Two-dimensional echocardiography in dogs

J

THORAC CARDIOV ASC SURG

90:430-440, 1985

Two-dimensional echocardiography in dogs Variation of left ventricular mass, geometry, volume, and ejection fraction on cardiopulmonary bypass Quantitative two-dimensional echocardiography was evaluated in 39 open-chest dogs placed on cardiopulmonary bypass. The correlation coefficient of left ventricular end-diastolic volume against postmortem pressure-volumecurves was r = 0.89 to 0.93 (347 measurements in 15 dogs, 0 to 24 mm Hg), Ejection fraction was validated against roller pump flow and echo left ventricular end-diastolic volume (r = 0.83, n = 13~ Left ventricular mass in vivo was compared with postmortem left ventricular mass (r = 0.81 in 21 early studies, r = 0.91 in 10 later studies with updated equipment) and was found to increase with ischemic injury as well as cardiopulmonary bypass with hemodilution. Left ventricular mass increased (p < 0.001) from 119 ± 5 (standard error of the mean) to 138 ± 6 gm (n = 23) after 2lf2 hours on cardiopulmonary bypass and moderate hemodilution. With the addition of ischemic arrest, left ventricular mass increased from 119 ± 7 to 148 ± 11 gm (p < 0.01, n = 8), and myocardial water content increased by 2.0 % ± 0.4 %, which accounted for at least 65 % of the observed mass change. Mean left ventricular wall thickness increased from 13.8 to 15.5 mm (p = 0.02) after ischemia. Ventricular shape became more spherical with increasing left ventricular end-diastolic pressure. We conclude that two-dimensional echocardiography can be reliably used for accurate, serial measurements in physiological studies. The demonstrated variability in left ventricular mass is important, yet frequently ignored. Recognizing left ventricular mass changes may facilitate the detection of myocardial injury reflected as edema.

George B. Haasler, M.D., Pa'!l C. Rodigas, M.D., Robert H. Collins, M.D., Jeng Wei, M.D., Frances J. Meyer, M.D., Alan J. Spotnitz, M.D., and Henry M. Spotnitz, M.D.,*

New York, N. Y.

RySiOlogical studies during cardiac operations have been hindered by the lack of simple, yet adequate methods for serial measurement of left ventricular (LV) mass, end-diastolic volume (LVEDV), and ejection fraction.':' Ultrasound techniques described in experimental animals, although promising," 7 are often cumbersome," invasive," or measure only single cardiac dimensions. I, 2 In addition, some methods of LVEDV

From the Department of Surgery, Columbia University College of Physicians and Surgeons, New York, N. Y. Supported in part by the U.S. Public Health Service Grant No. HL-22894. Received for publication Sept. 4, 1984, Accepted for publication Nov. 14. 1984. Address for reprints: H. M. Spotnitz, M.D., Columbia University College of Physicians and Surgeons, 630 West 168th St., New York, N. Y. 10032. *This work was performed in part during the tenure of an Established Investigatorship of the American Heart Association.

430

measurement implicitly assume that ventricular mass remains constant,':' without experimental confirmation. In studies of ventricular compliance, the implantation of measuring devices may alter the very properties under examination." Previous studies from our laboratory described a simplified echocardiographic technique for in vivo LVEDV and LV mass but did not validate the method over the wide range of volumes needed for physiological experiments. 10, 11 In the current investigation, therefore, we have used two-dimensional echocardiography to study LV mass, LVEDV, and ejection fraction in dogs with variable preload on cardiopulmonary bypass (CPB). In addition, this study confirms physiological variation in LV mass and correlates that observation with changes in LV geometry and myocardial water content. Methods Thirty-nine adult dogs (weighing 20 to 35 kg) were anesthetized with sodium pentobarbitol (30 rug/kg)

Volume 90 Number 3 September, 1985

Echo LV mass and volume 431

Fig. 1. Two-dimensional echocardiograms of the canine left ventricle. S I A and S2A are perpendicular long-axis sections. SS is a short-axis section at the largest internal left ventricular diameter perpendicular to the long axis. The long axis connects the apex with the junction of the mitral and aortic valve anuli. LA, Left atrium. L JI, Left ventricle. RJI, Right ventricle. Millar is the tip of the left ventricular pressure catheter.

administered intravenously, intubated, and mechanically ventilated. Through a median sternotomy, an open pericardial well was constructed and filled with soundconducting gel which was maintained at a reproducible level above the right ventricular surface. 10. 11 The animals were heparinized and cannulated for CPB. At bypass, the right ventricle was continually decompressed and the vessels in the pulmonary hila were ligated. Left ventricular end-diastolic pressure (LVEDP) was varied by regulating the flow through a cannula in the left atrium. LVEDP was recorded via a piezoelectric Mikro-Tip catheter (Model PC-740, No.7 Fr., Millar Instruments, Inc., Houston, Tex.) inserted through the right superior pulmonary vein. LVEDP, aortic pressure, and the electrocardiogram were recorded on an optical oscillograph (Model DR-12, Electronics for Medicine, Honeywell Inc., Pleasantville, N. Y.). Two-dimensional echocardiograms were recorded on videotape (30 frames/sec) with a hand-held 3.5 MHz ultrasound transducer and a V-3000 ultrasonograph (Diasonics CardioImaging, Salt Lake City, Utah) for the first 21 studies. A V-3400 ultrasonograph with digital image processing and improved video recorders was employed for the remaining 18 studies. Two

perpendicular long-axisechocardiograms, which exclude the papillary muscles (SlA and S2A), and the shortaxis LV cross section of largest internal diameter (SS) were recorded (Fig. 1) at several stable LVEDPs. The transducer was positioned anteriorly for short-axis sections and at the LV apex for long-axis sections. These sections were defined in a previous study in which volume measurements were validated in dogs without bypass.'? Ultrasound contrast effect was created when necessary to facilitate visualization of the endocardium by rapid infusion of saline solution with microbubbles through a lumen in the Millar Mikro-Tip catheter. In 23 dogs (Group A), LV mass and LVEDV were serially studied during prolonged CPB (mean 2 hours) in the presence of moderate to severe crystalloid hemodilution (before death, hematocrit values 15% to 20%). In eight dogs (Group B), LV mass was studied before and after 45 minutes of normothermic global ischemia followed by reperfusion. Transmural myocardial samples for water content were obtained with a high-speed drill (19 to 50 mg) and were frozen with Freon/liquid nitrogen. After samples were dried to stable weight, water content (%) was determined from the formula:

The Journal

43 2

Haasler et al.

of

Thoracic and Cardiovascular Surgery

by heart rate to obtain stroke volume. (2) Echocardiographic sections recorded included SIA, S2A, and SS described earlier as well as short-axis sections at the mitral valve (SSI), tip of papillary muscles (SS2), midpapillary muscle (SS3), and base of papillary muscles (SS4). . After all in vivo studies, hearts were arrested with systemic administration of potassium chloride (KCL, 160 mEq), and postmortem pressure-volume curves were determined. Clamps were applied to the aorta just distal to the aortic valve and slightly on the atrial side of the mitral valve to isolate the left ventricle. The right ventricle was opened widely. Two No. 16 Fr. plastic catheters in the LV apex were used to infuse known increments of saline solution while pressure was simultaneously recorded. Postmortem data were obtained with gel in situ to approximate in vivo immersion. The left ventricle was then excised (free wall plus interventricular septum) and weighed by a laboratory technician. The true LV mass was not divulged to the investigator until all calculations were complete. Data analysis

Fig. 2. Representative echo sections and calculations of left ventricular mass and end-diastolic volume. The common axis for S 1A and S2A is the left ventricular short axis of maximum internal diameter beyond the mitral valve. Water content (%) = ([wet weightdry weightj/wet weight) X 100

(I)

The standard error of water content estimate in five control hearts on which multiple drill biopsies were performed was 0.25%, with an average total range of 0.96% in any individual heart. In six dogs (Group C), LV mass measurements were taken sequentially at low LVEDP, at progressively higher LVEDP, then again at lower LVEDP to separate the effects of chamber volume and time on mass. Total CPB periods were much shorter in this group, and all sections were completed within 1 to 1V2 hours of the initial section, compared with an average total CPB time of 2 to 2V2 hours otherwise. Crystalloid hemodilution was minimized. In 13 dogs (Group D), sections for ejection fraction were validated as follows: (1) LV output was controlled by directing all flow from a calibrated, occlusive roller pump through the left atrium. The output was divided

Echocardiographic volume and mass were determined on stopped-frame images at end-diastole (defmed by the peak of the electrocardiographic R wave) on a computerized light pen (Diasonics, Inc.) with an algorithm based on Simpson's rule. LVEDV and LV mass calculations. Two volume and mass calculations, based on paired perpendicular sections, were done from each set of long-axis (SIA and S2A) and short-axis (SS) sections. Section pairs were SIA-S2A and SIA-SS (Fig. 2). The endocardial planimetric line was drawn just inside the apparent endocardial echoes and the epicardium was "planimetered" at the outer edge of the epicardial echoes. Epicardial and endocardial planimetric measurements were superimposed at the aortic and mitral valves. In the short-axis cross-section, papillary muscles were "planimetered" as part of the chamber volume. LV mass was calculated by the formula: LV mass (gm) = (epicardial volume LVEDV) X 1.055

(2)

where 1.055 represents the specific gravity of myocardium. Results of the two calculations of mass and volume were averaged.'? LV mass estimates temporally close to the termination of bypass ("late" masses) were correlated with postmortem mass by linear regression and were also compared with masses near the initiation of bypass ("early" masses) with paired Student's t test.

Volume 90 Number 3

Echo LV mass and volume

September, 1985

433

Table I. Effects of global ischemia and reperfusion on left ventricular mass and myocardial water content Myocardial water (%)

Echo LV mass (gm) Experiment

Pre

126 136 101

I 2 3 4

5 6

I

Post

150 150

81.9 82.1

+22 6

81.5

83.5

+21

130

+14

108

+ 9

124

7

161

195

142

198

Mean

125 7

147 12

I

Post

+34 +56

+14

122

7

Pre

84.0 84.8 84.7 83.2 83.5 81.8 82.2 83.6

+24

116 99 117

8

SEM

I

Change

+

80.3 82.2 81.8 81.9

82.3 79.8

0.33

0.38

Predicted mass change* (gm)

+28 +23 +19

+ 9

+

8 +12 + 3 +17

+15 3

Legend: Pre, Early in bypass run. prior to global ischemia. Post, Late in bypass run, following global ischemia and reperfusion. SEM, Standard error of mean. Data presented arc for mass determinations at matched left ventricular end-diastolic volumes (see text).

* Predicted

mass change is based on change in myocardial water content.

Ejection fraction. For each short-axis section, the endocardial area at end-diastole (peak of R wave), and the visually smallest systolic image were calculated. Several cycles of each section were measured by planimetry and the areas averaged. Averaged areas and diameters of combinations of the four sections were considered. Ejection fraction (EF) was then calculated from the areas as:

= lOa x (diastolic area systolic area)jdiastolic area

EF (%)

(3)

LVEDV was also calculated from the SIA, S2A, and SS sections. Stroke volume calculated from the known roller pump flow and heart rate was divided by LVEDV to yield a "true" value for ejection fraction. Predicted mass change. On the basis of observed changes in myocardial water content, the change in LV mass could be predicted from the relation: MI (l - PI) = M2 (l - P2)

(4)

where Ml and M2 are early and late mass, respectively, and PI and P2 are the corresponding water contents expressed as decimals (i.e., 78% = 0.78). M (1 - P) is the dry mass. With determined values for PI, P2, and late echo mass M2, a theoretical Ml can be calculated. The difference (M2 - M 1) is the mass change in grams predicted from the change in water content (Table I). Ventricular diameters and wall thickness. Shortaxis inner and outer equatorial ventricular diameters and wall thickness were defined by planimetric measurement from two-dimensional images during the mass and volume calculations (Fig. 2). The average of these dimensionsfrom the S 1A and SS sections were analyzed in six normal dogs (Group C) and in eight dogs with

global ischemia (Group B) to define their relation to each other and to LVEDV. The relationship of the chamber volume long axis to the short axis was defined for any given LVEDP as the eccentricity ratio, E = Long axis/short axis. Statistical comparisons. Paired data were compared with Student's paired t test and linear regression. The effect of LVEDP and time on mass determinations, wall thickness, and diameter comparisons were carried out with the computerized Statistical Analysis System (SAS Institute, Inc., Cary, N. C.) by Robert Sciacca, Eng.Sci.D. Corrections were made to a common volume where necessary. Calculations involving simple statistical parameters were carried out with a Commodore-S61 statistical calculator (Commodore Business Machines, Palo Alto, Calif.). Results LV mass and volume. Fig. 3 correlates echocardiographic LVEDV with the volume infused in postmortem pressure-volume curves to reproduce the LVEDP measured in vivo. Observer 1, with 189 individual LVEDV determinations in nine dogs, and Observer 2, working independently with 158 LVED V determinations in six other dogs, achieved nearly identical regressions and correlations (r = 0.89, 0.93) between in vivoechocardiographic LVEDV and postmortem infusion volume. The in vivo range of LVEDP was 0 to 24 mm Hg. The correlation of echocardiographic LV mass determinations late in the bypass period with postmortem LV mass is shown in Fig. 4. The correlation coefficient for mass determinations with the V-3000 ultrasonograph (n = 21, Group A) versus postmortem LV mass was 0.81 for the SIA-SS calculation and was 0.67 for the averaged (SIA-S2A and SIA-SS) calculation. With the

The Journal of Thoracic and Cardiovascular Surgery

434 Haasler et al.

110 100

110

n : 189

90 yo

eo Echo LVEDV in Vivo (ml)

n : 158

100

.78x+ 6.7 (r=.89)

10

50

.. " V'

eo

.: ,

10

Echo LVEDV in Vivo (ml)

60

, line of Identity

90

"

,,'.

.,

y=.7Sx+ 9.3 (r =.931

~

60

50

40

40

30

30

20

20

30

40

50 60

70

90

80

10

100 110 120

20

30

40

50

60

10

80

90

100

110 120

Post Mortem Volume (ml)

Post Mortem Volume (ml)

Fig. 3. Echocardiographically determined left ventricular end-diastolic volume (LVEDV) versus the postmortem left ventricular volume evoking left ventricular filling pressure observed in vivo. Left panel, 189 individual volume determinations in nine dogs by Observer I. Right panel. 158 volume determinations in six different dogs by Observer 2, working independently.

V-3000 220

V-3400 220

n , 21

~ 200

2' 200

en en

en

•• •

..: 180

::;

~ ISO

::; ISO w II:

....

....

0 ::; '40

....

en

0

~

Mas5pm : :

•

120

c,

.831 Massecho + 28.128 (r

~

.8ll

....

en

0

140 120

c,

ioo 80

•

~ 180 ::;

0::

•

n, 10

• • ••

• Masspm~

1.073 Mass av -1.712 (r ~ .9ll

•

100

•

ioo

• 120 140 ISO 180 ECHO MASS IgJ (S 1 A/SS only)

200

220

80

120 140 lOa 160 180 200 AVERAGEO ECHO MASS IgJ (S1A/52A & 51A 5S)

220

Fig. 4. Echocardiographic left ventricular mass late in bypass period versus postmortem weight in dogs studied with the V3000 (left panel) and V3400 (right panel) ultrasonographs. Each point is the average of two to four individual mass determinations.

V-3400 ultrasonograph (n = 10) in later studies, the correlation coefficient improved to 0.91. The regression line with the V-3400 machine did not differ significantly from the line of identity. A representative example of the effect of chamber volume on LV mass measurement (from Group C) is shown in Fig. 5. In six dogs studied, a small direct relationship could be seen: The apparent LV mass was greater at higher chamber volumes. The mean slope of this mass variation as a function of LVEDV between 14 and 110 ml was 0.34. Comparative mass determinations

are therefore most accurately assessed at closely matched LVEDV with our methods. The changes in LV mass between early and late bypass, with and without global ischemia, are shown in Fig. 6. With prolonged cardiopulmonary bypass and crystalloid hemodilution without ischemia (Group A), mean LV mass increased by 19 gm (p < 0.(01), from 119 ± 5 gm (SE) to 138 ± 6 gm (n = 23). With bypass and global ischemia (Group B, n = 8), mean LV mass increased 29 gm after reperfusion from 119 ± 7 to 148 ± 11 gm (p < 0.01). Because LV mass measure-

Volume 90 Number 3

Echo LV mass and volume 435

September, 1985

150 140

...........................................................•.........................

130 120 LV ECHO

1 10

MASS

100

(g. )

•

•

••

•

• •• • •

•

• • •

•

•

90 80 70 60 50

Ok

10

20

30

40

50

60

70

80

90

LVEDV (mI.)

Fig. 5. Representative effect of left ventricular chamber volume on echocardiographic measurement of left ventricular (LV) mass (experiment 4, weight 136 gm). The large points represent individual mass measurements as described in Fig. 2. The dotted line represents the postmortem mass, 136 gm. The measured mass increases artifactualiy with left ventricular end-diastolic volume (LVEDV); the slope of this increase is approximately 0.34. The absolute value of the echo mass is somewhat smaller than the postmortem left ventricular mass in this example. EDP, End-diastolic pressure.

ment varied with LVEDV, data were also compared at matched LVEDV (Fig. 7). LV mass then increased only 22 ± 6 gm, from 125 ± 7 gm before ischemia to 147 ± 12 gm after ischemia and reperfusion (p < 0.01). Because the experiments with prolonged bypass were among our earliest studies, it was not possible to reassess the mass change in this group at matched volumes. Myocardial water content in Group B increased from 81.5% ± 0.33% to 83.5% ± 0.38% after ischemia (p < 0.01, Table I). The corresponding predicted mass change was 15 ± 3 gm, 65% of the change observed at matched LVEDV (22 ± 6 gm, Table I). Ejection fraction. The best correlation between echocardiographically determined ejection fraction and ejection fraction based on echo LVEDV and known roller pump output was obtained with the SS section alone (Fig. 8). The correlation coefficient was r = 0.83 and the regression line was not statistically different from the line of identity. Ventricular geometry and wall thickness. Results for effects of global ischemia on LV wall thickness (Group B) were corrected to a common LVEDV of 61 mI. Mean wall thickness increased significantly (p = 0.02) from 1.38 ± 0.04 to 1.55 ± 0.07 em after ischemia (Table 11). Fig. 9 illustrates the relationship of the short-axis endocardial and epicardial ventricular diameters as LVEDV was varied in two ventricles of different mass.

Endocardial diameter appears linearly related to epicardial diameter in both experiments, but the slope of this relationship is variable. Fig. 10 shows the LV eccentricity ratio (E) for six dogs (Group C) in four intervals of LVEDP. E decreases with increasing LVEDP, implying a change in LV geometry from ellipsoidal toward spherical. Mean E decreased from 2.4 ± 0.16 (SE) to 1.64 ± 0.17 over a range of LVEDP from 2 to 20 mm Hg. A representative example of the effects of global ischemia and reperfusion on E is shown in Fig. 11. In eight experiments (Group B), mean E after ischemia was not statistically different from that determined before ischemia. Discussion The results of the present study validate the echocardiographic determination on CPB of LVEDV, LV mass, and ejection fraction in relation to standards of known accuracy. The methods developed are shown to be accurate and reproducible, although serial comparisons of LV mass should be done at matched LVEDV. In addition, we find significant mass increases in the canine left ventricle subjected to CPB both with and without ischemic injury. The majority of this mass gain correlates directly with increased myocardial water content. LV mass appears maximally increased after ischemia and reperfusion.

The Journal of Thoracic and Cardiovascular Surgery

43 6 Haasler et al.

210 210 200

200

190

190

180

180

170

170

160 LV Echo Mass (9)

160

I

150 140

± "

130 120 110

±"

I

LV Echo

Mass (9)

150

130 120

100

90

90

80 70 60

TIE

110

100

// .> Early Bypass

Late Bypass

I ..

-± Sf

140

I ~::: ........

~

-~

80 70 60

L--

-'-

-:--,--~-----'

Early Bypass

late Bypass

(Pre Ischemia)

(Post Ischemia)

Fig. 6. Mass change on cardiopulmonary bypass determined by echocardiography with prolonged bypass and hemodilution (left panel) and with 45 minutes of normothermic global ischemia and reperfusion (right panel). Each point represents several left ventricular mass determinations, and left ventricular end-diastolic volume is not considered. Means and standard errors are indicated. Mass increases are statistically significant (p < 0.01). SE. Standard error.

The increase in LV mass also results in an increase in diastolic wall thickness at any given LVEDV. This implies that epicardial or midwall dimensions may not accurately be used to measure LVEDV in the presence of developing myocardial edema.':' Moreover, a variable relationship noted between endocardial and epicardial ventricular diameters from dog to dog implies that there is no way to directly predict endocardial dimensions from epicardial measurements alone. LV mass and volume. Both Wyatt and associates? and Eaton and colleagues" demonstrated accurate measurement of LV mass in dogs with two-dimensional echocardiography, but both methods appear too complex for studies of changing mass and compliance in animals placed on CPB. In contrast, our three-section method allows multiple measurements to be obtained during the course of a short experiment. Thus, compliance curves can be determined and effects of hemodynamic and pharmacologic interventions can also be examined. 10. II Despite advantages of the method discussed, words of caution are appropriate. Quantitative echocardiography is a laborious, learned skill. Proper alignment of longaxis sections with appropriate landmarks, especially the

true apex, is essential. Planimetry for accurate results requires excellent images, time, and patience. Subjective errors in planimetry can alter results, so that independent corroboration (e.g. versus postmortem LV weight) or independent planimetry by a blinded second observer may be needed. Volume and mass are most reproducible at an LVEDP range of 10 to 15 mm Hg. Identification of the lateral epicardial and septal borders in long-axis sections has been markedly improved with digital image processing. Nevertheless, echo masses and volumes tend to be somewhat smaller than postmortem values. As in other studies, we find echo mass more reproducible than LVEDV.JO·12 At very low LVEDP, the LV cavity assumes a very irregular shape. The tendency toward inaccurately aligned, artifactually small sections is greater at low LVEDP. We believe mass is less sensitive than volume to sectioning errors because of its larger absolute value (100 gm versus 12 to 60 ml). Also, the regular epicardial surface is easier to measure planimetrically than the irregular endocardium. The converse argument, that mass is defined by two echo boundaries and should be more susceptible to error than LVEDV, which involves only one boundary, has not proved valid.

Volume 90 Number 3

Echo LV mass and volume

September, 1985

210

E Fp =

200

1,119 EFs s - 3.416· ( r = ,83)

100

190

"Identity

80

180

437

170 160 Echo

Mass (9)

150 140 130 ±5E 120 110 100 90 80

I

~ ~

I,,,

20

o

20

40

60 E Fss

_

L--_-----=-----,--"'~

40

Early Bypass

Late Bypass

(Pre Ischemia)

(Post Ischemia)

Fig. 7. Similar to Fig. 6, effects of global ischemia and reperfusion on left ventricular mass are assessed only at closely matched left ventricular end-diastolic volume. The mean increase from before to after ischemia is significant (p < 0.01) but slightly less than in Fig. 6.

100

Fig. 8. Ejection fraction determined from pump flow, heart rate, and echo LVEDV (EFp) is correlated with ejection fraction based on SS section area change (EFss) .

Table II. Effect of global ischemia and reperfusion on left ventricular wall thickness LV wall thickness by two-dimensional echocardiography (cm)

Preischemia

Postischemia

Change

I 2 3 4 6 7 8

1.48 1.33 1.28 1.34 1.20 1.55 1.47

1.45 1.62 1.33 1.36 1.47 1.70 1.89

-0.03 +0.29 +0.05 +0,02 +0.27 +0.15 +0.42

Mean SEM

1.38 0,04

1.55 0.07

+0.17 0,06

Experiment

A price of our simplified method for mass is preload dependence, which does not affect other methods." This may relate to technical factors affecting long-axis sections, including echo scatter and beam width'<" The methods described in this paper are not useful during cardiac operation in man because of difficulty in obtaining accurate long-axis sections. Water content and mass change. Myocardial edema may reflect either hemodilution or loss of cell volume regulation because of injury. Salisbury and associates" and Cross and colleagues" documented whole heart mass increases to 50 gm after injury. Causes include crystalloid hemodilution.v":" ventricular fibrillation and distention," global ischemia/a, 21 and acute heart failure." In addition to the present results, we have reported no increase in LV mass with CPB if crystalloid hemodilution is minimized. Large mass gains with global ischemic injury persist, however, even if osmolarity is increased during bypass.22 Edema may worsen or impede recovery from myocardial injury and adversely affect myocardial mechaniCS,16.21 effects which may be reversible with mannito1.19, 21 Clinically applicable echo methods for quantifying edema might therefore lead to the institution of appropriate therapy." Nearly two thirds of the mass increase in the present

80

(%)

Legend: Data obtained from two-dimensional echocardiograms at matched left ventricular end-diastolic volumes.

study can be accounted for by water absorption. The remaining increase may reflect increased protein content because of capillary membrane leaks and interstitial hemorrhage. t7 Small errors in measurement of or sampling for water content may also have caused discrepancies between observed and predicted mass change. LV mass may also be influenced to a small extent by aortic pressure and heart rate.": 23. 24 Systematic error with our echocardiographic method is unlikely because of good correlation with postmortem mass and independent confirmation within our laboratory. 10. 22 We estimate that specific gravity should decrease only 0.0025 for each 1% increase in myocardial water content, which makes alteration in specific gravity an unlikely factor in errors in mass calculation (less than 1 gm per 100

The Journal of Thoracic and Cardiovascular Surgery

43 8 Haasler et al.

E ~

5.0

60

0

y;1.7996x-8.01 / (,=.92) /

,

Qj

.

.,

.:: ••

Qj 40 E (ll

i5

/

/

I

-.

.4

Qj 30

e c

@

Exp#6 LVM." 1609

Exp#3 LV Mass 1209

6.0

/

2.0

I

I

/

10

5.0

y=1.087-3.43 (,=.97)

4.0 30

/

,r

6.0

/

~ ~"

""

""

I

50

.. .

,

'. " • /

60

, 70

8.0

Fig. 9. Endocardial left ventricular short-axis diameter versus epicardial diameter in two representative experiments. The diameters plotted are averaged from sections SIA and SS. The slope of the relation is steeper in the lighter (120 gm) ventricle.

+ ±SD

15+ LVEDP (mmHg)

Fig. 10. Mean end-diastolic eccentricity ratio versus left ventricular end-diastolic pressure (L VEDP) in six experiments, Group C. Both eccentricity ratio and the corresponding end-diastolic pressure have been averaged. Brackets represent standard deviations.

gm of LV weight for each 4% change in water content). Wall thickness and geometry. Decreasing ventricular wall thickness with increasing ventricular volume has been demonstrated previously.v'>" Although epicardial instrumentation for LV volume measurement has been used for many years, I the role of mass change has frequently been ignored. Suga and Sagawa" demonstrated that although internal equatorial diameter was related to chamber volume by a cubic relationship, a linear approximation of this

relationship produced only a 7% error in volume assessment. Additional documentation of a linear relationship between end-diastolic internal diameter, as measured by epicardial and subendocardial ultrasound crystals," has promoted the use of epicardial sonomicrometry for the determination of LV chamber volume in a wide variety of settings." Our results suggest that calculations from epicardial dimensions without regard to changing wall thickness may overestimate LVEDV by 25 ml or more, the volume of increased LV mass. Although the epicardial and endocardial equatorial LV diameters are reasonably approximated by a linear relation over the range of LVEDV seen in the present study (Fig. 9), it is also apparent that this relation varies from dog to dog and must change with increased ventricular mass and wall thickness. We conclude that the endocardial dimension of the ventricle and its rate of change cannot be predicted from epicardial dimensions' unless LV mass is known. Epicardial measurements therefore should not be used as an index of endocardial dimensions for compliance or systolic performance studies under conditions associated with myocardial edema unless changes in mass and/or diastolic wall thickness are measured. Further investigation will be required to define the physiological significance of preload at endocardial, epicardial, and midwall radii when LV mass increases. We observed a gradual change in E with increasing LVEDV. This change implies that the heart becomes less ellipsoidal and more spherical as volume increases. The axis ratio plays a major role in calculations of wall stress" but its precise behavior over the physiological range of LVEDV has not been described. Our normal values for the major and minor ventricular axes in dogs

Volume 90

Echo LV mass and volume

Number 3 September, 1985

2.4 Experiment

o

6

•

2.2

4 39

Before Ischemia After Ischemia

2.0 .0

Eccentricity Ratio

0

0

0

1.8

0

•

1.6

..• • • •

0

1.4

00

0\ 0

• •

0

1.2

1.0

0

10

20

30

40 LVEDV

50

60

70

80

90

(mil

Fig. I I. Representative experiment measuring eccentricity ratio versus left ventricular end-diastolic volume (LVEDVj before and after 45 minutes of aortic cross-clamping and reperfusion. Eccentricity is plotted against end-diastolic volume to avoid compliance effects. There is no apparent shift in the postischemic values compared with preischemic controls.

agree with those of Ross and associates" and the canine E appears comparable to that derived from available human studies." The additional finding that E does not change after global ischemia and reperfusion suggests that effects of this injury are homogeneous from the point of view of ventricular isotropy. Ejection fraction. Our studies indicate that, in normal dogs, an ejection fraction from the widest internal cross-sectional area change is as good as that derived from algorithms which employ several cross sections'l-" Accuracy is increased by averaging planimetric measurements from several cycles. Although the accuracy of multiple cross-section techniques has been unequivocally demonstrated, precise multiple sections are difficult to obtain in the open chest because of exaggerated cardiac motion during systole. Our method for ejection fraction is clearly not valid in the presence of large segmental wall motion abnormalities, as in regional myocardial ischemia, where multiple sectioning planes would be more accurate and consideration of the apical contribution to ejection fraction has been fruitful.32 In our previous study," ejection fraction with the short-axis section correlated well with angiographic data but diverged from the line of identity. Improvement in the present study is related to use of the raw value for ejection fraction, rather than raising it to the 3/2 power as done previously. Conclusions Two-dimensional echocardiography is applicable to physiological studies that require serial measurement, yet it avoids distortion of ventricular properties. LV

mass varies with hemodilution or ischemic injury in dogs placed on CPB. Because myocardial mass variation correlates with changes in water content, serial LV mass measurements may be useful as an index of cardiac injury and recovery. Variation in mass can be a source of significant experimental error when epicardial dimensions alone are used to measure endocardial diameter or volume. We thank Mr. Peter Bloom and Mr. Anthony Cuffy for technical assistance vital to running the experiments. Figures were prepared by Ms. Linda White, Ms. Angela LaValle, and Mr. John Karapelou. Photographs are by Mr. Renald Von Muchow. Ms. Lucille Stahr assisted in developing the methodology for the water content determinations. Statistical consultation was provided by Dr. Robert Sciacca, Eng. Sc. D. The assistance of Mrs. Rosemary Marx, Mrs. DorothyLowry, Ms. Sallie Stadlen, Mr. David Kirschenbaum, Ms. Lucia Paronetto,and Ms. Sara Jane Suder in the preparationof this manuscript has also been greatly appreciated. REFERENCES Braunwald E, Frye RL, Ross J Jr: Studies on Starling's law of the heart. Determinantsof the relationship between left ventricular end-diastolic pressure and circumference. Circ Res 8:1254-1263, 1960 2 OlsenCO, Attarian DE, Jones RN, Hill RC, Sink JD, Lee KL, Wechsler AS: The coronary pressure-flow determinants of left ventricular compliance in dogs. Circ Res 49:856-865, 1981 3 Van Trigt P, Spray TL, Pasque MK, Peyton RB, Pellom GL, Christian CM, Fagraeus L, Wechsler AS: The influence of time on the response to dopamine after coronary artery bypass grafting. Assessment of left ven-

440

Haasler et af.

tricular performance and contractility using pressure/ dimension analyses. Ann Thorac Surg 35:3-13, 1983 4 Rankin JS, Arentzen CE, Ring WS, Edwards CH II, McHale PA, Anderson RW: The diastolic mechanical properties of the intact left ventricle. Fed Proc 39: 141-147, 1980 5 Chitwood WR, Hill RC, Sink JD, Olsen CO, Wechsler AS: The comparative effects of intermittent ischemic arrest and potassium cardioplegia on diastolic ventricular compliance in patients (abstr). Am J Cardiol 47:417, 1981 6 Eaton LW, Maughan WL, Shoukas AA, Weiss JL: Accurate volume determination in the isolated ejecting canine left ventricle by two-dimensional echocardiography. Circulation 60:320-326, 1979 7 Wyatt HL, Heng JK, Meerbaum S, Hestenes JD, Cobo JM, Davidson RM, Corday E: Cross sectional echocardiography. I. Analysis of mathematical models for quantifying the mass of the left ventricle in dogs. Circulation 60: 1104-1113, 1979 8 Spotnitz HM: Profound topical hypothermia (letter). J THoRAc CARDIOVASC SURG 74:814, 1977 9 Hefner LL, Coghlan HC, Jones WB, Reeves TJ: Distensibility of the dog left ventricle. Am J Physiol 201:97-101, 1962 10 Collins RH, Haasler GB, Krug JH Jr, Colt C, Spotnitz AJ, Martin EC, Spotnitz HM: Canine left ventricular volume and mass during thoracotomy by two-dimensional echocardiography. J Surg Res 33:294-304, 1982 II Haasler GH, Rodigas PC, Spotnitz HM: The absence of temperature effects on end-diastolic pressure-volume relations in the canine left ventricle. J THORAC CARDIOVASC SURG 83:878-890, 1982 12 Helak JW, Reicheck N: Quantitation of human LV mass and volume by two-dimensional echocardiography. In vitro anatomic validation. Circulation 63:1398-1407, 1981 13 Feigenbaum H: Echocardiography, ed 3, Philadelphia, 1981, Lea & Febiger, pp I-50, 110-115 14 Roelandt J, van Dorp WG, Bom N, Laird JD, Hugenholtz PG: Resolution problems in echocardiology. A source of interpretation errors. Am J Cardiol 37:256-262, 1976 15 Salisbury PF, Cross CE, Rieben PA: Distensibility and water content of heart muscle before and after injury. Circ Res 8:788-793, 1960 16 Cross CE, Rieben PA, Salisbury PF: Influence of coronary perfusion and myocardial edema on pressure-volume diagram of left ventricle. Am J Physiol 201:102-108, 1961 17 Salisbury PF, Cross CE, Katsuhara K, Rieben PA: Factors which initiate or influence edema in the isolated dog's heart. Circ Res 9:601-606, 1961 18 Erdmann AJ, Geffin GA, Barret LV, Buckley MJ, Austen WG, Lowenstein E: Increased myocardial water content with acute Ringer's lactate hemodilution in dogs (abstr). Circulation 49,50:Suppl 3:18, 1974 19 Foglia RP, Lazar HLL, Steed DL, Follette DM, Manga-

The Journal of Thoracic and Cardiovascular Surgery

naro AJ, Deland ED, Buckberg GD: Iatrogenic myocardial edema with crystalloid primes. Effects on left ventricular compliance, performance, and perfusion. Surg Forum 29:312-314,1979 20 Nelson RL, Goldstein SM, McConnell DH, Maloney JV Jr, Buckberg GD: Studies on the effects of hypothermia on regional myocardial blood flow and metabolism during cardiopulmonary bypass. V. Profound topical hypothermia during ischemia in arrested hearts. J THoRAc CARDIOVASC SURG 73:201-207, 1977 21 DiBona DR, Powell WJ: Quantitative correlation between cell swelling and necrosis in myocardial ischemia in dogs. Circ Res 47:653-664, 1980 22 Spotnitz WD, Clark MB, Rosenblum HM, Lazar HL, Haasler GB, Collins RH, Spotnitz AJ, Wong CYH, Spotnitz HM: Effect of cardiopulmonary bypass and global ischemia on human and canine left ventricular mass. Evidence for interspecies differences. Surgery 96:230-239, 1984 23 Salisbury PF, Cross CE, Rieben PA: Physiological factors influencing coronary blood volume in isolated dog hearts. Am J Physiol 200:633-666, 1960 24 Glick G, Williams JF Jr, Harrison DC, Morrow AG, Braunwald E: Cardiac dimensions in intact unanesthetized man. VI. Effect of changes in heart rate. J Appl Physiol 21:947-958, 1966 25 Ross J Jr, Sonnenblick EH, Covell JW, Kaiser GA, Spiro D: The architecture of the heart in systole and diastole. Technique of rapid fixation and analysis of left ventricular geometry. Circ Res 21:409-421, 1967 26 Mitchell JH, WildenthaI K, Mullins CB: Geometric studies of the left ventricle utilizing biplane cinefluorography. Fed Proc 28:1334-1343,1969 27 Sasayama S, Franklin D, Ross J Jr, Kemper WS, McKown D: Dynamic changes in left ventricular wall thickness and their use in analyzing cardiac function in the conscious dog. Am J Cardiol 38:870-879, 1976 28 Suga H, Sagawa K: Assessment of absolute volume from diameter of the intact canine left ventricular cavity. J Appl Physiol 36:496-499, 1974 29 Janz RF: Effect of chamber eccentricity on equatorial fiber stress during systole. Fed Proc 39:183-187, 1980 30 Falsetti HL, Mates RE, Crant C, Greene DG, Bunnell IL: Left ventricular wall stress calculated from one-plane cineangiography. Circ Res 26:71-83, 1970 31 Beranek I, Moore R, Kim S, Formanek A, Amplatz K: Comparison of ejection fraction calculated by nine different volume calculation methods. Radiology 120:553-556, 1976 32 Quinones MA, Waggoner AD, Reduto LA, Nelson JG, Young JB, Winters WL Jr, Ribeiro LG, Miller RR: A new, simplified and accurate method for determining ejection fraction with two-dimensional echocardiography. Circulation 64:744-753,1981