Adult human valve dimensions and their surgical significance

VALVULAR HEART DISEASE

Adult Human Valve Dimensionsand Their Surgical Significance STEVE WESTABY,

MD,*

ROBERT B. KARP, MD, EUGENE H. BLACKSTONE,

MD,

and SANFORD P. BISHOP, DVM, PhD

After cardiac valve replacement, some patients may show little improvement in functional status, in part because their prosthesis Is restrictive. Guidelines were sought for valve replacement from measurements of valve circumference and calculated circularized orifice area in 160 postmortem hearts from adults with and without congestive cardiac failure. Multivariate analysis was used to relate valve area to age, sex, height, weight, body surface area and cardiac failure. Only sex and cardiac failure were significantly related to valve area. Body surface area and other variables were poorly related to valve area. The mean (f standard deviation) circularized orifice area for adult male (M) and female (F) heart valves in the absence of cardiac failure were: Aortic,

M 4.61 f 1.30, F 3.73 f 0.96; pulmonary, M 4.66 f 1.25, F 4.32 f 1.03; mitral, M 6.70 f 2.06, F 6.94 f 1.41; and tricuspid, M 11.9 f 2.72, F 9.33 f 2.02. In cardiac failure, atrioventricular valves enlarge (p <0.004). Guided by these dimensions, the surgeon can aim to insert a prosthesis of appropriate size. Comparison of these sizes with the manufacturer’s calculated area for current prostheses shows that most mechanical valves and bioprostheses are potentially restrictive at rest. Improved prosthestic design, valve repair whenever possible, and anular enlargement procedures would be required to eliminate this size disparity. (Am J Cardiol 1964;53:552-556)

Methods

In 1963, Rowlat et all reported the morphologic features and dimensional analysis of the heart valves of infants and children. They found that valve orifice sizes in any heart were related to body surface area, body weight and age. We believe that during valve repair or replacement in adults, however, these factors correlate poorly with assessed valve dimensions. Therefore, we studied valve orifice dimensions in postmortem adult hearts with and without congestive cardiac failure to determine their mean size and relation to demographic and other variables.

Valve orifice circumferences were measured in 160 human adult hearts from the autopsy records of the Specialized Center for Research in Ischemic Heart Disease at the University of Alabama in Birmingham. From these same records, demographic data, including height, weight, age, sex and race, were obtained, and body surface area was calculated. A detailed medical history and pathologic findings were available for each heart. Patients with primary valve disease were not included in the study. There were 111 men and 49 women. Body surface area ranged from 1.22 to 2.84 m2 (mean 1.84). Age at death ranged from 16 to 83 years (58 f 13, mean f standard deviation [SD]). For men, the mean age at death was 57 f 13 years (range 16 to 83) and for females 61 f 14 years (range 23 to 82). Height for the entire group ranged from 115 to 190 cm (mean 168) and weight from 36 to 151 kg (mean 72). Heart weights varied from 220 to 890 g (mean 493 f 124). Of the 160 patients, 14 died from trauma or cerebral hemorrhage and had no history of heart disease. Of the remaining 146 patients, 122 had ischemic heart disease, 81 myocardial infarction, 44 hypertension and 13 cardiomyopathy. Fortyseven of 111 men and 16 of 49 women had congestive cardiac failure.

From the Departments of Surgery and Pathology, The University of Alabama in Birmingham, The Medical Center, Birmingham, Alabama. This study was supported in part by lschemic Heart Disease SCAR Grant 5P50HL17667, National Heart, Lung, and Blood Institute, Bethesda, Maryland. Manuscript received August 3, 1983; revised manuscript received October 20, 1983, accepted October 25, 1983. Present address: Steve We&by, MD, Cardiovascular Surgery, Hammersmith Special Health Authority, Hammersmith Hospital, Du Cane Road, London W12 OHS, England. Address for reprints: Robert B. Karp, MD, Chief, Cardiac Surgery, Department of Surgery, University of Chicago Medical Center, Box 152, 5841 South Maryland Avenue, Chicago, Illinois 60637. l

552

February 1. 1984

TABLE I

THE AMERICAN JOURNAL OF CARDIOLOGY Volume 53

Mean Values for Mitral and Tricuspid Dimensions with and without Congestive Heart Failure (CHF) Without CHF Mitral

Cir;;wzence Male Female

553

(cm)*

With CHF

Triscuspid

9.79 f 1.23 10.15 f 1.24 9.11 f 0.86

Argvg;;) Male Female

7.76 f 8.33 f 6.66 f

1.93 1.98 1.26

Di;yet--(mm) Male Female

31.1 f 3.9 32.3 f 3.9 29.0 f 2.7

11.63 f 11.95 f 10.40 f

1.39 1.26 1.06

10.56 f 2.59 11.50 f 2.46 8.75 f 1.74 36.4 f 4.4 38.0 f 4.0 33.2 f 3.3

Mitral

10.46 f 10 68 f 9.71 f

Overall Tricuspid

12.22 f 12.48 f 11.46 f

1.26 1.26 1.00

8.83 f 2.12 9.21 f 2.13 7.58 f 1.55

Mitral

1.43 1.44 1.14

10.05 f 1.28 10.38 f 1.27 9.29 f 0.94

12.05 f 2.87 12.56 f 2.94 10.54 f 2.07

8.17 f 2.07 8.70 f 2.08 6.94 f 1.41

38.9 f 4.6 39.7 f 4.6 36.4 f 3.6

33.3 f 4.0 34.0 f 4.0 30.9 f 3.1

Tricuspid 11.74 f 12.12 f 10.77 f

1.45 1.36 1.18

11.14 f 2.79 11.90 f 2.72 9.33 f 2.02

32.0 f 4.0 33.0 f 4.0 29.6 f 3.0

37.3 f 4.6 38.8 f 4.3 34.3 f 3.6

Primary valve measurement. Area and diameter derived by calculation. Values are mean f standard deviation. l

TABLE ii

Mean Values for Aortic and Pulmonary Valve Dimensions Aortic

Circumference (cm)* Overall Male Female Area (cm2) Overall Male Female Diameter (mm) Overall Male Female

TABLE iii

Pulmonary

Multivariate Analysis of Mitral Valve Adult Heart Dissections (n = 160)

Variable Circumference

7.28 f 0.92 7.50 f 1.04 6.80 f 0.89

7.63 f 0.93 7.77 f 0.98 7.32 f 0.86

4.56 f 1.12 4.81 f 1.30 3.73 f 0.98

4.71 f 4.88 f 4.32 f

23.20 f 3.3 23.80 f 3.3 21.60 f 2.8

1.16 1.25 1.03

24.30 f 3.0 24.70 f 3.1 23.30 f 2.7

Primary valve measurement. Area and diameter derived by calculation. Values are mean f standard deviation. l

Intercept Male Conaestive failure

Results The dimensions for mitral and tricuspid valves are listed in Table I, and those for aortic and pulmonary valves in Table II. In hearts of patients without congestive heart failure, the measured circumference of the mitral valve ring was 9.79 f 1.23 cm, and the calculated valve area 7.76 f 1.93 cm2. For the tricuspid area the measured circumference was 11.63 f 1.39 cm and the valve area 10.56 f 2.59 cm2. Adult aortic valves measured 7.78 f 0.92 cm, with a calculated area of 4.56 f 1.12 cm2, and pulmonary valve circumference was 7.63 f 0.93 cm, with area 4.71 f 1.16 cm2. Corresponding derived valve diameters were mitral 31.1 f 3.9 mm,

p Value

(SD)*

(r* = 0.20; SD of regression

= 1.157)

9.125 f 0.178 1.021 f 0.20 0.5503 f 0.190

<0.0001 <0.0001 0.004

Area (r* = 0.20; SD of regression = 1.863) Intercept Male Conaestive failure

6.670 f 0.29 1.653 f 0.33 0.8924 f 0.3 1

<0.0001 <0.0001 0.004

Four significant figures presented for aid in using equation. SD = standard deviation; male: 1 if male, 0 if female; congestive failure: 1 if failure, 0 if no failure. l

TABLE IV

Valve measurements: The hearts were obtained soon after death and fixed with 10% formalin and examined by the pathologists of the Specialized Center of Research in Ischemic Heart Disease. The circumferences of the valve orifices were measured by opening the valve ring, laying it flat and measuring to the nearest millimeter with a ruler. Measurements of the valve orifices were made at the level of the insertion of the valve leaflets or cusps (corresponding to the positions of usual prosthetic valve insertion, at the base of the cusp insertion). Valve circumference was then used to calculate the circularized valve orifice area and diameter. (Circularized valve dimensions are appropriate for clinical comparison even for the atrioventricular positions, because insertion of a valve prosthesis also circularizes the anulus.)

Coefficient

Multivariate Analysis of Tricuspid Valve Adult Heart Dissections (n = 160)

Variable Circumference Intercept Male Congestive failure

Coefficient

(SD)*

(r* = 0.25; SD of regression

p Value = 1.27 1)

10.55 f 0.194 1.346 f 0.22 0.6664 f 0.21

<0.0001
Area (r* = 0.24; SD of regression = 2.456) Intercept Male Congestive failure

8.921 f 0.37 2.491 f 0.42 1.268 f 0.40

<0.0001 <0.0001 0.002

Four significant figures presented for aid in using equation. Abbreviations as in Table Ill. l

TABLE V

Multivariate Analysis of Aortic Valve Adult Heart Dissections (n = 160)

Variable Circumference Intercept Age (natural logarithm) Male X In(age)

Coefficient

(SD)*

p Value

(r* = 0.24; SD of regression 0.9167) 1.491 f 1.08 1.300 f 0.27 0.2025 f 0.039

0.17
Area (r* = 0.22: SD of rearession = 1.126) Intercept Age (natural logarithm) Male X In(age)

-2.280 f 1.33 1.474 f 0.33 0.2369 f 0.048

0.09 <0.0001 <0.0001

Four significant figures presented for aid in using equation. In = natural logarithm; other abbreviations as in Table III. l

ADULT HUMAN VALVE DIMENSIONS

554

NDWAL

VALVE

D I&ETERS

tricuspid 36.4 f 4.4 mm, aortic 23.2 f 3.3 mm and pulmonary 24.3 f 3.0 mm. Multivariate analysis indicates that mitral andtricuspid measurements were larger in men and in patients with congestive heart failure (Tables III and IV). For aortic and pulmonary valves, valve size was larger in men and in older subjects (Tables V and VI). These differences are small and account for very little (<25%) of the variability of heart valve size from person to person. Valve dimensions in these adult patients are only weakly correlated with body surface area (Table VII). Although all correlations are positive (that is, the larger the body surface area the larger the valve size), they account for little of the variability among subjects. All other variables, including height, body weight, race, ischemic heart disease and hypertension were not directly related to heart valve sizes. The overall mean calculated native valve orifice areas for aortic, mitral and tricuspid positions are compared with the calculated manufacturers’ prosthetic valve orifice areas for prostheses of equivalent diameter in Table VIII.

TRICUSPID

MITRAL

PULMONARY

c

50.0

AoRTiC

...,‘..,.I...

0.5

1.0

1.5

BODY SURFACE

, 2.5

2.0

AREA (meI

FIGURE1. Nomogam obtained from data of Rowtat et al1 relating mean valve diameter to body surface area in subjects aged 1 day to 17 years.

TABLE

Vi

Discussion Previous workers have identified the limitations of valve measurement at autopsy.2-5 The position in the cardiac cycle in which the ventricular wall is set by rigor mortis is determined by the amount of blood that distends the ventricular cavity after the last ventricular contraction. This can affect the dimensions of the atrioventricular but not the semilunar valves, and this is particularly so if valve diameter is measured by inserting a cone into the anulus, but is less important when the anulus is opened and laid flat as in this study. Several investigators1y5yg have performed detailed analysis of the relations of both atrioventricular and semilunar valves in children and have found that all valves can be related to one another and to body surface area, weight and age. A plot derived from the data of Rowlat appears in Figure 1, and Blackstone et al6 have used this information as guidelines for repair in tetralogy of Fallot and other congenital anomalies. We sought similar guidelines for sizing reparative procedures and for valve replacement in adults. However, although the shape of Rowlat’s curve might indicate correlation between body surface area, height and weight with projection into adult life, our studies indicate that for the adult, after the period of rapid growth, these correlations are weak. Previous investigators7*8 have reported

Multivariate Analysis of Pulmonary Valve Adult Heart bissections (n = 160)

Variable

Coefficient

Circumference

(SD)*

(r2 = 0.01; SD of regression

Intercept Age (natural logarithm) Male X In(age)

p Value = 0.9301)

4.522 f 1.10 0.6649 f 0.27 0.1276 f 0.040

<0.0001 0.01 0.0015

Area (r* = 0.09: SD of rearession = 1.161) Intercept Age (natural logarithm) Male X ln(age)

0.9645 f 1.37 0.8170 f 0.34 0.1581 f 0.049

0.5 0.02 0.002

Four significant figures presented for aid in using equation. Abbreviations as in Table Ill to V. l

TABLE Vii

Correlation Coefficients of Valve Size Versus Body Surface Area

Valves

Correlation Coefficient

Aortic Pulmonary Mitral Tricuspid

0.01 0.10 0.12 0.19

l

p value for difference

TABLE VIII

W(%) 0.02 1.0 1.5 3.7

p Value* ::;5 0.15 0.02

from 0.

Relation Between Mean Native Valve Area and Prosthetic Valve Area For an Anuius of Equivalent Diameter Manufacturer’s

Mean Native Value Diameter (mm)

Area (cm*)

Calculated Orifice Area (cm*)

S-E

C-E

B-S

I-S

SJM

Aortic Mitral

ZZ

4.8 8.2

1.67 3.24

3.50 6.98

2.50 4.60

2.96 6.79

2.55 5.15

Tricuspid

37

11.1

3.66

7.34

4.60

6.79

5.15

B-S = Bjdrk-Shiley;

C-E = Carpentier-Edwards;

I-S = lonescu-Shiley;

S-E Starr-Edwards;

(all size 23 prosthesis) (size 33 prosthesis apart from SJM, size 31) (largest available prosthesis for each valve)

SJM = St. Jude Medical.

February 1. 1984

wide individual variation in adult valve orifice diameters in postmortem hearts, and at operation, the size of the valve anulus appears to be unrelated to the size of the patient. Our findings show that for adults, valve sizes in general are not importantly related to body habitus. We and others8 found a continued but small increase in size of aortic and pulmonary valves with age, and Eckner et al9 found a correlation between valve size and height. However, in Eckner’s study it is unclear how many of his specimens were adults (older than 20 years). For both atrioventricular and semilunar valves, gender is more importantly related to valve size than other factors. Men have valves that are 10% larger than women in each position. We found that for the atrioventricular valves, congestive cardiac failure caused dilation of the valve orifices, but for practical purposes this difference was not large. The aortic and pulmonary valves were not affected. Hutchins and Araya7 studied hearts with a variety of cardiac diseases and could not establish significant differences in valve diameter for different conditions with or without congestive cardiac failure, although Bulkley and RobertslO found the mitral an&s greatly dilated in patients with floppy mitral valves and mitral regurgitation (autopsy measurements). Because valve size is unrelated to body surface area in adults, and the relation between men and women is consistent, it is realistic to use a mean figure for each valve dimension as a guide for valve surgery, bearing in mind that the expected size for the aortic and pulmonary valves increases slightly with age. Although there is no correlation in size between the atrioventricular and semilunar valves, the pulmonary valve diameter consistently exceeds the aortic diameter by a factor of 1.1, and the tricuspid valve is consistently larger than the mitral valve. A true comparison of mitral and tricuspid valves by area is difficult because the calculated valve area for the tricuspid valve is circularized and the tricuspid valve is oval during life and, therefore, of less area. In dealing with prosthetic valves, however, the comparison is valid because insertion of a prosthesis also circularizes the anulus. Because valve anulus size cannot be predicted from morphometric data, it is also unlikely that patientprosthetic mismatch is wholly avoidable using these data in prosthesis selection, as Rahimtoola suggests.” Because there are no grounds to relate valve area to body surface area, the so-called “valve area index” for prediction of the hemodynamic and functional adequacy of a prosthetic device in an individual patient is inappropriate, and mismatch may occur when the effective prosthetic valve area is less than that of the native human valve. Although there are no absolute figures for any individual patient, mean native valve size may provide better predictive guidelines. For instance, the substantial discrepancy between native and prosthetic valve orifice size is particularly striking for larger orifice areas, and in the triscuspid position all mechanical prostheses reduce the expected area by at least 50%. Clinical evidence from our institution supports this hypothesis, in that many patients who are in New York

THE AMERICAN

JOURNAL OF CARDIOLOGY

Volume 53

555

Heart Association functional class I or II after tricuspid or multiple valve replacement continue to require diuretic therapy for right heart failure. Many of these patients have a Starr-Edwards prosthesis whose orifice area is one-third that of an average adult tricuspid valve. Similarly, McIntosh et all2 reported an increase in the right atriomventricular mean diastolic gradient after tricuspid valve replacement with porcine heterografts. In a smaller-than-average aortic root, most prostheses, and particularly mechanical valves, are associated with gradients. The mean diameter of the male aortic root is 23 mm and the area is 4.56 cm2, compared with a Bjdrk-Shiley 23-mm valve area of 2.5 cm2 (calculated from manufacturer’s dimensions). This average aortic root is as likely to belong to a man who weighs 60 kg as to a man who weighs 120 kg. Because the average anulus size is much greater than that required for resting flow, a 23-mm prosthesis may be adequate for men of both sizes at rest. However, a skier who weighs 120 kg with a fixed valve orifice of 2.5 cm2 may well manifest evidence of prosthesis-patient mismatch during exercise. Usually, the reduction in effective valve orifice area after insertion of a valve prosthesis is mild to moderate in severity and of no immediate clinical significance. However, a patient may be hemodynamically and symptomatically worse after valve replacement, and this is especially troublesome because the effective prosthetic valve area may be further reduced by tissue ingrowth and endotheliazation so that the device becomes frankly stenotic. I3 When evaluating the late results of valve replacement (on survival, functional status and ventricular function), this should be kept in mind because poor results in some patients may result from the delayed effects of moderate to severe prosthetic stenosis, rather than ventricular dysfunction.14 In practice, the size of a prosthesis that can be inserted is usually determined by the size of the native valve anulus. The relations between valve size and body habitus have not been clearly defined for adults, and especially for adults with heart disease. Without a reliable guide to the optimal prosthetic size, it is not surprising that prosthesis-patient mismatch may produce disappointing functional results after an otherwise satisfactory valve replacement.ll Because the normal native valve area is much in excess of that required for resting flow, a restrictive prosthesis may result in suboptimal hemodynamics only on exercise, and it is therefore likely that the frequency of this problem is underestimated. When a smaller than predicted valve size is encountered at operation, especially in a young and active patient, perhaps greater effort should be made to enlarge the anulus rather than accepting a small prosthesis, which may be restrictive during exercise. In selecting the type of prosthesis, orifice area as well as durability and need for anticoagulation should be considered. Acknowledgment: We greatly appreciate the help of Mary Wirt in the preparation of this manuscript, and of Rob Brown for transcription of Rowlat’s data to the nomogram.

556

ADULT HUMAN VALVE DIMENSIONS

References 1. Rowlat UF, Rimaidi MJA, Lev J. The quantitative anatomy of the normal child’s heart. Pediatr Ciin North Am 1963;10:499-588. 2. Mallory FB, Wright JH. Pathological Technique: A Practical Manual for Workers in Pathological Histology and Bacteriology, including Directions for the Performance of Autopsies and for Clinical Diagnosis by Laboratory Methods. Philadelphia: WB Saunders, 1901. 3. Chapman CB. On the study of the heart-A comment on autopsy techniques. Arch Intern Med 1964;113:318-322. 4. Roberts WL. Examining the precordium and the heart. Chest 1970;57: 567-571. 5. De La Cruz MV, Anseiini A, Romero A, Monroe A. A Qualitative and auantitative study of the ventricles and great vessels of nor&i children. Am l&art J 1960:60:675-690. 6. Blackstone EH, Kirklin JW, Paclfico AD. Decision-making in repair of tetralogy of Fallot based on intraoperative measurements of pulmonary arterial outflow tract. J Thorac Cardiovasc Surg 1979;77:526-532.

7. Hulchins AM, Araya OA. Measurement of cardiac size, chamber volumes, and valve orifices at autopsy. Johns Hopkins Med J 1973;133:96-106. 6. Davies MJ. The aoftic valve; the mitral valve. In: Crawford T, ed. Pathology of Cardiac Valves. London-Boston: Butterworths, 1980:1-81, 62-99. 9. Eckner FAO, Brown BA, Davldson DL, Giagov S. Dimensions of normal human hearts. Arch Pathoi 1969;88:497-507. 10. Bulkley BH, Roberts WC. Dilation of the mitral annulus. A rare cause of mitral regurgitation. Am J Med 1975;59:457-463. 11. Rahimtoola SH. The problem of valve prosthesis-patient mismatch. Circulation 1978;58:20-24. 12. McIntosh CL, Mkhaleis LL, Mwow AG, itscotlz SB, Redwood MI, Epstein SE. Atrioventricular valve replacement with the Hancock porcine xenogaft: a five-year clinical experience. Surgery 1975;78:768-775. 13. Yoganathan AP, Corcoran WH, Harrison EC, Carl JR. The Bjork-Shiiey aortic prosthesis-flow characteristics, thrombus formation and tissue overgrowth. Circulation 1978;58:70-76. 14. Baxley WA, Karp RB, Dye LE. Evaluation of symptomatic patients with aortic valve prostheses (abstr). Circulation 1981;64:Suppl IV:IV-311.