- Email: [email protected]
A new use for M-mode echocardiography in detecting left ventricular diastolic dysfunction in coronary artery disease
A New Use for M-Mode Echocardiography in Detecting left Ventricular Diastolic Dysfunctionin Coronary Artery Disease WILLIAM
E. LAWSON, MD, EDWARD J. BROWN, Jr., MD, RITA D. SWINFORD, CLAIRE PROCTOR, RDMS, and PETER F. COHN, MD
Regional left ventricular (LV) diastolic function affects the global rate and pattern of LV filling. These changes may be detected by changes in the magnitude and timing of the increase in LV basal diameter during diastole. Because M-mode echocardiography possesses the high temporal and spatial resolution to detect such abnormalities, a group of 8 normal control subjects were compared with a group of 12 patients with coronary artery disease (CAD) to determine differences in the rate and timing of ventricular filling. The CAD patients had lower rates of fast filling expansion than the control subjects. The proportion of LV diastolic expansion during fast filling
BS,
was lower. During atrial systole the increased rate of LV expansion was not significantly higher in the CAD patients, but the proportion of diastolic expansion occurring with atrial systole was increased. These changes may reflect a decrease in the rate and magnitude of early diastolic filling in the noncompliant ventricle and an increased reliance on active atrial transport. Thus, CAD alters the rate and pattern of LV filling. Changes in LV diameter as measured by M-mode echocardiography may be useful in detecting altered patterns of LV diastolic filling and identifying patients with CAD. (Am J Cardiol 1986;58:21 O-21 3)
A
bnormalities of contraction and relaxation are known to coexist in patients with coronary artery disease (CAD]. Interest has focused on systolic contractile function, both in the diagnosis and therapy of CAD. However, in many patients, regional systolic function at rest may be normal. Even CAD patients with demonstrable abnormalities of regional contractile function may maintain normal global systolic function by compensatory hyperkinesia of normal segments. By comparison, the diastolic adaptive changes in global and regional function in response to ischemia appear less fully compensatory, and changes in left ventricular (LV) pressure, volume and geometry are detectable even at rest.1-3 Isovolumic diastolic abnormalities have been well demonstrated in CAD patients.4 However, regional filling abnormalities secondary to CAD may
also be expected to alter LV geometry and both global and regional filling in a predictable fashion. We evaluated the ability of the M-mode echocardiogram to reveal changes in diastolic LV dimensions at the base and to detect abnormal LV diastolic function secondary to either local or distant ischemia or infarction.
Methods Patients: The study population consisted of 20 patients who had cardiac catheterization and technically good M-mode echocardiograms. Of the 20 patients referred for diagnosis of CAD, 8 had normal LV angiographic and hemodynamic studies and normal coronary angiographic findings. The other 12 patients had coronary angiograms showing significant (at least 70% diameter) narrowing in at least 1 major coronary artery. Patients with identifiable valvular heart disease, conduction abnormalities such as atrioventricular, nodal or bundle branch blocks, or pericardial effusions were excluded. Echocardiography: With the patients positioned in the left lateral decubitus position, M-mode and 2-D echocardiograms were performed using an ATL Mark
From the Cardiology Division, Department of Medicine, State University of New York Health Sciences Center, Stony Brook, New York. Manuscript received January Z&1986; revised manuscript received April 3, 1986, accepted April 7,1986. Address for reprints: William E. Lawson, MD, Cardiology Division, SUNY Health Sciences Center T-17-020, Stony Brook, New York 11794. 210
August
V with a ~-MHZ transducer. The M-line cursor of the 2-D image was used to select an M-mode chord perpendicular to the ventricular septum and posterior LV wall of the left ventricle at the level of the tips of the mitral valve leaflets. With the patient breathing shallowly, a high-quality, rapid-speed (75 or 100 mm/s) Mmode recording of the LV cavity for at least 3 successive cardiac cycles was obtained. The endocardial edges of the ventricular septum and posterior wall bordering the LV cavity, seen on the M-mode echocardiogram, were traced and digitized using a Numonics graphics analyzer. The graphics analyzer allowed the LV internal diameter to be measured continuously over the cardiac cycle and automatically derived a velocity plot based on the rate of change in LV diameter (Fig. 1). LV filling was separated into phases of fast filling, slow filling and filling with atria1 systole as determined by inflection points on the plot of LV internal diameter during diastole. LV diameter was measured at minimal and maximal dimensions and the cube method was used to calculate a filling volume.5 The relative proportion of LV filling occurring during fast filling, slow filling and atria1 systolic phases were calculated using the initial and maximal diameters of each phase. A volume was computed for each phase of filling using the cube method.5 The fraction of total diastolic filling during each phase was derived by dividing the volume calculated for each phase by the calculated volume of total diastolic filling. The fraction of LV filling occurring during fast filling was designated the fast filling fraction. The contribution of atria1 systole was designated the fraction with atria1 systole. Statistical analysis: Values are mean f standard error of the mean. Comparison of values in normal vs CAD disease patients was done using a pooled Student t test. A p value <0.05 was considered significant.
Results Patient characteristics: Patients in the normal group were 36 to 62 years old [mean 48 f 3) and those in the CAD group were 34 to 72 years old [mean 51 f 4). There was 1 woman in the normal group and none in the CAD group. Average heart rate was 57 f 4 beats/min in the normal group and 65 f 2 beats/min in the CAD group. LV end-diastolic diameter as determined by M-mode echocardiography was 56 f 1 mm in the CAD group and 51 f 4 mm in the normal group. Mean fractional shortening was 0.36 f 0.03 and mean wall thickness was 10 f 0.54 mm in the CAD group and 0.42 f 0.03 and 6 f 0.36 mm, respectively, in the control group. None of these values were significantly different between the groups except for mean wall thickness (p <0.005). Echocardiographic filling parameters: Table I and Figures 2 and 3 show echocardiographic parameters for the normal and CAD groups. Significant differences were noted in the pattern of LV filling between the CAD group and the normal group. In the CAD group the peak rate of early diastolic expansion during fast filling was decreased. Most important, all 7 patients with peak velocity during fast filling of less than
1, i986
THE
AMERICAN
JOURNAL
OF CARDIOLOGY
Vo!ume
57
211
VELOCITY
FIGURE 1. Analysis of an M-mode echocardiogram recorded at the level of the left ventricular (LV) cavity in a normal subject (so/id line) and a patient with coronary artery disease (dashed line). A simultaneous electrocardiogram (EKG) defines each cardiac cycle. The graphics analyzer measures changes in cavity diameter to derive a difference plot and the rate of change in cavity size to yield a velocity plot. The fast filling period is defined as the initial early diastolic rapid increase in LV diameter. The fast filling phase was followed by an abrupt transition to a slow filling phase characterized by a slower, relatively constant increase in LV diameter. The slow increase in LV diameter is terminated at enddiastole by a rapid increase in LV diameter caused by atrial systole. The average peak velocity (rate of change in LV cavity diameter at the base) of at least 3 successive cardiac cycles was obtained for fast filling and atrial systole phases of diastole from the Numonics graphics analyzer velocity plot. The peak rate of change in LV cavity dimension during early diastolic fast filling was designated as peak fast filling. The peak rate of increase in LV dimension with atrial systole was designated as peak atrial systole.
125 mm/s had CAD (Fig. 2). Similarly, all 7 patients with a fast filling fraction of less than 53% had CAD (Fig. 3). During atria1 systole, the peak rate of LV diastolic expansion was not significantly increased in the CAD patients. However, atria1 systole contributed an increased fraction of the total diastolic filling. As seen in Figure 3, all 5 patients with fractional filling during atria1 systole greater than 0.27 had CAD.
212
ABNORMAL
VENTRICULAR
FILLING
TABLE Control
I Comparison of Values for Filling Subjects and Patients with Coronary
Fast filling fraction p <0.025 Peak fast filling (mm/s) p <0.05 Atrial systole filling fraction p <0.005 Peak filling with atrial systole NS = coronary
DISEASE
Variables in Normal Artery Disease
Normal
Variable
CAD
IN CORONARY
artery
0.62
(mm/s)
disease;
Abnormal Diastolic Measurement
CAD
f 0.02
0.51
152zt6 0.18
TABLE II Sensitivity, Specificity Abnormal Diastolic Measurements
f 0.02
124 f f 0.01
0.25
47 f 5
10
60 f 6
NS = not significant.
By defining limits based on our normal population, each M-mode echo measure [either alone or in combination) was sensitive and specific in identifying patients with CAD (Table II). Catheterization findings were reviewed in our CAD patients to characterize our population. The influence of global LV ejection fraction, presence and location of regional dysfunction and extent and site of CAD on diastolic function was .reviewed, though in most instances the small subgroup sample sizes precluded firm conclusions. Global LV ejection fraction was normal (20.55) in 9 CAD patients, with only mild reductions in the other 3 [II&, 0.45, 0.47). Regional wall motion abnormalities were absent in 3 CAD patients, distant to the M-mode sampling volume in 6 patients and involved the M-mode sampling r
Predictive Accuracy
0.67
0.88
0.75
0.75
0.88
0.80
0.67
0.63
0.65
0.75
0.88
0.80
1 .oo
0.50
0.80
I - control o- coronary
l
QO-
o- coronary artery
p
of
volume in 3 patients. Compared with the other CAD patients, the 3 patients with wall motion abnormalities in the area of M-mode sampling showed the most severe decrease in the peak rate of early diastolic filling. This subgroup also showed a smaller increase in both the peak rate and the fraction of filling with atria1 systole. The diastolic abnormalities observed in the 6 CAD patients with distant wall motion abnormalities were similar to those for the entire CAD group. Five patients had Wessel, 3 had .&vessel and 4 had l-vessel CAD. The vessels with at least 70% narrowings were also equally distributed between left anterior descending (8 patients), left circumflex (8 patients]
I
220-
Values
Specificity
Sensitivity
Peak fast filling (50 mm/s) Filling fraction with atrial systole (>0.20) Use of all the above criteria
f 0.01
240-
and Predictive
disease
disease
200-
artery
sol
0
lSO-
0 !
‘so140-
I
70-
:
60-
0
PEAK ,20
0 . ..L ____ o-. 0
VELOCITY omhc)
% FILLING
5o -
p
I
lOO-
i
40i 0
so-
30-
so2040lo-
ZOI
I
FAST FlLL!NB
I
I
I
1
I
1
I
ATRIAL SYSTDLE FAST FILLING
FIGURE 2. Diastolic filling velocity. The peak rate of change in left ventricular cavity size (peak velocity) was measured during the early fast filling portion of diastole and with atrial systole. Individual values of the normal and coronary artery disease patients are shown, together with the comparable mean f standard error of the mean. Significance levels are shown for each comparison. NS = not significant.
ATRIAL
SYSTOLE
FIGURE 3. Fractional diastolic filling. The percentage contributions of the early fast filling portion of diastole and atrial systole to total diastolic filling were expressed as percent filling. The individual values of the normal and coronary artery disease patients are shown, together with comparable mean f standard error of the mean. Significance levels are shown for each comparison.
August
and right coronary arteries (9 patients]. No apparent differences were seen in diastolic function when subgrouped according to involved vessels.
Discussion Previous radionuclide ventriculography studies have shown characteristic diastolic findings in patients with CAD. Bonow et all found rest LV diastolic abnormalities in peak filling rate and time to peak filling in 82% of patients with CAD in his study. Diastolic abnormalities were present despite normal systolic LV function, without evidence of ischemia and independent of the extent of CAD. They theorized that transient ischemia produces more lasting, and thus more readily detectable, effects on LV relaxation than it does on LV contraction. Echocardiography was used in dogs to study the effects of transient myocardial ischemia on LV diastolic function. Ischemia caused by temporary occlusion of the left circumflex or left anterior descending coronary artery results in a generalized reduction in both the rapid and slow filling phases and an increase in both the systolic and diastolic dimensions. In a study using sonomicrometers in chronically instrumented dogs, Smalling et al3 examined the effect of local ischemia on regional function. Segmental systolic dysfunction in the distribution of the left circumflex artery caused by partial occlusion of the vessel was almost completely offset by augmented systolic function in the distribution of the nonischemic left anterior descending artery. By contrast, the diastolic impairment of relaxation in the left circumflex distribution was minimally offset by enhanced anteroseptal thinning. This study showed that regional systolic dysfunction may be completely offset by augmented function elsewhere. Abnormal regional rapid diastolic filling was, by contrast, minimally compensated for by facilitated relaxation of normal segments, leading to detectable alterations in global diastolic function. This study indicated that abnormalities of diastolic function, particularly on a global basis, should prove more sensitive than abnormalities of systolic function in detecting CAD in humans. Gibson and Brown6 reported that in patients without segmental LV dysfunction, digitized M-mode echocardiograms yield estimates of peak systolic and diastolic wall motion similar to those obtained by ventriculography. Measures of myocardial relaxation obtained with digitized M-mode echocardiography have been shown to be as reliable as radionuclide angiography in identifying non-CAD patients with normal regional and global systolic function but abnormal diastolic function.7 Our study demonstrates that M-mode echocardiography remains a useful technique in the CAD patient. CAD characteristically altered diastolic filling as manifested by changes in LV basal cavity dimensions. The normal rapid early diastolic increase in basal LV dimension was decreased both absolutely and as a fraction of total diastolic expansion. Alterations in the diastolic rate and pattern of LV filling as manifested by
1, 1986
THE AMERICAN
JOURNAL
OF CARDIOLOGY
Volume
57
213
changes in LV internal diameter occurred despite normal M-mode regional function in the CAD patients. M-mode echocardiographic analysis was thus sensitive to distant alterations in regional LV function secondary to CAD. During atria1 systole, our CAD patients showed no significant change in the peak rate of expansion in LV cavity size, but atria1 systole was responsible for an increased proportion of the total increase in diastolic dimension. This is in accord with studies by Carroll8 and Matsudag and their colleagues showing active atria1 transport to be of increased importance in maintaining LV filling in the acutely ischemic and previously infarcted left ventricle. This increased contribution of the left atrium to LV filling is apparently mediated by the Frank-Starling mechanism and results from elevated left atria1 pressure. Our findings are in accord with radionuclide and experimental studies suggesting that CAD alters the LV diastolic filling pattern, even with normal systolic function. BryhnlO and Bahler et all1 showed that age, heart rate, fractional shortening and LV hypertrophy may also all affect LV relaxation. Further work is needed to evaluate the impact of each of these variables individually and in combinations on LV function. Clinical implications: Changes in LV diastolic function are detectable by digitized M-mode echocardiography. These changes appear to be sensitive markers of CAD and are found in CAD patients with normal systolic function on M-mode echocardiography. The M-mode echocardiographic findings in CAD patients suggest that ischemic heart disease impairs initial diastolic filling of the left ventricle, causing atria1 systole to assume increased importance in maintaining enddiastolic volume and cardiac output.
References 1. Bonow RO, Bacharach SL, Green MV, Kent KM, Rosing DR, Lipson LC, Leon MB, Epstein SE. Impaired left ventricular diastolic filling in patients with coronary artery disease: assessment with radionuclide angiography. Circulation 1983;64:315-323. 3. Reduto LA, Wickemeyer WJ, Young JB, DelVentura LA, Reid JW, Glaeser DH, Ouinones MA, Miller RR. Left ventricular diastolic performance at rest and during exercise in patients with coronary artery disease. Circulation 1981;63:1228-1237. 3. Smalling RW, Kelley KO, Kirkeeide RL, Gould KL. Comparison of early systolic and early diastolic regional function during regional ischemia in a chronicallv instrumented canine model. JACC 1985:2:2&+2t%. 4. LawsonWE, Proctor CE. Echocardiograpy. In: Cohn, PF, ed. Diagnosis and Therapy of Coronary Artery Disease. 2nd ed. Boston: Martinus Niihoff. ,, 1985:169-189.
5. Pombo JF, Troy BL, Russell RO Jr. Left ventricular volume and ejection fractions by echocardiography. Circulation 1971;49:489-499. 6. Gibson DG, Brown DJ. Measurement of peak rates of left ventricular wall movement in man. Comparison of echocardiography with angiography. Br Heart J 1975;37:677-683. 7. Bryhn M. Abnormal left ventricular filling in patients with sustained myocardial relaxation: assessment of diastolic parameters using radionuclide angiography and echocardiography. CJin CardioJ 1984;7:639-646. 8. Carroll ID, Hess OM. Hirzel HO, Kravenbuehl HP. Dvnamics of left ventricular fiJJing at rest and during exercise. Circulation i983;68:59167,’ 6. Matsuda Y, Toma Y, Ogawa H, Matsuzaki M, Katayama K, Fujii T, Yoshino F. Moritani K, Kumada T, Kusukawa R. Importance of left atrial function in patients with myocardial infarction. Circulation 1983~67:ii66-571,. 10. Brvhn M. Echocardiographic assessment of left ventricular diastolic function ina normal population and a group of patients with myocardial hypertrophy. CJin CardioJ 1984;7:935-349. 11. Bahler RC, Vrobel TR, Martin P. The relation of heart rate and shortening fraction to echocardiographic indexes of left ventricular relaxation in normal subjects. JACC 1985;2:926-933.
E. LAWSON, MD, EDWARD J. BROWN, Jr., MD, RITA D. SWINFORD, CLAIRE PROCTOR, RDMS, and PETER F. COHN, MD
Regional left ventricular (LV) diastolic function affects the global rate and pattern of LV filling. These changes may be detected by changes in the magnitude and timing of the increase in LV basal diameter during diastole. Because M-mode echocardiography possesses the high temporal and spatial resolution to detect such abnormalities, a group of 8 normal control subjects were compared with a group of 12 patients with coronary artery disease (CAD) to determine differences in the rate and timing of ventricular filling. The CAD patients had lower rates of fast filling expansion than the control subjects. The proportion of LV diastolic expansion during fast filling
BS,
was lower. During atrial systole the increased rate of LV expansion was not significantly higher in the CAD patients, but the proportion of diastolic expansion occurring with atrial systole was increased. These changes may reflect a decrease in the rate and magnitude of early diastolic filling in the noncompliant ventricle and an increased reliance on active atrial transport. Thus, CAD alters the rate and pattern of LV filling. Changes in LV diameter as measured by M-mode echocardiography may be useful in detecting altered patterns of LV diastolic filling and identifying patients with CAD. (Am J Cardiol 1986;58:21 O-21 3)
A
bnormalities of contraction and relaxation are known to coexist in patients with coronary artery disease (CAD]. Interest has focused on systolic contractile function, both in the diagnosis and therapy of CAD. However, in many patients, regional systolic function at rest may be normal. Even CAD patients with demonstrable abnormalities of regional contractile function may maintain normal global systolic function by compensatory hyperkinesia of normal segments. By comparison, the diastolic adaptive changes in global and regional function in response to ischemia appear less fully compensatory, and changes in left ventricular (LV) pressure, volume and geometry are detectable even at rest.1-3 Isovolumic diastolic abnormalities have been well demonstrated in CAD patients.4 However, regional filling abnormalities secondary to CAD may
also be expected to alter LV geometry and both global and regional filling in a predictable fashion. We evaluated the ability of the M-mode echocardiogram to reveal changes in diastolic LV dimensions at the base and to detect abnormal LV diastolic function secondary to either local or distant ischemia or infarction.
Methods Patients: The study population consisted of 20 patients who had cardiac catheterization and technically good M-mode echocardiograms. Of the 20 patients referred for diagnosis of CAD, 8 had normal LV angiographic and hemodynamic studies and normal coronary angiographic findings. The other 12 patients had coronary angiograms showing significant (at least 70% diameter) narrowing in at least 1 major coronary artery. Patients with identifiable valvular heart disease, conduction abnormalities such as atrioventricular, nodal or bundle branch blocks, or pericardial effusions were excluded. Echocardiography: With the patients positioned in the left lateral decubitus position, M-mode and 2-D echocardiograms were performed using an ATL Mark
From the Cardiology Division, Department of Medicine, State University of New York Health Sciences Center, Stony Brook, New York. Manuscript received January Z&1986; revised manuscript received April 3, 1986, accepted April 7,1986. Address for reprints: William E. Lawson, MD, Cardiology Division, SUNY Health Sciences Center T-17-020, Stony Brook, New York 11794. 210
August
V with a ~-MHZ transducer. The M-line cursor of the 2-D image was used to select an M-mode chord perpendicular to the ventricular septum and posterior LV wall of the left ventricle at the level of the tips of the mitral valve leaflets. With the patient breathing shallowly, a high-quality, rapid-speed (75 or 100 mm/s) Mmode recording of the LV cavity for at least 3 successive cardiac cycles was obtained. The endocardial edges of the ventricular septum and posterior wall bordering the LV cavity, seen on the M-mode echocardiogram, were traced and digitized using a Numonics graphics analyzer. The graphics analyzer allowed the LV internal diameter to be measured continuously over the cardiac cycle and automatically derived a velocity plot based on the rate of change in LV diameter (Fig. 1). LV filling was separated into phases of fast filling, slow filling and filling with atria1 systole as determined by inflection points on the plot of LV internal diameter during diastole. LV diameter was measured at minimal and maximal dimensions and the cube method was used to calculate a filling volume.5 The relative proportion of LV filling occurring during fast filling, slow filling and atria1 systolic phases were calculated using the initial and maximal diameters of each phase. A volume was computed for each phase of filling using the cube method.5 The fraction of total diastolic filling during each phase was derived by dividing the volume calculated for each phase by the calculated volume of total diastolic filling. The fraction of LV filling occurring during fast filling was designated the fast filling fraction. The contribution of atria1 systole was designated the fraction with atria1 systole. Statistical analysis: Values are mean f standard error of the mean. Comparison of values in normal vs CAD disease patients was done using a pooled Student t test. A p value <0.05 was considered significant.
Results Patient characteristics: Patients in the normal group were 36 to 62 years old [mean 48 f 3) and those in the CAD group were 34 to 72 years old [mean 51 f 4). There was 1 woman in the normal group and none in the CAD group. Average heart rate was 57 f 4 beats/min in the normal group and 65 f 2 beats/min in the CAD group. LV end-diastolic diameter as determined by M-mode echocardiography was 56 f 1 mm in the CAD group and 51 f 4 mm in the normal group. Mean fractional shortening was 0.36 f 0.03 and mean wall thickness was 10 f 0.54 mm in the CAD group and 0.42 f 0.03 and 6 f 0.36 mm, respectively, in the control group. None of these values were significantly different between the groups except for mean wall thickness (p <0.005). Echocardiographic filling parameters: Table I and Figures 2 and 3 show echocardiographic parameters for the normal and CAD groups. Significant differences were noted in the pattern of LV filling between the CAD group and the normal group. In the CAD group the peak rate of early diastolic expansion during fast filling was decreased. Most important, all 7 patients with peak velocity during fast filling of less than
1, i986
THE
AMERICAN
JOURNAL
OF CARDIOLOGY
Vo!ume
57
211
VELOCITY
FIGURE 1. Analysis of an M-mode echocardiogram recorded at the level of the left ventricular (LV) cavity in a normal subject (so/id line) and a patient with coronary artery disease (dashed line). A simultaneous electrocardiogram (EKG) defines each cardiac cycle. The graphics analyzer measures changes in cavity diameter to derive a difference plot and the rate of change in cavity size to yield a velocity plot. The fast filling period is defined as the initial early diastolic rapid increase in LV diameter. The fast filling phase was followed by an abrupt transition to a slow filling phase characterized by a slower, relatively constant increase in LV diameter. The slow increase in LV diameter is terminated at enddiastole by a rapid increase in LV diameter caused by atrial systole. The average peak velocity (rate of change in LV cavity diameter at the base) of at least 3 successive cardiac cycles was obtained for fast filling and atrial systole phases of diastole from the Numonics graphics analyzer velocity plot. The peak rate of change in LV cavity dimension during early diastolic fast filling was designated as peak fast filling. The peak rate of increase in LV dimension with atrial systole was designated as peak atrial systole.
125 mm/s had CAD (Fig. 2). Similarly, all 7 patients with a fast filling fraction of less than 53% had CAD (Fig. 3). During atria1 systole, the peak rate of LV diastolic expansion was not significantly increased in the CAD patients. However, atria1 systole contributed an increased fraction of the total diastolic filling. As seen in Figure 3, all 5 patients with fractional filling during atria1 systole greater than 0.27 had CAD.
212
ABNORMAL
VENTRICULAR
FILLING
TABLE Control
I Comparison of Values for Filling Subjects and Patients with Coronary
Fast filling fraction p <0.025 Peak fast filling (mm/s) p <0.05 Atrial systole filling fraction p <0.005 Peak filling with atrial systole NS = coronary
DISEASE
Variables in Normal Artery Disease
Normal
Variable
CAD
IN CORONARY
artery
0.62
(mm/s)
disease;
Abnormal Diastolic Measurement
CAD
f 0.02
0.51
152zt6 0.18
TABLE II Sensitivity, Specificity Abnormal Diastolic Measurements
f 0.02
124 f f 0.01
0.25
47 f 5
10
60 f 6
NS = not significant.
By defining limits based on our normal population, each M-mode echo measure [either alone or in combination) was sensitive and specific in identifying patients with CAD (Table II). Catheterization findings were reviewed in our CAD patients to characterize our population. The influence of global LV ejection fraction, presence and location of regional dysfunction and extent and site of CAD on diastolic function was .reviewed, though in most instances the small subgroup sample sizes precluded firm conclusions. Global LV ejection fraction was normal (20.55) in 9 CAD patients, with only mild reductions in the other 3 [II&, 0.45, 0.47). Regional wall motion abnormalities were absent in 3 CAD patients, distant to the M-mode sampling volume in 6 patients and involved the M-mode sampling r
Predictive Accuracy
0.67
0.88
0.75
0.75
0.88
0.80
0.67
0.63
0.65
0.75
0.88
0.80
1 .oo
0.50
0.80
I - control o- coronary
l
QO-
o- coronary artery
p
of
volume in 3 patients. Compared with the other CAD patients, the 3 patients with wall motion abnormalities in the area of M-mode sampling showed the most severe decrease in the peak rate of early diastolic filling. This subgroup also showed a smaller increase in both the peak rate and the fraction of filling with atria1 systole. The diastolic abnormalities observed in the 6 CAD patients with distant wall motion abnormalities were similar to those for the entire CAD group. Five patients had Wessel, 3 had .&vessel and 4 had l-vessel CAD. The vessels with at least 70% narrowings were also equally distributed between left anterior descending (8 patients), left circumflex (8 patients]
I
220-
Values
Specificity
Sensitivity
Peak fast filling (
f 0.01
240-
and Predictive
disease
disease
200-
artery
sol
0
lSO-
0 !
‘so140-
I
70-
:
60-
0
PEAK ,20
0 . ..L ____ o-. 0
VELOCITY omhc)
% FILLING
5o -
p
I
lOO-
i
40i 0
so-
30-
so2040lo-
ZOI
I
FAST FlLL!NB
I
I
I
1
I
1
I
ATRIAL SYSTDLE FAST FILLING
FIGURE 2. Diastolic filling velocity. The peak rate of change in left ventricular cavity size (peak velocity) was measured during the early fast filling portion of diastole and with atrial systole. Individual values of the normal and coronary artery disease patients are shown, together with the comparable mean f standard error of the mean. Significance levels are shown for each comparison. NS = not significant.
ATRIAL
SYSTOLE
FIGURE 3. Fractional diastolic filling. The percentage contributions of the early fast filling portion of diastole and atrial systole to total diastolic filling were expressed as percent filling. The individual values of the normal and coronary artery disease patients are shown, together with comparable mean f standard error of the mean. Significance levels are shown for each comparison.
August
and right coronary arteries (9 patients]. No apparent differences were seen in diastolic function when subgrouped according to involved vessels.
Discussion Previous radionuclide ventriculography studies have shown characteristic diastolic findings in patients with CAD. Bonow et all found rest LV diastolic abnormalities in peak filling rate and time to peak filling in 82% of patients with CAD in his study. Diastolic abnormalities were present despite normal systolic LV function, without evidence of ischemia and independent of the extent of CAD. They theorized that transient ischemia produces more lasting, and thus more readily detectable, effects on LV relaxation than it does on LV contraction. Echocardiography was used in dogs to study the effects of transient myocardial ischemia on LV diastolic function. Ischemia caused by temporary occlusion of the left circumflex or left anterior descending coronary artery results in a generalized reduction in both the rapid and slow filling phases and an increase in both the systolic and diastolic dimensions. In a study using sonomicrometers in chronically instrumented dogs, Smalling et al3 examined the effect of local ischemia on regional function. Segmental systolic dysfunction in the distribution of the left circumflex artery caused by partial occlusion of the vessel was almost completely offset by augmented systolic function in the distribution of the nonischemic left anterior descending artery. By contrast, the diastolic impairment of relaxation in the left circumflex distribution was minimally offset by enhanced anteroseptal thinning. This study showed that regional systolic dysfunction may be completely offset by augmented function elsewhere. Abnormal regional rapid diastolic filling was, by contrast, minimally compensated for by facilitated relaxation of normal segments, leading to detectable alterations in global diastolic function. This study indicated that abnormalities of diastolic function, particularly on a global basis, should prove more sensitive than abnormalities of systolic function in detecting CAD in humans. Gibson and Brown6 reported that in patients without segmental LV dysfunction, digitized M-mode echocardiograms yield estimates of peak systolic and diastolic wall motion similar to those obtained by ventriculography. Measures of myocardial relaxation obtained with digitized M-mode echocardiography have been shown to be as reliable as radionuclide angiography in identifying non-CAD patients with normal regional and global systolic function but abnormal diastolic function.7 Our study demonstrates that M-mode echocardiography remains a useful technique in the CAD patient. CAD characteristically altered diastolic filling as manifested by changes in LV basal cavity dimensions. The normal rapid early diastolic increase in basal LV dimension was decreased both absolutely and as a fraction of total diastolic expansion. Alterations in the diastolic rate and pattern of LV filling as manifested by
1, 1986
THE AMERICAN
JOURNAL
OF CARDIOLOGY
Volume
57
213
changes in LV internal diameter occurred despite normal M-mode regional function in the CAD patients. M-mode echocardiographic analysis was thus sensitive to distant alterations in regional LV function secondary to CAD. During atria1 systole, our CAD patients showed no significant change in the peak rate of expansion in LV cavity size, but atria1 systole was responsible for an increased proportion of the total increase in diastolic dimension. This is in accord with studies by Carroll8 and Matsudag and their colleagues showing active atria1 transport to be of increased importance in maintaining LV filling in the acutely ischemic and previously infarcted left ventricle. This increased contribution of the left atrium to LV filling is apparently mediated by the Frank-Starling mechanism and results from elevated left atria1 pressure. Our findings are in accord with radionuclide and experimental studies suggesting that CAD alters the LV diastolic filling pattern, even with normal systolic function. BryhnlO and Bahler et all1 showed that age, heart rate, fractional shortening and LV hypertrophy may also all affect LV relaxation. Further work is needed to evaluate the impact of each of these variables individually and in combinations on LV function. Clinical implications: Changes in LV diastolic function are detectable by digitized M-mode echocardiography. These changes appear to be sensitive markers of CAD and are found in CAD patients with normal systolic function on M-mode echocardiography. The M-mode echocardiographic findings in CAD patients suggest that ischemic heart disease impairs initial diastolic filling of the left ventricle, causing atria1 systole to assume increased importance in maintaining enddiastolic volume and cardiac output.
References 1. Bonow RO, Bacharach SL, Green MV, Kent KM, Rosing DR, Lipson LC, Leon MB, Epstein SE. Impaired left ventricular diastolic filling in patients with coronary artery disease: assessment with radionuclide angiography. Circulation 1983;64:315-323. 3. Reduto LA, Wickemeyer WJ, Young JB, DelVentura LA, Reid JW, Glaeser DH, Ouinones MA, Miller RR. Left ventricular diastolic performance at rest and during exercise in patients with coronary artery disease. Circulation 1981;63:1228-1237. 3. Smalling RW, Kelley KO, Kirkeeide RL, Gould KL. Comparison of early systolic and early diastolic regional function during regional ischemia in a chronicallv instrumented canine model. JACC 1985:2:2&+2t%. 4. LawsonWE, Proctor CE. Echocardiograpy. In: Cohn, PF, ed. Diagnosis and Therapy of Coronary Artery Disease. 2nd ed. Boston: Martinus Niihoff. ,, 1985:169-189.
5. Pombo JF, Troy BL, Russell RO Jr. Left ventricular volume and ejection fractions by echocardiography. Circulation 1971;49:489-499. 6. Gibson DG, Brown DJ. Measurement of peak rates of left ventricular wall movement in man. Comparison of echocardiography with angiography. Br Heart J 1975;37:677-683. 7. Bryhn M. Abnormal left ventricular filling in patients with sustained myocardial relaxation: assessment of diastolic parameters using radionuclide angiography and echocardiography. CJin CardioJ 1984;7:639-646. 8. Carroll ID, Hess OM. Hirzel HO, Kravenbuehl HP. Dvnamics of left ventricular fiJJing at rest and during exercise. Circulation i983;68:59167,’ 6. Matsuda Y, Toma Y, Ogawa H, Matsuzaki M, Katayama K, Fujii T, Yoshino F. Moritani K, Kumada T, Kusukawa R. Importance of left atrial function in patients with myocardial infarction. Circulation 1983~67:ii66-571,. 10. Brvhn M. Echocardiographic assessment of left ventricular diastolic function ina normal population and a group of patients with myocardial hypertrophy. CJin CardioJ 1984;7:935-349. 11. Bahler RC, Vrobel TR, Martin P. The relation of heart rate and shortening fraction to echocardiographic indexes of left ventricular relaxation in normal subjects. JACC 1985;2:926-933.