- Email: [email protected]
Predicting aortic valve prosthesis size
J THORAC CARDIOVASC SURG 81:450-454, 1981
Predicting aortic valve prosthesis size A step toward better matching of patient and prosthesis We have developed a simple method of predicting aortic valve prosthesis size from supravalvular aortic cineangiograms, taken in the 30 degree right anterior oblique projection, during routine diagnostic cardiac catheterization. Measurements made on the aortogram were corrected to true size by means of a 1 em? grid reference plate filmed at the conclusion of the catheterization. This method was utilized prospectively to predict the anulus size in 26 patients undergoing aortic valve replacement alone or as part of a more extensive cardiac procedure. The catheterization prediction of anulus size was perfectly predictive of the prosthesis size used in eight of 26 (31%) cases, within 1 mm of the prosthesis size used in 20 of 26 (77%), and within one prosthesis size in all 26 (100%) cases. Comparing the measured anulus diameter to the actual prosthesis diameter yielded r = 0.93. When both systolic and diastolic frames were available for analysis, the diastolic frame was more accurately predictive than the systolic frame (diastolic r = 0.93; systolic r = 0.88), although both yielded acceptable results. Utilization of this method has significantly altered operative plans in four patients with a small anulus. Such preoperative assessment should allow improved matching of patient and valve prosthesis and thereby yield improved long-term results in aortic valve replacement.
Joseph D. Babb, M.D., Grant V. S. Parr, M.D., and Martin J. O'Neill, Jr., M.D., Hershey, Pa.
T
he morbidity and long-term mortality of valve prosthesis-patient mismatch have received increasing attention. 1-6 This is especially true in patients with a small aortic anulus, in whom inadequate prosthesis orifice area may result in postoperative aortic stenosis. Thus we and others 1 have perceived the need for a means of accurate preoperative prediction of realistic valve size in individual patients. This report describes a technique of determining valve size from a supravalvular aortic cineangiogram taken in the 30 degree right anterior oblique projection. The accuracy and clinical application of this technique are discussed.
Patients and methods Beginning in February, 1979, all patients undergoing cardiac catheterization at The Milton S. Hershey MediFrom the Division of Cardiology, Department of Medicine, the Bridgeport Hospital, Bridgeport, Conn., and the Division of Cardiothoracic Surgery, Department of Surgery, The Milton S. Hershey Medical Center, The Pennsylvania State University, Hershey, Pa. Received for publication Feb. 19,1980. Accepted for publication July 15, 1980. Address for reprints: Joseph D. Babb, M.D., Section of Cardiology, the Bridgeport Hospital, 267 Grant St., Bridgeport, Conn. 06602.
450
cal Center for evaluation of aortic valve disease had a supravalvular aortic cineangiograrn taken in the 30 degree right anterior oblique projection, regardless of the precatheterization impression of the presence or absence of significant aortic regurgitation. Injection was performed with 20 to 40 ml of meglumine diatrizoate (Renografin-76) and recorded on 35 mm cine film at 60 frames/sec. At the conclusion of the study, a I em" grid plate was filmed to provide correction for magnification in the manner utilized for ventriculographic volume analysis. t Subsequently, with a stable projector to screen distance, a representative frame from the supravalvular aortic cineangiogram was selected and the outline of the aortic root and cusps were traced. This generally was a frame early in the injection phase before sufficient contrast had regurgitated into the left ventricle to obscure sharp definition of the cusps. Care was taken to follow precisely the sharp margins of the aortic root and cusps. In like manner, with the same projector to screen distance, the grid was traced at a level corresponding to the portion of the aortic root on the supravalvular film. From this, a correction factor was determined by the formula:
= Correction factor True grid linear distance Measured grid linear distance
0022-5223/81/030450+05$00.50/0 © 198I The C. V. Mosby Co.
Volume 81 Number 3 March,1981
Predicting aortic valve prosthesis size
45 1
Fig. 1. The described technique for anulus sizing with two sinuses of Valsalvaseen. Left. A representativeframe from a supravalvular aortic cineangiogram taken in the 30 degree right anterior oblique projection. Right, The tracing of the aortic root and cusps with the transverse line drawn I mm above the lowermostcusp margins. The distance between the vertical hash marks is measuredand multipliedby the correction factor to yield the predicted anulus diameter.
Fig. 2. The described technique for anulus sizing with three sinuses of Valsalva seen. For further legend see Fig. 1. Depending on the aortic root anatomy, either two or three sinuses were visualized in the right anterior oblique projection. When two sinuses were visualized, as in Fig. I, the transverse diameter of the traced aortic root was measured I mm above the base of these cusps. When three sinuses were visualized, as in Fig. 2, the transverse diameter was measured at a level 1 mm above only the two lowermost cusps (noncoronary and right coronary sinuses). By means of the grid correction factor, this measured diameter was converted to a true anulus diameter. Placement of the transverse line 1 mm above the bases of the cusps was found on empiric
grounds to yield the best in vivo correlation early in our series. While some minor variation of 0.5 mm more or less has been found in some patients, the 1 mm level has proved to be the most convenient and consistent. Aortic valve replacement was performed by itself or in combination with other procedures. Standard techniques of cardiopulmonary bypass and cold potassium cardioplegia were employed. The surgeon placed the largest prosthesis possible into the patient's anulus. The decision of prosthesis type was based on anulus size, patient age and lifestyle, possible contraindications to anticoagulants, and surgeon preference. The size of the
The Journal of Thoracic and Cardiovascular Surgery
452 Babb, Parr, O'Neill
33 E
.
31
E
a: UJ .... UJ
29
s Cl
27
:;;
~
en
25
0
23
.... en a::
c, ....J
<0: ::::l
....
U <0:
/
~
/
/
/
/
/
/
/
/
/
/
// /
/'
21 ./ /
19 17
y=2.52+0.89x r=0.93 Sy.x = 1.17
",,,
/,
/,
/
/
17
/
/
/,
/
,
/,
.<
•
/,
19
21
23
25
27
Table I. Characteristics of 26 patients undergoing preoperative determination of valve size Prosthesis size (OD in mm)
~
UJ
::c
•
/
/
/
/
29
31
33
ANGIOGRAPHIC ANNULUS DIAMETER(mm)
Fig. 3. Comparison of angiographic anulus diameter (predictedvalvesize) to actualprosthesis diameter. The solid line is the line of unity. The dotted line is that determined by the regression equation, which is shown along with the correlation coefficient, r and the standard error of estimate of yon x, Sy.x·
prosthesis actually employed was correlated with the angiographic prediction. Late in the series, as experience with the technique showed predictive accuracy, patients with small predicted anulus sizes were called to the attention of the operating surgeon to assist in preoperative planning. Results
Since this procedure was put into practice, 45 adult patients with aortic valve disease have been studied at our institution. Of these, II have not undergone valve replacement. An aortic root injection was not performed in two patients and a grid film was not made in six. The remaining study group comprised 26 patients with an average age of 59.5 years (range 23 to 78 years) who underwent aortic valve replacement. Procedures included isolated aortic valve replacement (13); aortic valve replacement with coronary grafting (four); combined aortic and mitral valve replacement (four); aortic, mitral, and tricuspid valve replacement (one); aortic and mitral valve replacement with tricuspid annuloplasty (one); aortic and mitral valve replacement with coronary grafting (one); composite graft replacement of the aortic valve and ascending aorta (one); and aortic valve replacement with tricuspid annuloplasty (one). The only perioperative death occurred in the patient who had aortic valve replacement and tricuspid an-
Sex Male Female Cause of valve disease Bicuspid Rheumatic Degenerative Miscellaneous Aortic valve calcification None Mild Moderate Severe Aortic regurgitation Yes No
No.
Mean ± SEM
13 13
25.3 ± 0.9 22.3 ± 0.6*
I
Range
19-31 19-25
8 8 6 4
24.5 23.9 22.3 24.5
± ± ± ±
1.2 1.2 1.1 1.5
19-29 21-31 19-25 21-27
8
24.5 21.8 23.6 25.3
± ± ± ±
0.9 1.9 0.8 2.0
21-27 19-29 19-27 22-31
24.4 ± 0.7 21.4 ± 1.2
19-31 19-25
5 9 4 21 5
Legend: 00, Outer diameter. *p = 0.02.
nuloplasty. This death occurred late postoperatively because of perforation of a duodenal ulcer. Characteristics of the 26 patients undergoing anulus sizing are shown in Table I. Valves were significantly smaller in women than men (p = 0.02). Patients with aortic regurgitation tended to have larger valves, but the difference was not significant. Variability was such that these factors were of no practical use in predicting the size of the aortic valvular prosthesis. Neither the degree of aortic valvular calcification nor the cause of the valvular disease correlated with prosthesis size. The comparison of predicted valve size to actual valve size is shown in Fig. 3. In eight of 26 (31%) patients, predicted and actual size coincided exactly; in 20 of 26 (77%) they were within I mm of each other. In all 26 (100%) cases, predictions were accurate within one valve size. In general, diastolic frames were better defined and easier to draw than systolic frames. In some cases, however, both could be drawn accurately. Analysis of predictive accuracy with systolic frames, when available, or diastolic frames, when available, disclosed correlation coefficients (r values) of 0.88 and 0.93, respectively. Although plots of actual versus predicted size revealed more frequent exact concordance when the systolic frames were used preferentially, the distribution of nonconcordant cases served to produce a more accurate linear regression when diastolic frames were used. However, since the respective correlation
Volume 81 Number 3 March,1981
coefficients are not markedly different, it appears that neither frame is significantly better than the other (systolic frame y = 2.50 + 0.92X, r = 0.88, Sy . x = 1.54; diastolic frame y = 2.52 + 0.89X, r = 0.93, Sy . x = 1.17).
Discussion Selection of an aortic valve prosthesis presented few problems when the choice was limited. With the development of acceptable new valve prostheses, however, the cardiac surgeon currently has a valid choice of valves for a given patient. This has generated recent interest in valve prosthesis-patient mismatch, wherein valve replacement leaves a patient with a "stenotic" prosthesis for his cardiovascular needs. 1 This problem is most likely to occur in patients with a small anulus. However, the experience at our institution over the past 2 years indicates that 40% of 85 patients received aortic valve prosthesis of 25 mm in diameter or less. This is the range wherein significant gradients can occur at normal flow rates across the valve. The ability to predict the prosthesis size might help to overcome this problem. Advance knowledge of a small anulus would allow the surgeon to plan an anulus-enlarging procedure and discuss it with the patient preoperatively. 2. 3 In some patients who have an extremely small anulus and in children or adolescents, such a procedure may be the only suitable alternative to leaving the patient with an unacceptably small prosthesis. After having calculated that the aortic anulus would be small, we have performed three such procedures in children without any deaths. However, the reported procedural risk of 24%5 with anulus enlargement is generally higher than the 2.5% 30 days mortality experienced at this institution for aortic valve replacement using cold potassium cardioplegia (unpublished observations). For most adult patients, an acceptable alternative to enlarging the anulus is to select a prosthesis with acceptable flow characteristics in small sizes. Although this solution may not totally overcome the problem of valve prosthesis-patient mismatch, it should help to minimize it, particularly if the operating surgeon has a strong preference for a specific prosthesis that has unacceptable flow characteristics in smaller sizes. Additionally, the accurate preoperative selection ·of valve type will give advance knowledge of the necessity for long-term anticoagulation. For patients in whom such therapy would be hazardous, alternative valves could be utilized for an anulus-enlargement procedure planned. As an example, a 63-year-old woman had severe aortic stenosis necessitating valve replacement. She also had
Predicting aortic valve prosthesis size
453
a history of peptic ulcer disease with a recent acute bleeding episode making the long-term use of anticoagulants unacceptable. Her prosthesis size was predicted to be 21 rom. Normally, in such a small aortic root, a disc-type prosthesis would have been used, but this would necessitate long-term warfarin sodium (Coumadin) therapy. With this prior knowledge, the operation was delayed until an Ionescu-Shiley valve could be obtained. Subsequently, a 21 mm IonescuShiley valve was implanted without difficulty and the patient has done well without requiring anticoagulants. We have found this technique to be accurate in all forms of aortic valvular disease, including Marfan aneurysms of the ascending aorta. We have gained sufficient confidence in this procedure that we did not recommend valve replacement in a patient with New York Heart Association Class II symptoms of pulmonary venous hypertension and a calculated aortic anulus of 19 mm. With this small anulus, the estimated postoperative transvalvular gradient would have been similar to her existing gradient. Thus an operation would have offered little benefit in this particular patient. We have identified two possible causes for imperfect measurement of valve size. First, in the presence of very heavy calcification, it may be impossible or unwise to remove all of the annular calcium. Thus, after debridement, the aortic anulus may be somewhat smaller than the angiographic estimate. Second, although the patient is consistently positioned in a 30 degree right anterior oblique projection, the intrathoracic aortic root position may vary somewhat from patient to patient. The radiographic appearance of two versus three sinuses is not always related to the presence of a bicuspid versus a tricuspid valve. Instead, the view may vary somewhat from the true perpendicular and cause variation in sinus visualization. Depending upon the degree and angle of variation, the angiographic estimate of anulus size may be greater or less than actual size. These inherent errors have not been great, however, not exceeding I to 2 mm. Thus, despite methodologic limitations, this technique has proved reliable and useful. Prospectively determining the valve size can facilitate the appropriate preoperative selection of valve type and/or operative technique. Although in many cases this may simplify the surgeon's task only minimally, in some cases it can make a significant contribution to achieving a satisfactory operative result. REFERENCES Rahimtoola SH: The problem of valve prosthesis-patient mismatch. Circulation 58:20-24, 1978
The Journal of
454 Babb, Parr, O'Neill
Thoracic and Cardiovascular Surgery
2 Konno S, Imai Y, Iida Y, Nakajima M, Tatsuno K: A new method for prosthetic valve replacement in congenital aortic stenosis associated with hypoplasia of the aortic valve ring. J THORAC CARDIOVASC SURG 70:909-917, 1975 3 Blank R, Pupello DF, Bessoe LN, Harrison EE, Sbar S: Method of managing the small aortic anulus during valve replacement. Ann Thorac Surg 22:356-361, 1976 4 Symbas PN, Ware RE, Hatcher CR Jr, Temesy-Armos PN: An operation for relief of severe left ventricular outflow tract obstruction. J THORAC CARDIOVASC SURG 71:245249, 1976 5 Rastan H, Abu-Aishah N, Rastan D, Heisig B, Koncz J,
Bjornstad PG, Beuren AJ: Results of aortoventriculoplasty in 21 consecutive patients with left ventricular outflow tract obstruction. J THORAC CARDIOVASC SURG 75:659-669, 1978 6 Rittenhouse EA, Sauvage LR, Stamm SJ, Mansfield PB, Hall DG, Herndon PS: Radical enlargement of the aortic root and outflow tract to allow valve replacement. Ann Thorac Surg 27:367-373, 1979 7 Kasser IS, Kennedy JW: Measurement of left ventricular volumes in man by single-plane cineangiocardiography. Invest Radiol 4:83-90, 1969
Predicting aortic valve prosthesis size A step toward better matching of patient and prosthesis We have developed a simple method of predicting aortic valve prosthesis size from supravalvular aortic cineangiograms, taken in the 30 degree right anterior oblique projection, during routine diagnostic cardiac catheterization. Measurements made on the aortogram were corrected to true size by means of a 1 em? grid reference plate filmed at the conclusion of the catheterization. This method was utilized prospectively to predict the anulus size in 26 patients undergoing aortic valve replacement alone or as part of a more extensive cardiac procedure. The catheterization prediction of anulus size was perfectly predictive of the prosthesis size used in eight of 26 (31%) cases, within 1 mm of the prosthesis size used in 20 of 26 (77%), and within one prosthesis size in all 26 (100%) cases. Comparing the measured anulus diameter to the actual prosthesis diameter yielded r = 0.93. When both systolic and diastolic frames were available for analysis, the diastolic frame was more accurately predictive than the systolic frame (diastolic r = 0.93; systolic r = 0.88), although both yielded acceptable results. Utilization of this method has significantly altered operative plans in four patients with a small anulus. Such preoperative assessment should allow improved matching of patient and valve prosthesis and thereby yield improved long-term results in aortic valve replacement.
Joseph D. Babb, M.D., Grant V. S. Parr, M.D., and Martin J. O'Neill, Jr., M.D., Hershey, Pa.
T
he morbidity and long-term mortality of valve prosthesis-patient mismatch have received increasing attention. 1-6 This is especially true in patients with a small aortic anulus, in whom inadequate prosthesis orifice area may result in postoperative aortic stenosis. Thus we and others 1 have perceived the need for a means of accurate preoperative prediction of realistic valve size in individual patients. This report describes a technique of determining valve size from a supravalvular aortic cineangiogram taken in the 30 degree right anterior oblique projection. The accuracy and clinical application of this technique are discussed.
Patients and methods Beginning in February, 1979, all patients undergoing cardiac catheterization at The Milton S. Hershey MediFrom the Division of Cardiology, Department of Medicine, the Bridgeport Hospital, Bridgeport, Conn., and the Division of Cardiothoracic Surgery, Department of Surgery, The Milton S. Hershey Medical Center, The Pennsylvania State University, Hershey, Pa. Received for publication Feb. 19,1980. Accepted for publication July 15, 1980. Address for reprints: Joseph D. Babb, M.D., Section of Cardiology, the Bridgeport Hospital, 267 Grant St., Bridgeport, Conn. 06602.
450
cal Center for evaluation of aortic valve disease had a supravalvular aortic cineangiograrn taken in the 30 degree right anterior oblique projection, regardless of the precatheterization impression of the presence or absence of significant aortic regurgitation. Injection was performed with 20 to 40 ml of meglumine diatrizoate (Renografin-76) and recorded on 35 mm cine film at 60 frames/sec. At the conclusion of the study, a I em" grid plate was filmed to provide correction for magnification in the manner utilized for ventriculographic volume analysis. t Subsequently, with a stable projector to screen distance, a representative frame from the supravalvular aortic cineangiogram was selected and the outline of the aortic root and cusps were traced. This generally was a frame early in the injection phase before sufficient contrast had regurgitated into the left ventricle to obscure sharp definition of the cusps. Care was taken to follow precisely the sharp margins of the aortic root and cusps. In like manner, with the same projector to screen distance, the grid was traced at a level corresponding to the portion of the aortic root on the supravalvular film. From this, a correction factor was determined by the formula:
= Correction factor True grid linear distance Measured grid linear distance
0022-5223/81/030450+05$00.50/0 © 198I The C. V. Mosby Co.
Volume 81 Number 3 March,1981
Predicting aortic valve prosthesis size
45 1
Fig. 1. The described technique for anulus sizing with two sinuses of Valsalvaseen. Left. A representativeframe from a supravalvular aortic cineangiogram taken in the 30 degree right anterior oblique projection. Right, The tracing of the aortic root and cusps with the transverse line drawn I mm above the lowermostcusp margins. The distance between the vertical hash marks is measuredand multipliedby the correction factor to yield the predicted anulus diameter.
Fig. 2. The described technique for anulus sizing with three sinuses of Valsalva seen. For further legend see Fig. 1. Depending on the aortic root anatomy, either two or three sinuses were visualized in the right anterior oblique projection. When two sinuses were visualized, as in Fig. I, the transverse diameter of the traced aortic root was measured I mm above the base of these cusps. When three sinuses were visualized, as in Fig. 2, the transverse diameter was measured at a level 1 mm above only the two lowermost cusps (noncoronary and right coronary sinuses). By means of the grid correction factor, this measured diameter was converted to a true anulus diameter. Placement of the transverse line 1 mm above the bases of the cusps was found on empiric
grounds to yield the best in vivo correlation early in our series. While some minor variation of 0.5 mm more or less has been found in some patients, the 1 mm level has proved to be the most convenient and consistent. Aortic valve replacement was performed by itself or in combination with other procedures. Standard techniques of cardiopulmonary bypass and cold potassium cardioplegia were employed. The surgeon placed the largest prosthesis possible into the patient's anulus. The decision of prosthesis type was based on anulus size, patient age and lifestyle, possible contraindications to anticoagulants, and surgeon preference. The size of the
The Journal of Thoracic and Cardiovascular Surgery
452 Babb, Parr, O'Neill
33 E
.
31
E
a: UJ .... UJ
29
s Cl
27
:;;
~
en
25
0
23
.... en a::
c, ....J
<0: ::::l
....
U <0:
/
~
/
/
/
/
/
/
/
/
/
/
// /
/'
21 ./ /
19 17
y=2.52+0.89x r=0.93 Sy.x = 1.17
",,,
/,
/,
/
/
17
/
/
/,
/
,
/,
.<
•
/,
19
21
23
25
27
Table I. Characteristics of 26 patients undergoing preoperative determination of valve size Prosthesis size (OD in mm)
~
UJ
::c
•
/
/
/
/
29
31
33
ANGIOGRAPHIC ANNULUS DIAMETER(mm)
Fig. 3. Comparison of angiographic anulus diameter (predictedvalvesize) to actualprosthesis diameter. The solid line is the line of unity. The dotted line is that determined by the regression equation, which is shown along with the correlation coefficient, r and the standard error of estimate of yon x, Sy.x·
prosthesis actually employed was correlated with the angiographic prediction. Late in the series, as experience with the technique showed predictive accuracy, patients with small predicted anulus sizes were called to the attention of the operating surgeon to assist in preoperative planning. Results
Since this procedure was put into practice, 45 adult patients with aortic valve disease have been studied at our institution. Of these, II have not undergone valve replacement. An aortic root injection was not performed in two patients and a grid film was not made in six. The remaining study group comprised 26 patients with an average age of 59.5 years (range 23 to 78 years) who underwent aortic valve replacement. Procedures included isolated aortic valve replacement (13); aortic valve replacement with coronary grafting (four); combined aortic and mitral valve replacement (four); aortic, mitral, and tricuspid valve replacement (one); aortic and mitral valve replacement with tricuspid annuloplasty (one); aortic and mitral valve replacement with coronary grafting (one); composite graft replacement of the aortic valve and ascending aorta (one); and aortic valve replacement with tricuspid annuloplasty (one). The only perioperative death occurred in the patient who had aortic valve replacement and tricuspid an-
Sex Male Female Cause of valve disease Bicuspid Rheumatic Degenerative Miscellaneous Aortic valve calcification None Mild Moderate Severe Aortic regurgitation Yes No
No.
Mean ± SEM
13 13
25.3 ± 0.9 22.3 ± 0.6*
I
Range
19-31 19-25
8 8 6 4
24.5 23.9 22.3 24.5
± ± ± ±
1.2 1.2 1.1 1.5
19-29 21-31 19-25 21-27
8
24.5 21.8 23.6 25.3
± ± ± ±
0.9 1.9 0.8 2.0
21-27 19-29 19-27 22-31
24.4 ± 0.7 21.4 ± 1.2
19-31 19-25
5 9 4 21 5
Legend: 00, Outer diameter. *p = 0.02.
nuloplasty. This death occurred late postoperatively because of perforation of a duodenal ulcer. Characteristics of the 26 patients undergoing anulus sizing are shown in Table I. Valves were significantly smaller in women than men (p = 0.02). Patients with aortic regurgitation tended to have larger valves, but the difference was not significant. Variability was such that these factors were of no practical use in predicting the size of the aortic valvular prosthesis. Neither the degree of aortic valvular calcification nor the cause of the valvular disease correlated with prosthesis size. The comparison of predicted valve size to actual valve size is shown in Fig. 3. In eight of 26 (31%) patients, predicted and actual size coincided exactly; in 20 of 26 (77%) they were within I mm of each other. In all 26 (100%) cases, predictions were accurate within one valve size. In general, diastolic frames were better defined and easier to draw than systolic frames. In some cases, however, both could be drawn accurately. Analysis of predictive accuracy with systolic frames, when available, or diastolic frames, when available, disclosed correlation coefficients (r values) of 0.88 and 0.93, respectively. Although plots of actual versus predicted size revealed more frequent exact concordance when the systolic frames were used preferentially, the distribution of nonconcordant cases served to produce a more accurate linear regression when diastolic frames were used. However, since the respective correlation
Volume 81 Number 3 March,1981
coefficients are not markedly different, it appears that neither frame is significantly better than the other (systolic frame y = 2.50 + 0.92X, r = 0.88, Sy . x = 1.54; diastolic frame y = 2.52 + 0.89X, r = 0.93, Sy . x = 1.17).
Discussion Selection of an aortic valve prosthesis presented few problems when the choice was limited. With the development of acceptable new valve prostheses, however, the cardiac surgeon currently has a valid choice of valves for a given patient. This has generated recent interest in valve prosthesis-patient mismatch, wherein valve replacement leaves a patient with a "stenotic" prosthesis for his cardiovascular needs. 1 This problem is most likely to occur in patients with a small anulus. However, the experience at our institution over the past 2 years indicates that 40% of 85 patients received aortic valve prosthesis of 25 mm in diameter or less. This is the range wherein significant gradients can occur at normal flow rates across the valve. The ability to predict the prosthesis size might help to overcome this problem. Advance knowledge of a small anulus would allow the surgeon to plan an anulus-enlarging procedure and discuss it with the patient preoperatively. 2. 3 In some patients who have an extremely small anulus and in children or adolescents, such a procedure may be the only suitable alternative to leaving the patient with an unacceptably small prosthesis. After having calculated that the aortic anulus would be small, we have performed three such procedures in children without any deaths. However, the reported procedural risk of 24%5 with anulus enlargement is generally higher than the 2.5% 30 days mortality experienced at this institution for aortic valve replacement using cold potassium cardioplegia (unpublished observations). For most adult patients, an acceptable alternative to enlarging the anulus is to select a prosthesis with acceptable flow characteristics in small sizes. Although this solution may not totally overcome the problem of valve prosthesis-patient mismatch, it should help to minimize it, particularly if the operating surgeon has a strong preference for a specific prosthesis that has unacceptable flow characteristics in smaller sizes. Additionally, the accurate preoperative selection ·of valve type will give advance knowledge of the necessity for long-term anticoagulation. For patients in whom such therapy would be hazardous, alternative valves could be utilized for an anulus-enlargement procedure planned. As an example, a 63-year-old woman had severe aortic stenosis necessitating valve replacement. She also had
Predicting aortic valve prosthesis size
453
a history of peptic ulcer disease with a recent acute bleeding episode making the long-term use of anticoagulants unacceptable. Her prosthesis size was predicted to be 21 rom. Normally, in such a small aortic root, a disc-type prosthesis would have been used, but this would necessitate long-term warfarin sodium (Coumadin) therapy. With this prior knowledge, the operation was delayed until an Ionescu-Shiley valve could be obtained. Subsequently, a 21 mm IonescuShiley valve was implanted without difficulty and the patient has done well without requiring anticoagulants. We have found this technique to be accurate in all forms of aortic valvular disease, including Marfan aneurysms of the ascending aorta. We have gained sufficient confidence in this procedure that we did not recommend valve replacement in a patient with New York Heart Association Class II symptoms of pulmonary venous hypertension and a calculated aortic anulus of 19 mm. With this small anulus, the estimated postoperative transvalvular gradient would have been similar to her existing gradient. Thus an operation would have offered little benefit in this particular patient. We have identified two possible causes for imperfect measurement of valve size. First, in the presence of very heavy calcification, it may be impossible or unwise to remove all of the annular calcium. Thus, after debridement, the aortic anulus may be somewhat smaller than the angiographic estimate. Second, although the patient is consistently positioned in a 30 degree right anterior oblique projection, the intrathoracic aortic root position may vary somewhat from patient to patient. The radiographic appearance of two versus three sinuses is not always related to the presence of a bicuspid versus a tricuspid valve. Instead, the view may vary somewhat from the true perpendicular and cause variation in sinus visualization. Depending upon the degree and angle of variation, the angiographic estimate of anulus size may be greater or less than actual size. These inherent errors have not been great, however, not exceeding I to 2 mm. Thus, despite methodologic limitations, this technique has proved reliable and useful. Prospectively determining the valve size can facilitate the appropriate preoperative selection of valve type and/or operative technique. Although in many cases this may simplify the surgeon's task only minimally, in some cases it can make a significant contribution to achieving a satisfactory operative result. REFERENCES Rahimtoola SH: The problem of valve prosthesis-patient mismatch. Circulation 58:20-24, 1978
The Journal of
454 Babb, Parr, O'Neill
Thoracic and Cardiovascular Surgery
2 Konno S, Imai Y, Iida Y, Nakajima M, Tatsuno K: A new method for prosthetic valve replacement in congenital aortic stenosis associated with hypoplasia of the aortic valve ring. J THORAC CARDIOVASC SURG 70:909-917, 1975 3 Blank R, Pupello DF, Bessoe LN, Harrison EE, Sbar S: Method of managing the small aortic anulus during valve replacement. Ann Thorac Surg 22:356-361, 1976 4 Symbas PN, Ware RE, Hatcher CR Jr, Temesy-Armos PN: An operation for relief of severe left ventricular outflow tract obstruction. J THORAC CARDIOVASC SURG 71:245249, 1976 5 Rastan H, Abu-Aishah N, Rastan D, Heisig B, Koncz J,
Bjornstad PG, Beuren AJ: Results of aortoventriculoplasty in 21 consecutive patients with left ventricular outflow tract obstruction. J THORAC CARDIOVASC SURG 75:659-669, 1978 6 Rittenhouse EA, Sauvage LR, Stamm SJ, Mansfield PB, Hall DG, Herndon PS: Radical enlargement of the aortic root and outflow tract to allow valve replacement. Ann Thorac Surg 27:367-373, 1979 7 Kasser IS, Kennedy JW: Measurement of left ventricular volumes in man by single-plane cineangiocardiography. Invest Radiol 4:83-90, 1969