2 31 2.4 20 6

± ± ± ± ± ± ± ± ±

NoM!

6 4 16 44 0.8 12 1.9 5 2

21 16 63 60 2.3 25 2.3 23 7

± ± ± ± ± ± ± ± ±

9 7 15 24t 0.8 10 1.6 10 3

Legend: LVEDP, Left ventricular end-diastolic pressure. PCWP, Pulmonary capillary wedge pressure. EF, Ejection fraction. Systolic gradient, Systolic aortic valve gradient. CI, Cardiac index. SVR, Systemic vascular resistance. PVR, Pulmonary vascular resistance. MPAP, Mean pulmonary arterial pressure. RAP, Right atrial pressure. 'No LOS different from LOS, p < 0.05. tNo MI different from MI, p < 0.05.

infarction, because those who underwent coronary bypass grafting had an 8% incidence compared to a 0% incidence in those who had coronary artery disease but did not undergo coronary bypass operations (not significant). The peak systolic valve gradient was higher in those who had a perioperative myocardial infarction than those who did not (Table I). Finally, the incidence of perioperative myocardial infarction was higher in women, possibly because they were older (63 ± 10 years) than the men (58 ± 14 years, p < 0.02). Nonpredictors of postoperative morbidity and m~r­ tality. A variety of important factors did not influence the postoperative outcome. Cardiac catheterization variables other than the systolic aortic valve gradient did not predict perioperative mortality, ventricular dysfunction, or myocardial infarction (Table I). The predominant presenting symptom, the aortic valve disease, the severity of aortic valve calcification, the preoperative rhythm, and previous valve operations did not influence the outcome (Table 11). In addition, there were no differences among the five surgeons. Patients undergoing urgent aortic valve replacement had a higher incidence of postoperative low output syndrome because they had a higher incidence of endocarditis (19%) than those undergoing elective operations (2%, p < 0.001). Discussion This prospective study was designed to identify the risk factors for operative mortality, postoperative ventricular dysfunction, and perioperative myocardial infarction for patients undergoing contemporary aortic valve replacement. Identification of high-risk patients may permit cardiac surgeons to design strategies to improve the surgical results. Age was identified as the most important independent risk factor for operative

mortality by a multivariate analysis. Coronary artery disease, peak systolic aortic valve gradient, and age predicted postoperative ventricular dysfunction. The extent of coronary artery disease was the only independent predictor of perioperative myocardial infarction. Most studies that have assessed the risk factors for operative mortality employed patient populations encountered between 5 and 13 years in order to have sufficient numbers for valid statistical analysis.v'' These studies have demonstrated that the year of the operation influenced the outcome. Therefore, these studies may not be representative of the current referral patterns or contemporary surgical practice. The routine use of coronary revascularization and cold potassium cardioplegia may have improved the recent results. However, the evaluation of contemporary populations reduces the number of patients available for statistical analysis and may obscure the importance of some risk factors. We evaluated 277 consecutive patients, all of whom underwent coronary angiography and were operated on with modem anesthetic techniques, extensive perioperative hemodynamic monitoring, cold potassium cardioplegia with vein graft infusions during valve replacement, and currently available prostheses. Age. Age was the only independent risk factor for operative mortality. Other studies have also identified age as an important risk factor for aortic valve replacement.S.15-17 Reed and colleagues" demonstrated a 23% mortality in a group of patients over 70 years of age. The results of aortic valve replacement in the elderly have improved in recent years. Between 1960 and 1969, the operative mortality for aortic valve replacement in patients 60 years or older ranged between 8% and 33%.IS-21 Recent reports have described excellent results

Volume 92 !\lumber 1

Aortic valve replacement

July, 1986

in carefully selected septuagenarians undergoing aortic valve replacement.":" Jamieson and colleagues" reported a 4.7% operative mortality in 190 patients older than 65 years. In carefully selected patients with a low incidence of coronary artery disease and without severe hemodynamic derangement, excellent results can be obtained from aortic valve replacement. Copeland and associates 17. 26 identified age, the type of valve lesion, the presence of coronary artery disease, and indices of left ventricular function as the predictors of operative mortality in a study of 1,127 patients undergoing aortic valve replacement between 1963 and 1973. Improvements in myocardial protection may have reduced the significance of coronary artery disease and the type of valve lesion on operative mortality, but age has remained the most significant predictor of outcome. The reason for the higher mortality in older patients remains obscure. Older patients may have had more extensive ventricular dysfunction because of a prolonged course of aortic valve disease. We found a negative correlation between age and cardiac index, which suggested more extensive preoperative ventricular dysfunction. In addition, the friable tissues and the increased susceptibility to cardiac edema frequently encountered inelderly patients may contribute to the higher mortality. Age was also an independent predictor of postoperative ventricular dysfunction. Age at the time of the operative procedure has been reported as a significant risk factor for long-term survival.":" Because age contributes to the risk of both perioperative and long-term mortality, the risk-benefit ratio must be carefully assessed in elderly patients being considered for aortic valve replacement. However, both ventricular function and cardiac symptoms improve after aortic valve replacement in the elderly." 24·26 Therefore, the quality of life as well as longevity must be considered when recommending surgical treatment in elderly patients. Coronary artery disease. Although patients with symptomatic coronary artery disease had a higher operative mortality, neither the presence of coronary artery disease, the extent of coronary artery disease, nor the symptoms associated with coronary artery disease were independently predictive. Earlier studies have suggested that coronary artery disease was associated with a higher mortality.i 27·]3 Coronary artery disease may have been an important risk factor when methods of perioperative myocardial protection were incomplete. Miller and colleagues" employed topical hypothermia and found a 10% mortality in those with coronary artery disease as compared to a 1% mortality for those

43

Table II. The contribution of clinical presentation to the incidence of mortality, low output syndrome (LOS), and myocardial infarction (MI)

a

Pathology Rheumatic/idiopathic calcific Congenital bicuspid Prosthetic dysfunction Infective endocarditis Myxomatous Sex Female Male Timing of operation Elective Urgent Previous operation Yes No Rhythm Atrial fibrillation Normal sinus Valvular calcification None/mild Moderate Severe Preoperative symptoms Dyspnea Fatigue Asymptomatic Presyncope

% Mortality

I % LOS

~

190

3.1

12.9

5.4

51 15 8 13

3.9 6.6 12.5 0

11.7 13.3 12.5 0

5.9 0 0 0

77 200

5.1 3.0

12.9 12.0

9.0* 3.0

241 36

3.3 5.5

11.0* 22.5

4.0 9.7

17 260

5.8 3.4

12.2 12.5

4.7 4.1

36 241

0 4.1

18.6 11.2

5.5 4.6

109 55 113

4.5 3.6 2.7

8.2 12.7 15.9

2.4 3.6 6A

148 24 23 17

4.0 0 0 0

13.5 4.2 8.7 17.6

5.4 4.2 0 11.8

"Differences within each group. p < 0.05

undergoing isolated aortic valve replacement. Loop and colleagues" employed normothermic anoxic arrest and found a 9% mortality in these undergoing aortic valve replacement and coronary bypass grafting, which they believed was three times greater than the mortality for aortic valve replacement alone. Recent studies employing cold crystalloid cardioplegia have found that coronary bypass grafting did not increase the mortality of aortic valve replacement," 15. J4 Both Reed and colleagues" and Macmanus and associates" attributed the improved results in patients undergoing concomitant aortic valve replacement and coronary bypass grafting to better myocardial protection. However, other factors may have contributed to the improved results in patients requiring combined operations. Nunley, Grunkemeier, and Starr" suggested that improved anesthetic techniques and better postoperative care accounted for the improved results in these patients. We found that coronary artery disease was a significant risk factor for postoperative low output syndrome and perioperative myocardial infarction. Coronary

The Journal of Thoracic and Cardiovascular Surgery

4 4 Christakis et al.

artery disease was a risk factor for operative mortality in the 19605. 27-30 The current methods of myocardial preservation may permit these patients to survive with a higher incidence of postoperative complications. Other studies have confirmed the adverse effect of coronary artery disease on postoperative ventricular dysfunction and myocardial ischemia," 14,27-33 Reed and colleagues" demonstrated that the incidence of myocardial infarction was increased in patients who required concomitant coronary bypass grafting. Lytle and associates' reported a higher incidence of new Q waves and more patients with cardiac enzyme elevations when coronary bypass grafting was required in addition to aortic valve replacement. Bonow and colleagues" found no difference in the incidence of myocardial infarction (based on electrocardiographic changes) in those who underwent aortic valve replacement and coronary bypass grafting (9%) than those who underwent aortic valve replacement alone (8%). However, in their study, cardiac enzymes were not measured and some patients with a perioperative infarction may have been missed. Miller and colleagues" found a higher incidence of perioperative myocardial infarction in those patients undergoing concomitant coronary bypass operations (8.3%) compared to patients undergoing isolated aortic valve replacement (2.2%). Most of these studies did not employ a multivariate statistical analysis, and therefore the independent influence of coronary artery disease on postoperative complications was difficult to assess. . Valve lesion. The valve lesion did not influence perioperative mortality, but the peak systolic aortic valve gradient was an important independent risk factor for postoperative ventricular dysfunction. Patients with aortic stenosis or mixed valve disease who had a low peak gradient had an incidence of postoperative ventricular dysfunction that was not different from that of patients with aortic regurgitation. Other studies have also found that the type of valve lesion did not influence perioperative mortality.6,14-16,36 Earlier studies found that patients with aortic insufficiency had a higher operative risk. Copeland and colleagues" found an II % operative mortality for those with aortic insufficiency and a 6% mortality for those with aortic stenosis. Hirshfield and associates' reported a 17% operative mortality for those with aortic insufficiency and a 10% mortality for those with aortic stenosis. Lytle and associates' employed a multivariate analysis and found that aortic insufficiency was an independent risk factor for perioperative mortality. They reported a mortality of II % for those with aortic insufficiency and 4% for those with aortic stenosis. The influence of valve lesion on operative mortality was dependent on the year of the operation. In the early

days of cardiac surgery, aortic valve replacement was only recommended in symptomatic patients despite severe ventricular dysfunction. Recently, aortic valve replacement has been recommended in asymptomatic patients with even mild ventricular dysfunction."!' The advent of noninvasive methods to assess ventricular function in patients with aortic regurgitation has permitted earlier referral in asymptomatic patients." Aortic valve replacement in asymptomatic patients who exhibit ventricular dysfunction at exercise prevents delaying the operation until ventricular function is depressed at rest. Patients with aortic insufficiency were younger and had a lower incidence of coronary artery disease than patients with aortic stenosis or mixed aortic valve disease. Current referral patterns may have improved the results of aortic valve replacement in patients with aortic insufficiency and may have eliminated the type of valve lesion as a risk for valve replacement. Peak systolic aortic valve gradient was a significant risk factor for postoperative ventricular dysfunction. Other perioperative hemodynamic measurements were not predictive of outcome. Patients with severe aortic stenosis frequently have severe left ventricular hypertrophy. Current methods of perioperative myocardial protection may be incomplete, because abnormalities in ventricular function, metabolism, and high-energy phosphate stores have been identified in these patients." Schaff and colleagues" demonstrated that the extent of perioperative ischemic injury correlated with the degree of myocardial hypertrophy in patients undergoing aortic valve replacement.

Improving the results of aortic valve replacement.

Operating before irreversible ventricular dysfunction in patients with aortic valve disease will improve the operative results." 39 Our patients had a mean preoperative ejection fraction of 55% and a cardiac index of 2.2 L/min/m 2• In addition, very few patients were in NYHA CLass IV (49/277 or 18%). Noninvasive assessment of ventricular function?" has permitted referral for operation before irreversible ventricular dysfunction. In addition, exercise testing has resulted in an even better noninvasive assessment of ventricular

function." Improved methods of perioperative myocardial protection may further reduce the incidence of postoperative complications. Patients with concomitant coronary artery disease and aortic valve disease present a challenge for myocardial protection. Aggressive preoperative treatment of patients with severe symptomatic coronary artery stenoses may reduce intraoperative ischemic injury. Intravenous nitroglycerin or the intra-aortic balloon pump may permit anesthetic induction without

Volume 92 Number 1 July, 1986

ventricular decompensation or ischemic injury. Because coronary stenoses limit the distribution of cardioplegic solutions delivered into the coronary ostia, alternative methods are required." Vein graft infusions may cool and protect the most ischemic regions during aortic valve replacement." The use of topical hypothermia or retrograde cardioplegia delivery into the right atrium or coronary sinus42, 4J may permit better cooling of the hypertrophied left ventricle. Blood cardioplegia may be particularly effective in patients with hypertrophy and coronary stenoses. Blood cardioplegia may preserve subendocardialhigh-energy phosphate stores and permit restoration of normal metabolic recovery after crossclamp removal.44 Although age cannot be considered a contraindication to aortic valve replacement, the estimated life expectancy, the quality of life, and the natural history of the disease must be balanced against the immediate and late postoperative risks of aortic valve replacement in the elderly. The results of aortic valve replacement have improved dramatically over the past 15 years. Further improvements may be possible through modification of patient referral patterns, improvements in myocardial protection, and more critical assessment of the timing and indications for operation. We extend our appreciation to Mr. Brian W. Chong and Mr. Jordan A. Herst for assistance with data compilation and analysis and to Ms. Susan Slattery for preparation of the manuscript. REFERENCES 1 Kay EB, Zimmerman HA, Mendelsohn D Jr: Aortic valvular surgery with artificial valves. An analysis of the first 100 patients. Circulation 27:767-771,1963 2 Callard GM, F1ege JB Jr, Todd JC: Combined valvular and coronary artery surgery. Ann Thorac Surg 22:338342, 1976 3 Hirshfeld JW Jr, Epstein SE, Roberts AJ, Glancy DL, Morrow AG: Indices predicting long-term survival after valve replacement in patients with aortic regurgitation and patients with aortic stenosis. Circulation 50: 1190-1199, 1974 4 MacmanusQ, Grunkemeier GL, Lambert LE, Tepley JF, Harlan BJ, Starr A: Year of operation as a risk factor in the late results of valve replacement. J THoRAc CARDlOVAse SURG 80:834-841, 1980 5 Dale J, Levang 0, Enge I: Long-term results after aortic valve replacement with four different prostheses. Am Heart J 99: 155-162, 1980 6 Cheung D, Flemma RJ, Mullen DC, Lepley D Jr, Anderson AJ, Weirauch E: Ten-year follow-up in aortic valve replacement using the Bjork-Shiley prosthesis. Ann Thorac Surg 32: 138-145, 1981

Aortic valve replacement

45

7 Lytle BW, Cosgrove DM, Loop FD, Taylor PC, Gill CC, Golding LAR, Goormastic M, Groves LK: Replacement of aortic valve combined with myocardial revascularization. Determinants of early and late risk for 500 patients, 1967-1981. Circulation 68:1149-1162, 1983 8 Janusz MT, Jamieson WRE, Allen P, Munro AI, Miyagishima RT, Tutassura H, Burr LH, Gerein AN, Tyers GFO: Experience with the Carpentier-Edwards porcine valve prosthesis in 700 patients. Ann Thorac Surg 34:625633, 1982 9 Henry WL, Bonow RO, Borer JS, Ware JH, Kent KM, Redwood DR, McIntosh CL, Morrow AG, Epstein SE: Observations on the optimum time for operative intervention for aortic regurgitation. I. Evaluation of the results of aortic valve replacement in symptomatic patients. Circulation 61:471-483, 1980 10 Henry WL, Bonow RO, Rosing DR, Epstein SE: Observations on the optimum time for operative intervention for aortic regurgitation. II. Serial echocardiographic evaluation of asymptomatic patients. Circulation 61:484-492, 1980 II Bonow RO, Rosing DR, Kent KM, Epstein SE: Timing of operation for chronic aortic regurgitation. Am J Cardiol 50:325-336, 1982 12 Cunha CLP, Giuliani ER, Fuster V, Seward JB, Brandenburg RO, McGoon DC: Preoperative M-mode echocardiography as a predictor of surgical results in chronic aortic insufficiency. J THORAC CARDIOVASC SURG 79:256265, 1980 13 Weisel RD, Hoy F, Baird RJ, Ivanov J, Burns RJ, Madonik M, McLaughlin PR: Improved myocardial protection during a prolonged cross-clamp period. Ann Thorae Surg 36:664-670, 1983 14 Nunley DL, Grunkemeier GL, Starr A: Aortic valve replacement with coronary bypass grafting. Significant determinants of ten-year survival. J THoRAc CARDIOVASC SURG 85:705-711,1983 15 Reed GE, Sanoudos GM, Pooley RW, Moggio RA, McClung JA, Somberg ED, Praeger PI: Results of combined valvular and myocardial revascularization operations. J THORAC CARDIOVASC SURG 85:422-426, 1983 16 Jacobs ML, Fowler BN, Vezeridis MP, Jones N. Daggett WM: Aortic valve replacement. A 9-year experience. Ann Thorac Surg 30:439-447, 1980 17 Copeland JG, Griepp RB, Stinson EB, Shumway NE: Long-term follow-up after isolated aortic valve replacement. J THoRAc CARDIOVASC SURG 74:875-888, 1977 18 Ahmad A, Starr A: Valve replacement of geriatric patients. Br Heart J 31:322-325, 1969 19 Austen WG, De Sanctis RW, Buckley MJ, Mundth ED, Scanell JG: Surgical management of aortic valve disease in elderly. JAMA 211:624-628, 1970 20 Barnhorst DA, Giuliani ER, Pluth JR, Danielson GK, Wallace RB, McGoon DC: Open heart surgery in patients more than 65 years old. Ann Thorac Surg 18:81-84, 1974 21 Oh W, Hickman R, Emanuel R, McDonald L, Somerville

46

22

23

24

25

26

27

28

29

30

31

32

33

34

The Journal of Thoracic and Cardiovascular Surgery

Christakis et al.

J, Ross D, Ross K, Gonzalez-Lavim L: Heart valve surgery in 114 patients over the age of 60. Br Heart J 35:174-177,1973 Jamieson WRE, Dooner J, Munro AI, Janusz MT, Burgess JJ, Miyagishima RR, Gerein AN, Allen P: Cardiac valve replacement in the elderly. A review of 320 consecutive cases. Circulation 64:Supp1 2:177-183,1981 Bessone LN, Pupello DF, Blank RH, Harrison EE, Sbar S: Valve replacement in patients over 70 years. Ann Thorac Surg 24:417-421,1977 Hochberg MS, Morrow AG, Michaelis LL, McIntosh CL, Redwood DR, Epstein SE: Aortic valve replacement in the elderly. Encouraging postoperative clinical and hemodynamic results. Arch Surg 112:1475-1480, 1977 Quinlan R, Cohn LH, Collins JJ: Determinants of survival following cardiac operations in elderly patients. Chest 68:498-502, 1975 Copeland JG, Griepp RB, Stinson EB, Shumway NE: Isolated aortic valve replacement in patients older than 65 years. JAMA 237:1578-1581, 1977 Berndt TB, Hancock EW, Shumway NE, Harrison DC: Aortic valve replacement with and without coronary artery bypass surgery. Circulation 50:967-971, 1974 Berger TJ, Karp RB, Kouchoukos NT: Valve replacement and myocardial revascularization. Results of combined operation in 59 patients. Circulation 51, 52:Suppl 1:126131, 1975 Rossiter SJ, Hultgren HN, Kosek JC, Wuerflein RD, Angell WW: Myocardial damage in combined valvular and coronary bypass surgery. Circulation 51, 52:Suppl 1:119-125,1975 Craver JM, Jones EL, Hatcher CR Jr, Farmer JH: Concomitant aortic valve replacement and myocardial revascularization. Ann Surg 185:713-716, 1977 Miller DC, Stinson EB, Oyer PE, Rossiter SJ, Reitz BA, Shumway NE: Surgical implications and results of combined aortic valve replacement and myocardial revascularization. Am J Cardiol 43:494-501, 1979 Loop FD, Phillips DF, Roy M, Taylor PC, Groves LK, Effler DB: Aortic valve replacement combined with myocardial revascularization. Late clinical results and survival of surgically-treated aortic valve patients with and without coronary artery disease. Circulation 55: 169-173, 1977 Richardson JV, Kouchoukos NT, Wright JO 1Il, Karp RB: Combined aortic valve replacement and myocardial revascularization. Results in 220 patients. Circulation 59:75-81,1979 Macmanus Q, Grunkemeier G, Lambert L, Dietl C, Starr A: Aortic valve replacement and aorta-coronary bypass

35

36

37

38

39

40

41

42

43

44

surgery. Results with perfusion of proximal and distal coronary arteries. J THoRAc CARDIOVASC SURG 75:865869, 1978 Bonow RO, Kent KM, Rosing DR, Lipson LC, Borer JS, McIntosh CL, Morrow AG, Epstein SE: Aortic valve replacement without myocardial revascularization in patients with combined aortic valvular and coronary artery disease. Circulation 63:243-253, 1981 Schwarz F, Flameng W, Lagebartels F, Sesto M, Walter P, Schlepper M: Impaired left ventricular function in chronic aortic valve disease. Survival and function after replacement by Bjork-Shiley prosthesis. Circulation 60:48-58, 1979 Forman R, Firth BG, Barnard MS: Prognostic significance of preoperative left ventricular ejection fraction and valve lesion in patients with aortic valve replacement. Am J Cardiol 45: 1120-1125, 1980 Schaff HV, Bixler TJ, Flahery JT, Brawley RK, Donahoo JS, Goldman RA, Gott VL, Gardner TJ: Identification of persistent myocardial ischemia in patients developing left ventricular dysfunction following aortic valve replacement. Surgery 86:70-77, 1979 Samuels DA, Curfman GD, Friedlich AL, Buckley MJ, Austen WG: Valve replacement for aortic regurgitation. Long-term follow-up with factors influencing the results. Circulation 60:647-654, 1979 Borer JS, Rosing DR, Kent KM, Bacharach SL, Green MV, McIntosh CJ, Morrow AG, Epstein SE: Left ventricular function at rest and during exercise after aortic valve replacement in patients with aortic regurgitation. Am J Cardiol 44:1297-1305, 1979 Weisel RD, Hoy FBY, Baird RJ, Ivanov J, Hilton JD, Burns RJ, Mickle DAG, Mickleborough LL, Scully HE, Goldman BS, McLaughlin PR: A comparison of alternative cardioplegic techniques. J THORAC CARDIOVASC SURG 86:97-107, 1983 Fabiani IN, Carpentier A: Comparative evaluation of retrograde cardioplegia through the coronary sinus and the right atrium (abstr). Circulation 68:Suppl 3:25, 1983 Menasche P, Kural S, Fauchet M, Lavergne A, Commin P, Bercot M, Touchot B, Georgiopoulos G, Piwnica A: Retrograde coronary sinus perfusion. A safe alternative for ensuring cardioplegic delivery in aortic valve surgery. Ann Thorac Surg 34:647-658, 1982 Fremes SE, Christakis GT, Weisel RD, Mickle DAG, Madonik MM, Ivanov J, Harding R, Seawright 5J, Houle S, McLaughlin PR, Baird RJ: A clinical trial of blood and crystalloid cardioplegia. J THORAC CARDIOVASC SURG 88:726-741, 1984