- Email: [email protected]
Optimal determination of right ventricular filling dynamics in systemic hypertension
Optimal determination of right ventricular filling dynamics in systemic hypertension Antonio Cittadini, MD, Serafino Fazio, MD, a Hinrik StrSmer, MD, Alberto Cuocolo, MD, a Domenico Sabatini, MD, a Massimo Imbriaco, MD, a Luigi Saccfi, MD, a and Pamela S. Douglas, MD
Boston, Mass., and Naples, Italy To determine the optimal method of normalizing peak filling rate (PFR) determinations and apply it to the assessment of right ventricular (RV) and left ventricular (LV) filling characteristics and their interactions, 41 subjects with hypertension and 40 matched normals underwent echo-Doppler and nuclear study. Conventional normalization of PFR to end-diastolic volume (EDV) yielded poor correlations between nuclear- and echo-derived PFR (RV, r= 0.34; LV, r= 0.42), whereas nuclear and echo PFR normalized to stroke volume (SV) were closely correlated (RV, r = 0.87; LV, r = 0.92). Further, use of PFR normalized to SV revealed a close relation between RV and LV filling characteristics. Multivariate analysis confirmed that, in contrast to normalization to EDV or early to late filling-velocity ratios (E/A), peak filling rate normalized to SV was independent of ejection fraction and heart rate. In addition, RV filling impairment was related to LV filling impairment, and the effects of hypertension eliminated the independent influence of age on both LV and RV filling. In conclusion, normalization of PFR to SV may be preferable to use of EDV or E/A in evaluating RV and LV filling dynamics. (AM HEART J 1995;130:1074-82.)
Abnormal right ventricular (RV) filling dynamics have been described in a variety of cardiovascular diseases, often associated with significant clinical consequences. TM However, the best method for assessing RV filling is unknown. Conventional assessment ofventricular filling rates using normalization to end-diastolic volume (EDV) is difficult to determine for the irregularly shaped and poorly imaged right ventricle, and even in the left ventricle yields From the Charles A. Dana Research Institute and the Harvard-Thorndike Laboratory, Department of Medicine (Cardiovascular Division) of Beth Israel Hospital and Harvard Medical School, and the aDepartment of Internal Medicine of Federico II University Medical School Supported in part by a grant from ConsigHo Nazionale delle Ricerche, Italy (A.C.) Received for publication April 6, 1995; accepted May 24, 1995. Reprint requests: Pamela S. Douglas, MD, Cardiovascular Division, Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215. Copyright © 1995 by Mosby-Year Book, Inc. 0002-8703/95/$5.00 + 0 4/1/66515
1074
only modest correlations between nuclear and echocardiographic methods. 5, 6 To eliminate the influence of chamber volume, two studies have suggested normalization of Doppler flow velocities to tricuspid stroke volume (SV), 7, s but no scintigraphic validation of this index has been performed. In part because of these methodologic difficulties, little information is available concerning RV filling dynamics in systemic hypertension, in spite of evidence t h a t the right ventricle may be affected in left ventricular (LV) pressure overload states. 9-11 Specifically, no study to date has examined the determinants of filling or has taken into account the impact of RV systolic function on filling. Thus this study was designed first to validate a potentially more accurate method of assessing RV filling and then to apply it to patients with systemic hypertension. Specifically, we sought (1) to determine whether normalization of peak filling rate to tricuspid SV yields a reliable index of analyzing RV filling dynamics; (2) to assess RV and LV filling dynamics in normal adults and in patients with mild-to-moderate systemic hypertension; and (3) to determine the relation between RV and LV filling patterns and t h e clinical, echocardiographic, and scintigraphic characteristics of this group. METHODS Patient selection. Forty-one subjects aged 25 to 67 years
(49 +- 9 years) with mild-to-moderate essential systemic hypertension (diastolic blood pressure off medication ranged from 90 to 119 mm Hg) were enrolled in this study. Blood pressure was determined at the beginning of each study as the mean of three sphygmomanometer readings, taken 5 minutes apart with the patient in the sitting position for at least 30 minutes, according to the recommendations of the American Heart Association. 12 Duration of hypertension ranged from 2 to 22 years, and antihypertensive therapy had been withdrawn for at least 4 weeks. The exclusion criteria were as follows: heart rate <50 or >90 beats/min, atrioventricular block, supraventricular arrhythmias, evidence of secondary hypertension, renal insufficiency, chronic obstructive pulmonary disease, dia-
Volume 130, Number 5 American Heart Journal
betes mellitus, and clinical, electrocardiographic, or Doppler echocardiographic evidence of valvular heart disease, coronary artery disease, or congestive heart failure. Forty age- and sex-matched normal subjects served as the control group. These subjects were normal volunteers with no evidence of underlying cardiovascular disorder by physical examination, exercise electrocardiography, and exercise thallium-201 myocardial scintigraphy. The study was approved by the local ethical committee. Echocardiographic examination. Complete M-mode, two-dimensional, and Doppler echocardiographic studies were performed using an ultrasound mechanical system equipped with 2.5 and 3.5 MHz transducers (Apogee CX, Interspec, Inc., Ambler, Penn.). M-mode and two-dimensional recordings were made with the subjects in a lateral recumbent position according to the standardization of the American Society of Echocardiography. 13 On M-mode tracings, measurements were made from three consecutive cycles at a paper speed of 50 mm/sec. LV dimensions and wall thickness were measured according to the recommendations of the American Society of Echocardiography. Calculation of LV mass was performed using the Penn convention as described by Devereux et al. 14 From a fourchamber view, mitral annulus diameters were measured at end-diastole from the base of the posterior leaflet to the point of the attachment of the anterior leaflet. For measurements of tricuspid annulus diameter, the transducer was placed at an intermediate position between the LV apex and the left lower sternal border to obtain the longaxis plane of the RV inflow tract, and annulus diameter was measured as the distance between leaflet attachments. The cross-sectional areas of the mitral and tricuspid annuli were determined assuming a circular geometry according to the formula ~r × D2/4. LV volume was calculated using the method ofTortoledo et al. 15 and RV volume according to a biplane area and length approach. 16 RV free-wall thickness and dimensions were measured as previously described. 4 Briefly, from an apical fourchamber view, the major end-diastolic (long axis) of the right ventricle, defmed as the distance between the RV apex to the midpoint of the tricuspid valve annulus (mm) and the maximum end-diastolic dimension (short axis) of the body defined as width at the middle third of the right ventricle (mm), were determined. RV free-wall thickness was derived from RV outflow tract views. Interobserver and intraobserver variability of RV free-wall thickness measurements in our laboratory is 10% _+ 3% and 13% -+ 5%, respectively. 4 Dopplerexamination. Two-dimensionally guided pulsed Doppler interrogation of the LV inflow was carried out from the apical two- or four-chamber views. RV inflow velocities were recorded from the short axis, low parasternal, or apical four-chamber view. The sample volume was placed at the mitral and tricuspid annulus level, and its position was optimized to obtain better defined Doppler waveforms with minimal spectral broadening and with no angle correction. To assess the influence of the sample volume position on tricuspid flowmetry, a subgroup of 20
Cittadini et al.
1075
Table I. Clinical characteristics
Age (yr) Gender (male/female) SBP (mm Hg) DBP (ram Hg) BMI (kg/m2) Heart rate (beats/min)
Normals (n = 40)
Hypertensives (n = 41)
48 _+8 30/10 124 _+12 73 _+9 24 _+5 72 _+3
49 _+9 31/10 154 _+20 101 _+12 25 _+4 73 _+2
p value
n.s. n.s. <0.001 <0.001 n.s. n.s.
SBP, Systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index.
subjects was randomly chosen from both the normal and the hypertensive population. The sample volume was first positioned at the annular level and then moved at the leaflets' tips. Tricuspid early to late diastolic flow velocity (E/A) ratio and RV peak filling rate normalized to SV were assessed at both locations. On the Doppler tracings, measurements were made from six consecutive cycles at a paper speed of 50 or 100 mndsec, to minimize errors resulting from beat-to-beat variability during the respiratory cycles. Tricuspid inflow velocities were recorded during end-tidal apnea. The E/A ratio was derived from mitral and tricuspid flow-velocity curves. Doppler echocardiographic absolute peak filling rate was calculated as the product of peak early-diastolic velocity times atrioventricular annulus cross-sectional area. Absolute peak filling rate was then divided by EDV to obtain normalized filling rates (EDV/sec). Doppler echocardiographic peak filling rates normalized to mitral and tricuspid SV (SV/sec) were calculated by dividing the early peak inflow velocity by diastolic time-velocity integral. 6 In our laboratory, we studied 50 normal subjects (not including those in this study) aged 17 to 71 years (mean, 45 --+ 10) and found that the average values for LV and RV peak filling rates normalized to SV were 5.1 _+ 0.8 and 4.6 -+ 0.7 SV/sec, respectively; therefore, we considered a LV and RV peak filling rate <3.5 and <3.2 SV/sec, respectively, to be abnormal (2 SD below mean values) and the basis for the classification of dysfunctional LV and RV filling. Intraobserver reproducibility of echo-derived LV and RV peak filling rates normalized to SV was assessed by comparing the measurements of the same set of six cardiac cycles, obtained by the same investigator (A.C.) on two separate occasions 2 weeks apart. Interobserver reproducibility was determined by comparing paired readings of two observers (A.C., S.F.), who independently measured a set of six cardiac cycles. Pulmonary artery pressures were not determined because few patients had an adequate tricuspid regurgitation jet, and other methods poorly discriminate mild elevation. 9
Gated blood-pool cardiac scintigraphy. W i t h i n 1 h o u r from the completion of the echocardiography, equilibrium radionuclide angiography was performed at rest using red blood cells labeled in vivo with 15 to 20 mCi oftechnetium99m and a small field-of-view gamma camera (Starcam 300 A/M, GE) equipped with a low-energy all-purpose col-
1076
November 1995 American Heart Journal
Cittadini et CI].
Table II. Doppler-echocardiographic findings
Aortic root (cm) Left atrium (cm) Left ventricle-echo LV posterior wall thickness (cm) Septal thickness (cm) LV mass (gm) LV mass index (gm/m2) Relative wall thickness (cm) Right ventricle-echo RV long axis (cm) RV short axis (cm) RV free-wall thickness (cm) Left ventricle-Doppler Mitral E/A ratio LV PFR/EDV (EDV/sec) LV PFR/SV (SV/sec) Right ventricle-Doppler Tricuspid E/A ratio RV PFR/EDV (EDV/sec) RV PFR/SV (SV/sec)
Normals (n = 40)
Hypertensive (n = 41)
p value
3.5 _+0.5 3.6 ± 0.4
3.6 -~ 0.7 3.8 ± 0.5
n.s. <0.05
0.82 ± 0.1 0.86 ± 0.1 134 +_29 74 ± 15 0.34 ± 0.05
1.0 ± 0.1 1.17 +_0.18 193 -+ 41 107 -_ 24 0.40 ± 0.06
<0.001 <0.001 <0.001 <0.001 <0.01
4 +_0.5 t.8 ± 0.3 0.4 _+0.2
4.5 ± 0.5 1.9 ± 0.2 0.7 ± 0.2
n.s. n.s. <0.001
1.8 _+0.5 3.6 ± 0.6 4.9 _+0.8
0.9 +_0.4 2.9 _+1 3.7 ± 0.8
<0.001 <0.001 <0.001
1.7 _+0.3 2.9 ± 0.8 4.4 '± 0.6
1.0 _+0.2 2.3 ± 0.9 3.5 ± 0.7
<0.001 <0.001 <0.001
LV, Lei~ ventricle; RV, right ventricle; PFR/EDV, peak filling rate normalized to end-diastolic volume; PFR/SV, peak filling rate normalized to stroke volume.
limator. LV ejection fraction (%), peak ejection rate (EDV/ sec), time to peak ejection rate (msec), peak filling rate normalized to EDV (EDV/sec) and to SV (SV/sec), and time to peak filling rate were measured as previously described. 17 RV function indices were assessed in a modified left anterior oblique position, which provides the best separation between ventricles and atria, thus avoiding underestimation of RV ejection fraction.iS The average values in 50 normal subjects aged 17 to 71 years for LV and RV . • peak filhng rates corrected by SV were 5.5 + 0.9 and 5.1 -+ 0.9 SV/sec, respectively; thus we considered LV and RV peak filling rate <3.7 and <3,4 SV/sec (2 SD below these mean values), respectively, to be abnormal. Statistics. Data were handled and stored using the DBASE III plus software, and the statistical analysis was performed using the SPSS package. Independent sample t test for unpaired data, linear regression analyses, and Fisher's exact test (two-tailed) were used when appropriate. Multiple regression analysis with the backward elimination procedure was carried out. To evaluate determinants of each index of filling, the following clinical, echocardiographic, and scintigraphic characteristics of the patient groups were the 10 independent variables tested: age, sex, duration of hypertension, heart rate, LV mass index, RV free-wall thickness, systolic and diastolic blood pressures, peak filling rate of t h e opposite ventricle, and ejection fraction. A univariate p value <0.05 was required for each independent variable to enter the equation. Data are reported as mean +- SD.
RESULTS Clinical characteristics. (Table I) Age, weight, and h e i g h t did not significantly differ b e t w e e n the nor-
motensive a n d the h y p e r t e n s i v e groups. As expected, systolic a n d diastolic p r e s s u r e s were significantly h i g h e r in the h y p e r t e n s i v e t h a n in the control group. T w e n t y - o n e p a t i e n t s h a d mild h y p e r t e n s i o n , as defined by a diastolic blood p r e s s u r e v a l u e b e t w e e n 90 a n d 104 m m Hg, a n d 20 h a d m o d e r a t e hypertension, as defined b y a diastolic blood p r e s s u r e value bet w e e n 105 a n d 119 m m Hg. H e a r t r a t e was not significantly different b e t w e e n the two groups. Echocardiographic findings (Table II). Left atrial dimensions a n d i n t e r v e n t r i c u l a r septal a n d posterior wall thickne§ses were h i g h e r in the h y p e r t e n sives t h a n in t h e controls, as were relative wall thickness, LV mass, a n d LV m a s s index. LV hypertrophy, as defined by a LV m a s s >134 grn]m 2 for m e n a n d 110 gm]m 2 for w o m e n , 14 w a s p r e s e n t in 32% (13 of 41 patients). RV c h a m b e r size revealed no differences in long-axis a n d short-axis m e a s u r e m e n t s , w h e r e a s RV free-wall thickness was h i g h e r in the h y p e r t e n s i v e s as c o m p a r e d w i t h n o r m a l subjects. Doppler findings (Table II). Overall, the filling dyn a m i c s of both left a n d r i g h t ventricles in h y p e r t e n sion were c h a r a c t e r i z e d by a decrease in early p e a k filling r a t e s a n d the E/A ratio. Sixteen (39%) of 41 p a t i e n t s h a d a n a b n o r m a l RV p e a k filling r a t e normalized to SV, w h e r e a s 19 (46%) h a d a n a b n o r m a l LV p e a k filling rate. I n the s u b g r o u p of patients exa m i n e d for assessing the influence of s a m p l e v o l u m e positioning, m o v i n g the s a m p l e v o l u m e from the tricuspid a n n u l u s to the tip c h a n g e d RV E/A ratio from 1.35 +_ 0.4 to 1.71 _+ 0.6 (p < 0.001), w h e r e a s RV
Volume 130, Number 5 American Heart Journal
Cittadini et al.
1077
Table III. Scintigraphic findings Normals (n = 40) Left ventricle LV ejection fraction (%) LV p e a k ejection r a t e (EDV/sec) LV t i m e to p e a k filling r a t e (msec) LV p e a k filling r a t e (EDV/sec) LV p e a k filling r a t e (SV/sec) R i g h t ventricle RV ejection fraction (%) RV p e a k ejection r a t e (EDV/sec) RV t i m e to p e a k filling r a t e (msec) RV p e a k filling r a t e (EDV/sec) RV p e a k filling r a t e (SV/sec)
Hypertensives (n = 41)
p value
66 3.8 153 3.5 5.3
_+ 7 _+ 0.6 _+ 55 -+ 0.4 -+ 1.1
67 3.6 183 2.7 4.0
_+ 8 _+ 0.7 _+ 61 _+ 0.7 -+ 1.1
n.s. n.s. <0.02 <0.001 <0.001
58 2.5 133 2.9 5.0
_+ 8 _+ 0.8 _+ 45 +_ 0.4 _+ 0.8
57 2.4 160 2.2 3.8
_+ 7.8 _+ 0.6 _+ 70 _+ 0.7 _+ 1.0
n.s. n.s. <0.04 <0.001 <0.001
LV, Left ventricle; RV, right ventricle; EDV/sec, end-diastolic volumes/second; SV/sec, stroke volumes/second.
peak filling rate normalized to SV did not change (3.95 _+ 0.7 vs 3.91 ± 1, p -- n.s.) Reproducibility of echo-derived peak filling rate corrected by SV was as follows: intraobserver reproducibility 5% _+ 4% and 6% ± 4% for left and right ventricle, respectively; interobserver reproducibility was 7% -+ 5% and 7% _+ 6% for left and right ventricle, respectively. Gated blood-pool cardiac scintigraphy (Table III). Overall, indexes of LV and RV systolic function did not significantly differ between normotensive and hypertensive subgroups. Conversely, indexes of diastolic filling dynamics were impaired in both the left and right sides of the heart in t h e p a t i e n t s with systemic hypertension. In particular, average LV and RV peak filling rates were 2.7 and 2.2 EDV/sec as compared with 3.5 and 2.9 EDV/sec, respectively, in control subjects. LV and RV peak filling rates normalized to SVwere 4.0 - 1.1 and 3.8 ± 1.0 SV/sec in hypertension as compared with 5.3 _+ 1.1 and 5.0 ± 0.8 SV/sec in the controls, respectively (p < 0.01 for both). Fourteen (34%) of the 41 hypertensive subjects had an abnormal RV peak filling rate corrected by SV, whereas 17 (43%) had an abnormal LV peak filling rate. Univariate correlates of nuclear and Doppler-derived peak filling rates. The correlation between Doppler
and nuclear LV peak filling rates normalized to EDV was poor in the overall population (r = 0.42) (Fig. 1, A). Even worse was the correlation between echo and nuclear RV peak filling rate normalized to EDV (r = 0.34) (Fig. 1, B). On the contrary, there were excellent correlations between Doppler- and nuclearderived peak filling rate normalized to SV in the overall population (both hypertensives and normals) with r values of 0.92 (Fig. 1, C) and 0.87 (Fig. 1, D) for left and right ventricle, respectively. There were also excellent correlations in the hypertensive sub-
Table IV. Univariate correlations between RV and LV in-
dexes of filling Normals (n = 40) Echo p e a k filling r a t e (SV/sec) Echo p e a k filling r a t e (EDV/sec) N u c l e a r p e a k filling r a t e (SV/sec) N u c l e a r p e a k filling r a t e (EDV/sec)
r r r r
= = = =
0.68 0.51 0.65 0.61
Hypertensives (n = 41) r r r r
= = = =
0.66 0.49 0.63 0.57
RV, Right ventrieular; LV, left ventrieular; SV, stroke volume; EDV, enddiastolic volume.
jects alone, with r values of 0.89 for left ventricle and of 0.86 for right ventricle (p < 0.0001). Correlations between a variety of RV and LV indexes of filling. RV and LV peak filling rate indexes were
related (r = 0.49 to 0.68) (Table IV), suggesting that RV filling was often abnormal in patients with LV filling impairment (Fig. 2). These relations were found regardless of the technique or specific index used to measure filling or of the population in which they were examined. However, it should be noted that echo and peak filling rates corrected by SV produced closer relations between RV and LV filling in both normals and hypertensives than did peak filling rate corrected by EDV (r = 0.68 vs 0.51 in normals and r = 0.66 vs 0.49 in hypertensives) (Figure 3, A-D). Multivariate analysis of LV and RV diastolic filling.
Multivariate analyses of the determinants of peak filling rate by both echo and nuclear methods showed that, in both normals and hypertensives, normalization of peak filling rate to SVinstead of EDV removed the influence of ejection fraction on systolic function (Table V). In normals, assessment of both RV and LV filling dynamics by E/A ratio was affected by heart rate; this variable did not influence early peak filling rates normalized to SV. Filling dynamics in both
1078
November 1 9 9 5 American Heart Journal
Cittadini et al.
A
]B
I~. . . . . . . . . 9-°.'21
6y=O. 17x+2.44
+
4-
i 2
i 4
Echo-Doppler
C
~
4-
Z
2-
Echo-Doppler
LV-PFR (EDV/e)
D 1.18x-0.43
r=0.92
i
i
Echo-Doppler
•
i 6
8y=
]
r=0.34
LV-PFR (SV/s)
8-
~y
=1.22-0.41
RV-PFR (EDV/$)
r=0.87
i
I
I
I
I
2
4
6
8
Echo-Doppler
RV-PFR (SV s)
Fig, 1. Scatterplot of regression analysis of left ventricular estimated by echo-Doppler and nuclear techniques in both hypertensives (crosses, n = 41) and normal (dots, n = 40) populations. A and B, Left ventricular and right ventricular peak filling rates normalized to end-diastolic volume, respectively, are shown. C and D, Left ventricular (LV) and right ventricular (RV) peak filling rate (PFR) normalized to stroke volume (SV), respectively, are shown.
<3.5
ventricles in normals were affected by age (Fig. 4), regardless of method. In contrast, in hypertensive subjects, age and heart rate were not independently related to any filling parameter; their influence was superseded by t h a t of systolic blood pressure. Thus in the hypertensive population, the only independent determinant of the Doppler-derived RV peak filling rate normalized to SV was the LV filling rate.
8-
U) nu. D.
6-
•
+,o ¢,j
°
DISCUSSION
4-
o 0
.
.
.
.
.
.
.
.
.
.
.
6t o ul
2-
I
2
I I I I
I
I
I
4
6
8
Echo-Doppler
LV-PFR (SV/s)
Fig. 2. Scatterplot of left ventricular and right ventricular echo-detected peak filling rates values in hypertensives (crosses) and normal subjects (dots). D a s h e d lines indicate lower limits of left and right echo-detected peak filling rate normalized to stroke volume. Therefore, 19 and 16 hypertensives fell in abnormal range for left ventricular and right ventricular filling, respectively.
Although determination of the relative utility of any index of diastolic function is problematic because of the complexity of diastole and because of the lack of a gold standard, the results of this study suggest t h a t normalization of early peak filling rate to SV may be preferable to other methods in both right and left ventricles. In our study, this method eliminated the confounding influences on filling of systolic function, which confounds use of peak filling rate normalized to EDV, and of heart rate, which confounds use of the E/A ratio. Further, comparison between RV and LV filling rates was improved by normalization of Doppler-derived values to SV. The utility of this method in revealing pathophysiology was then demonstrated by application to a hypertensive pop-
Volume 130, Number 5 American
Cittadini et al.
HeartJournal
t
6y=O,65x+057 r=0.51
]
B
"6 g
•
~
O•
i
8-
1079
J~=o.,gx+2.o, r=9.~ i
64-
~ -
oo
~o u
Echo-Doppler nV-PFR (EDV/s)
C
6-
,,=
'=-
o.
1.0,~,07
.... 9 I
-t-
Echo-Doppler LV-PFR (SV/s)
D
6-
=.
+ 4-
+ +
.
,o
+ 2
Echo-Doppler LV-PFR (EDV/sv)
u 4
Echo-Doppler LV-PFR (SV~s)
Fig. 3. Scatterplot of regression analysis between right ventricular (RV) and left ventricular (LV) indexes of filling. A and B, Peak filling rate (PFR) normalized to end-diastolic volume (EDV) and stroke volume (SV), respectively, are considered in the normal population (dots). C and D, same indexes in hypertensive population (crosses) are shown.
ulation. Related abnormalities in both RV and LV filling were found, suggesting ventricular diastolic interdependence, and the well-described influence of age on filling was displaced by that of systolic blood pressure. Selection of the best index of filling. Doppler echocardiography and radionuclide ventriculography are the two most common noninvasive techniques that allow assessment of LV and RV diastolic filling. The correlation between these methods is rather poor when absolute or EDV-normalized peak filling rates are considered, 5, 6 perhaps because of the geometric assumptions necessary to estimate peak filling rates and their dependence on cavity size. When this technique is applied to the right ventricle, all these assumptions become even less reliable, as it is a structurally complex chamber, and poorly imaged. Thus it is not surprising that we found poor correlations between n u c l e a r - a n d echo-detected LV and especially RV peak filling rates normalized to EDV (r = 0.42 and r = 0.34, respectively). Therefore, a method of normalizing peak filling rate that does not depend on volume estimates would be highly desireable for both ventricles and would be particularly suited to RV peak filling rate assessment. Normalization to SV eliminates many of the sources of error of normalization to EDV. It is easily
calculated using both Doppler (E/total diastolic volume thickness index [VTI]) and nuclear (peak filling rate [PFR]/EDV + ejection fraction [EF])techniques, does not require visualization and measurement of annulus or chamber sizes, and does not depend on ejection fraction. 19, 20 The close correlations we found between Doppler and nuclear results in both left and right ventricles also help confirm the validity of our approach. Another index of filling, the E/A ratio, is also easy to obtain and less dependent on chamber size and systolic function. Nevertheless, this study suggests that there are several advantages of normalized peak filling rate relative to E/A ratio. These include independence of Doppler sample volume location, independence of H R influence in normal subjects, and values more comparable to those obtained by nuclear techniques. Although it is impossible definitively to determine the superiority of one technique over the other, our findings suggest that Doppler-derived peak early filling rates, when normalized to SV, are easy to obtain, accurate, and relatively independent of other physiologic influences, suggesting that this index is one of the best available. RV function in hypertension. RV filling dynamics have only recently been evaluated in normal subjects and in selected clinical settings, such as cardiac amyloidosis, acromegaly, and myocardial infarc-
November 1995 American Heart Journal
1080 Cittadini et al. Table V. Multivariate predictors of peak filling rates Dependent variable Normal subjects nLV-PFR (EDV/sec) nRV-PFR (EDV/sec) nLV-PFR (SV/sec) nRV-PFR (SV/sec) eLV-PFR (SV/sec) eRV-PFR (SV/sec) eLV-E/A eRV-E/A ' Hypertensive patients nLV-PFR (EDV/sec) nRV-PFR (EDV/sec) nLV-PFR (SV/sec) nRV-PFR (SV/sec) eLV-PFR (SV/sec) eRV-PFR (SV/sec) eLV-E/A eRV-E/A
Independent predictor(s)
Multiple r value
nLV-EF nRV-EF Age Age Age Age Age Age
nRV-PFR nLV-PFR nRV-PFR nLV-PFR eRV-PFR eLV-PFR H e a r t rate H e a r t rate
Age Age
nLV-EF nRV-EF SBP nLV-PFR SBP eLV-PFR SBP eLV-E/A
nRV-PFR nLV-PFR nRV-PFR
SBP
eRV-PFR eRV-E/A
eRV-E/A eLV-E/A
0.75 0.77 0.67 0.71 0.74 0.72 0.79 0.77 0.77 0.79 0.66 0.65 0.65 0.62 0.67 0.65
n, Nuclear; e, echocardiographic; LV, left ventricular; RV, right ventricular; PFR, peak filling rate; SV, stroke volume; EDV, end-diastolic volume; SBP, systolic blood pressure; EF, ejection fraction.
tion. 1-4 The complex anatomy of right ventricle, as well as the assumption that this chamber plays a minor role in overall cardiovascular function, have limited investigations about t h e effects of systemic pressure overload on RV function. Nevertheless, two recent studies have reported abnormal Doppler echocardiographic RV variables in hypertensive subjects as compared with age- and sex-matched healthy normals. 7,s The difficulty of obtaining reliable echocardiographic estimates of RV volumes led these investigators to exclude analysis of RV systolic function, This omission is significant, as it is reasonable to assume that RV systolic and diastolic function are related, as is the case with the LV. Moreover, no study thus far has constructed a multivariate analysis to find out which variables are truly independent in predicting RV peak filling rates. RV ejection fraction has been reported to be either normal or decreased in systemic hypertension. 9, 10 Our findings of normal resting values of RV ejection fraction in our patients are consistent with previous observations that, in the absence of LV dysfunction, RV ejection fraction is usually normal. 1° In spite of a normal ejection fraction, this study shows that impairment of RV diastolic filling is a common finding in systemic hypertension. In particular, as far as peak filling rate corrected by SV is concerned, 34% and 39% of the hypertensive patients had RV filling abnormalities by Doppler and nuclear techniques, respectively, as compared with an only slightly higher prevalence of LV filling impairment (46% and 43%, respectively). In spite of estimates of similar RV cavity size in
normals and hypertensives, RV hypertrophy was consistently present in our patients. This finding has been previously reported, 7 although the pathogenesis is still an open issue. It has been proposed that hemodynamic overload of the right ventricle might result from elevated LV diastolic pressure, or that unidentified local or circulating growth factors such as catecholamines, angiotensin, insulin-like growthfactor-l, and so on, may act on both ventricles. 21 Because we were unable to make an accurate noninvasive measurement of RV or LV pressures in our ambulatory patients (see Methods), we could provide no data correlating either RV hypertrophy or RV filling indexes to pulmonary or left heart pressures. Relations between RV and LV filling. An important finding of this study is that RV and LV filling indexes are closely interrelated in both the normal and hypertensive populations. Those patients with the most markedly impaired LV diastolic function also had the most striking RV abnormalities. Several lines of evidence support the concept of a mechanical and hemodynamic interaction between right and left ventricles, both in diastole and in systole. 22 From an anatomic standpoint, the two ventricles share common circular and spiral muscle fiber bundles, interventricular septum, and a common pericardium. Furthermore, the pulmonary circulation indirectly connects the two ventricles. Although our finding of close correlations between LV and RV indexes of filling by both indexes provides further in vivo evidence of this diastolic interplay, the results of the multivariate analysis are even more striking. For both LV and RV, and in both hypertensives and normals, ev-
Volume 130, Number 5 American Heart Journal
Cittc~ini et al.
n
U
7-
1081
6-
,,= a.
5r=0.70
r=0.74
o 6
~-+
:,
~ _ + j~H ~ r=0.40
, o
r=0.31
3-
4++ 2
2'0
20
go
8'0
Age
2
2'0
,'o
+ 8'0
Age
Fig. 4. Univariate correlation between age and left ventricular peak filling rate normalized to stroke volume (LV-PFR) in A; in B correlation between age and right ventricular peak filling rate normalized to stroke volume (RV-PFR) is shown. Crosses, Hypertensives; dots, normals.
ery index of filling had, as an independent determinant, the corresponding filling index of the opposite chamber. Thus RV filling was dependent on LV filling and vice versa. Determinants of LV and RV filling in hypertension. Although age is well described to influence LV filling in
normals, 23 its importance in disease states is unclear. Our findings are consistent with other studies that found that, in the presence of cardiomyopathy and hypertension, the effects of age on LV filling are no longer detected. 24 To our knowledge, no previous studies have examined the multiple influences of age and disease state on RV function. Our results demonstrate that, in the normal right ventricle, similar to the normal left ventricle, age produces intrinsic changes in myocardial stiffness or relaxation or both so as to impair RV filling 25, 26 However, we have also shown t h a t the effects of hypertension include abnormal RV as well as LV filling. This is of such a magnitude that the effects of hypertension on the RV can supersede those of normal aging. This is particularly remarkable because the primary hemodynamic abnormality is one of systemic pressure overload, whose effects should be more marked on the left ventricle. Study limitations. Because the validation of our method required comparison of several techniques in the same individual, m a n y of the patient-related factors t h a t affect diastole should have been constant (including HR, left atrial and LV pressures, PR interval, etc.) and not influenced our findings. In this study, we chose an ellipsoidal model for estimation of RV volumes; although the RV is clearly not a prolate ellipse, this formula does include the infundibular contribution to volume, is relatively simple, and has been demonstrated to yield an excellent correlation (r = 0.95) with h u m a n RV cast measurements. 16 Several different approaches have been proposed for calculation of RV volume, but none of them appears
to offer a superior estimate, and they usually require additional calculations. Further, most of our conclusions are based on the use of peak filling rate normalized to SV, a method that eliminates dependence on RV volume estimation. Although equilibrium radionuclide angiography does not suffer from the same geometric assumptions as does 2D echocardiography, a source of error has been described in difficulties in separating the right atrium from the right ventricle. We used a modified left anterior oblique position to avoid underestimation of RV ejection fraction; in addition, any underestimation should have been similarly present in both the study groups. Measurements of RV freewall thickness by echocardiography have been reported to have a high variability; however, previous data from our laboratory showed an acceptable reproducibility.4 Clinical implications. This study demonstrates that of the several indexes of RV filling that can be easily measured by clinically available techniques, RV peak filling rate normalized to SV might be preferred because of ease of calculation, repeatability, freedom from physiologic influences, and an excellent correlation between echo and radionuclide angiographic assessments. In applying this method, we showed that more t h a n one third of hypertensive subjects had RV filling abnormalities. Although altered RV filling has been described in several disease states, 1-4 the clinical relevance and the prognostic significance of RV diastolic impairment is unclear. In this respect, the availability of better indexes of RV filling, such as peak filling rate normalized to SV, should facilitate a detailed noninvasive assessment of RV diastolic function in various clinical settings and a better understanding of its importance. The fact that LV and RV filling are both dependent on diastolic performance in the other chamber strongly supports the concept ofventricular diastolic
1082
November 1995 American Heart Journal
Cittadini et al.
interdependence. Whether this interaction is the result of anatomic factors or extrinsic factors (i.e., hormonal influences acting on both ventricles) or both is unclear. However, it is readily apparent that RV abnormalities can be just as much a part of the complex of cardiac effects of systemic hypertension as are LV filling abnormalities, and that the evaluation of RV filling may provide additional insight into LV diastolic function. Also clinically relevant is the effect of age on RV filling in normal subjects; this finding suggests that determinations of right ventricular filling should include reference to age-related standards. However, the removal of such effects in the hypertensive population must also be taken into account when assessing RV filling in disease states; further research is necessary to determine the relative importance of age and disease in determining RV filling. Conclusions. In this study we examined the utility of echo-Doppler-derived peak filling rate normalized to tricuspid SV as an index of RV diastolic filling in normal and hypertensive populations. This index appears to be particularly suitable for the investigation of RV filling dynamics because of the excellent correlation with radionuclide angiographic techniques, the ease of calculation, and relative independence of geometric assumption, ventricular size, sample volume location, heart rate, and systolic function. Because RV filling abnormalities are common in hypertension and appear related to both LV abnormalities and the severity of systemic blood pressure elevation, RV diastolic abnormalities should be further investigated as an important component of hypertensive heart disease and a manifestation of ventricular interaction. REFERENCES
1. Nishimura RA, Abel MD, Hatle LK, Tajik AJ. Assessment of diastolic function of the heart: background and current application of Doppler echocardiography. Part II. Clinical studies. Mayo Clin Proc 1989;64:181204. 2. Klein AL, Hatle LK, Burstew DJ, Taliercio CP, Seward JB, Kyle RA, Bailey KR, Gertz MA, Tajik AJ. Comprehensive Doppler assessment of right ventricular diastolic function in cardiac amyloidosis. J Am Cell Cardiol 1990;15:99-108. 3. Fujii J, Yazaki Y, Sawada H, Aizawa T, Watanabe H, Kate K. Noninvasive assessment of left and right ventricular filling in myocardial infarction with a two-dimensional Doppler echocardiographic method. J Am Cell Cardiol 1985;5:1155-60. 4. Fazio S, Cittadini A, Sabatini D, Merola B, Colao AM, Biondi B, Lombardi G, Sacch L. Evidence for biventricular involvement in acromegaly: a Doppler echocardiographic study. Eur Heart J 1993;14:26-33. 5. Friedman BJ, Drinkovic N, Miles H, Shih WJ, Mazzoleni A, DeMaria
6.
7.
8.
9. 10.
11. 12.
13.
14.
15.
16.
17.
18.
19. 20.
21. 22. 23.
24.
25.
26.
AN. Assessment of left ventricular diastolic function: comparison of Doppler echocardiography and gated blood pool scintigraphy. J Am Cell Cardiol 1986;8:1348-54. Bowman LK, Lee FA, Jaffe CC, Mattera J, Wackers FJ, Zaret BL. Peak filling rate normalized to mitral stroke volume: a new Doppler echocardiographic filling index validated by radionuclide angiographic technique. J Am Cell Cardiol 1988;12:937-43. Chakko S, de Marchena E, Kessler KM, Materson BJ, Myerburg RJ. Right ventricular diastolic function in systemic hypertension. Am J Cardiol 1990;65:1117-20. Habib GB, Zoghbi WA. Doppler assessment of right ventricular filling dynamics in systemic hypertension: comparison with left ventricular filling. A~ HEARTJ 1992;124:1313:20. Ferlinz J. Right ventricular performance in essential hypertension. Circulation 1980;61:156-62. Nunez BDI Messerli FH, Amodeo C, Garavaglia GE, Frohlic ED. Biventricular cardiac hypertrophy in essential hypertension. AMHEART J 1987;111:813-18. Neubar T, ScwartzkopffB, Strauer BE. Right ventricular performance in arterial hypertension. Eur Heart J 1992;13(suppl D):33-8. Kirkendall WM, Burton HC, Epstein FH, Freis ED. Recommendation for human blood pressure determination by sphygmomanometer. Circulation 1967;36:980-8. Sahn DJ, DeMaria A, Kisslo J, Weyman A. The committee on M-mode standardization of the American Society of Echocardiography. Recommendations regarding quantification in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation 1978; 58:1072-83. Devereux RB. Detection of left ventricular hypertrophy by M-mode echocardiography. Anatomic validation, standardization and comparison to other methods. Hypertension 1987;9(suppl II):19-26. Tortoledo FA, Quinones MA, Fernandez GC, Waggoner AD, Winters WL Jr. Quantification of left ventricular volume by two-dimensional echocardiography: a simplified and accurate approach. Circulation 1983;67:579-84. Levine RA, Gibson TC, Aretz T, Gillam LD, Guyer DE, King ME, Weym a n AE. Echocardiographic measurement of right ventricular volume. Circulation 1984;69:497-505. Cuocol0 A, Frederick LS, Brush JE, Maron BJ, Bacharach SL, Bonow RO. Left ventricular hypertrophy and impaired diastolic filling in essential hypertension. Circulation 1990;81:978-86. Legrand V, Checigne M, Foulon J, Rigo P. Evaluation of right-ventricular function by gated blood-pool scintigraphy. J Nucl Med 1983;24:88693. Bonow RO. Radionuclide angiographic evaluation of left ventricular diastolic function. Circulation 1991;84(suppl I):I-208-I15. Smith VE, Schulman P, Karimeddini MK, White WB, Meeran MK, Katz AM. Rapid ventricular filling in left ventricular hypertrophy: II. Pathologic hypertrophy. J Am Cell Cardiol 1985;5:869-74. Morgan HE, Baker KM. Cardiac hypertrophy. Mechanical, neural, and endocrine dependence. Circulation 1991;83:13-25. Bove A, Santamore W. Ventricular interdependence. Prog Cardiovasc Dis 1981;23:365-87. Benjamin EJ, Levy D, Anderson KM, Wolf PA, Plehn JF, Evans JC, Comai K, Fuller DL, St. John Stutten M. Determinants of Doppler indexes of left ventricular diastolic function in normal subjects (the Framingham Heart Study). Am J Cardio] 1992;70:508-15. Sartory MP, Quinones MA, Kuo LC. Relation of Doppler-derived left ventricular filling parameters to age and radius/thickness ratio in normal and pathologic states. Am J Cardiol 1987;59:1179-82. Berman GO, Reichek N, Br°ws°n D, D°uglas PS" Effects °f sample v°lume location, imaging view, heart rate and age on tricuspid velocimetry in normal subjects. Am J Cardiol 1990;65:1026-30. Zoghbi WA, Habib GB, Qninones MA. Doppler assessment of right ventricular filling in a normal population: comparison with left ventricular filling dynamics. Circulation 1990;82:1316-24.
Boston, Mass., and Naples, Italy To determine the optimal method of normalizing peak filling rate (PFR) determinations and apply it to the assessment of right ventricular (RV) and left ventricular (LV) filling characteristics and their interactions, 41 subjects with hypertension and 40 matched normals underwent echo-Doppler and nuclear study. Conventional normalization of PFR to end-diastolic volume (EDV) yielded poor correlations between nuclear- and echo-derived PFR (RV, r= 0.34; LV, r= 0.42), whereas nuclear and echo PFR normalized to stroke volume (SV) were closely correlated (RV, r = 0.87; LV, r = 0.92). Further, use of PFR normalized to SV revealed a close relation between RV and LV filling characteristics. Multivariate analysis confirmed that, in contrast to normalization to EDV or early to late filling-velocity ratios (E/A), peak filling rate normalized to SV was independent of ejection fraction and heart rate. In addition, RV filling impairment was related to LV filling impairment, and the effects of hypertension eliminated the independent influence of age on both LV and RV filling. In conclusion, normalization of PFR to SV may be preferable to use of EDV or E/A in evaluating RV and LV filling dynamics. (AM HEART J 1995;130:1074-82.)
Abnormal right ventricular (RV) filling dynamics have been described in a variety of cardiovascular diseases, often associated with significant clinical consequences. TM However, the best method for assessing RV filling is unknown. Conventional assessment ofventricular filling rates using normalization to end-diastolic volume (EDV) is difficult to determine for the irregularly shaped and poorly imaged right ventricle, and even in the left ventricle yields From the Charles A. Dana Research Institute and the Harvard-Thorndike Laboratory, Department of Medicine (Cardiovascular Division) of Beth Israel Hospital and Harvard Medical School, and the aDepartment of Internal Medicine of Federico II University Medical School Supported in part by a grant from ConsigHo Nazionale delle Ricerche, Italy (A.C.) Received for publication April 6, 1995; accepted May 24, 1995. Reprint requests: Pamela S. Douglas, MD, Cardiovascular Division, Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215. Copyright © 1995 by Mosby-Year Book, Inc. 0002-8703/95/$5.00 + 0 4/1/66515
1074
only modest correlations between nuclear and echocardiographic methods. 5, 6 To eliminate the influence of chamber volume, two studies have suggested normalization of Doppler flow velocities to tricuspid stroke volume (SV), 7, s but no scintigraphic validation of this index has been performed. In part because of these methodologic difficulties, little information is available concerning RV filling dynamics in systemic hypertension, in spite of evidence t h a t the right ventricle may be affected in left ventricular (LV) pressure overload states. 9-11 Specifically, no study to date has examined the determinants of filling or has taken into account the impact of RV systolic function on filling. Thus this study was designed first to validate a potentially more accurate method of assessing RV filling and then to apply it to patients with systemic hypertension. Specifically, we sought (1) to determine whether normalization of peak filling rate to tricuspid SV yields a reliable index of analyzing RV filling dynamics; (2) to assess RV and LV filling dynamics in normal adults and in patients with mild-to-moderate systemic hypertension; and (3) to determine the relation between RV and LV filling patterns and t h e clinical, echocardiographic, and scintigraphic characteristics of this group. METHODS Patient selection. Forty-one subjects aged 25 to 67 years
(49 +- 9 years) with mild-to-moderate essential systemic hypertension (diastolic blood pressure off medication ranged from 90 to 119 mm Hg) were enrolled in this study. Blood pressure was determined at the beginning of each study as the mean of three sphygmomanometer readings, taken 5 minutes apart with the patient in the sitting position for at least 30 minutes, according to the recommendations of the American Heart Association. 12 Duration of hypertension ranged from 2 to 22 years, and antihypertensive therapy had been withdrawn for at least 4 weeks. The exclusion criteria were as follows: heart rate <50 or >90 beats/min, atrioventricular block, supraventricular arrhythmias, evidence of secondary hypertension, renal insufficiency, chronic obstructive pulmonary disease, dia-
Volume 130, Number 5 American Heart Journal
betes mellitus, and clinical, electrocardiographic, or Doppler echocardiographic evidence of valvular heart disease, coronary artery disease, or congestive heart failure. Forty age- and sex-matched normal subjects served as the control group. These subjects were normal volunteers with no evidence of underlying cardiovascular disorder by physical examination, exercise electrocardiography, and exercise thallium-201 myocardial scintigraphy. The study was approved by the local ethical committee. Echocardiographic examination. Complete M-mode, two-dimensional, and Doppler echocardiographic studies were performed using an ultrasound mechanical system equipped with 2.5 and 3.5 MHz transducers (Apogee CX, Interspec, Inc., Ambler, Penn.). M-mode and two-dimensional recordings were made with the subjects in a lateral recumbent position according to the standardization of the American Society of Echocardiography. 13 On M-mode tracings, measurements were made from three consecutive cycles at a paper speed of 50 mm/sec. LV dimensions and wall thickness were measured according to the recommendations of the American Society of Echocardiography. Calculation of LV mass was performed using the Penn convention as described by Devereux et al. 14 From a fourchamber view, mitral annulus diameters were measured at end-diastole from the base of the posterior leaflet to the point of the attachment of the anterior leaflet. For measurements of tricuspid annulus diameter, the transducer was placed at an intermediate position between the LV apex and the left lower sternal border to obtain the longaxis plane of the RV inflow tract, and annulus diameter was measured as the distance between leaflet attachments. The cross-sectional areas of the mitral and tricuspid annuli were determined assuming a circular geometry according to the formula ~r × D2/4. LV volume was calculated using the method ofTortoledo et al. 15 and RV volume according to a biplane area and length approach. 16 RV free-wall thickness and dimensions were measured as previously described. 4 Briefly, from an apical fourchamber view, the major end-diastolic (long axis) of the right ventricle, defmed as the distance between the RV apex to the midpoint of the tricuspid valve annulus (mm) and the maximum end-diastolic dimension (short axis) of the body defined as width at the middle third of the right ventricle (mm), were determined. RV free-wall thickness was derived from RV outflow tract views. Interobserver and intraobserver variability of RV free-wall thickness measurements in our laboratory is 10% _+ 3% and 13% -+ 5%, respectively. 4 Dopplerexamination. Two-dimensionally guided pulsed Doppler interrogation of the LV inflow was carried out from the apical two- or four-chamber views. RV inflow velocities were recorded from the short axis, low parasternal, or apical four-chamber view. The sample volume was placed at the mitral and tricuspid annulus level, and its position was optimized to obtain better defined Doppler waveforms with minimal spectral broadening and with no angle correction. To assess the influence of the sample volume position on tricuspid flowmetry, a subgroup of 20
Cittadini et al.
1075
Table I. Clinical characteristics
Age (yr) Gender (male/female) SBP (mm Hg) DBP (ram Hg) BMI (kg/m2) Heart rate (beats/min)
Normals (n = 40)
Hypertensives (n = 41)
48 _+8 30/10 124 _+12 73 _+9 24 _+5 72 _+3
49 _+9 31/10 154 _+20 101 _+12 25 _+4 73 _+2
p value
n.s. n.s. <0.001 <0.001 n.s. n.s.
SBP, Systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index.
subjects was randomly chosen from both the normal and the hypertensive population. The sample volume was first positioned at the annular level and then moved at the leaflets' tips. Tricuspid early to late diastolic flow velocity (E/A) ratio and RV peak filling rate normalized to SV were assessed at both locations. On the Doppler tracings, measurements were made from six consecutive cycles at a paper speed of 50 or 100 mndsec, to minimize errors resulting from beat-to-beat variability during the respiratory cycles. Tricuspid inflow velocities were recorded during end-tidal apnea. The E/A ratio was derived from mitral and tricuspid flow-velocity curves. Doppler echocardiographic absolute peak filling rate was calculated as the product of peak early-diastolic velocity times atrioventricular annulus cross-sectional area. Absolute peak filling rate was then divided by EDV to obtain normalized filling rates (EDV/sec). Doppler echocardiographic peak filling rates normalized to mitral and tricuspid SV (SV/sec) were calculated by dividing the early peak inflow velocity by diastolic time-velocity integral. 6 In our laboratory, we studied 50 normal subjects (not including those in this study) aged 17 to 71 years (mean, 45 --+ 10) and found that the average values for LV and RV peak filling rates normalized to SV were 5.1 _+ 0.8 and 4.6 -+ 0.7 SV/sec, respectively; therefore, we considered a LV and RV peak filling rate <3.5 and <3.2 SV/sec, respectively, to be abnormal (2 SD below mean values) and the basis for the classification of dysfunctional LV and RV filling. Intraobserver reproducibility of echo-derived LV and RV peak filling rates normalized to SV was assessed by comparing the measurements of the same set of six cardiac cycles, obtained by the same investigator (A.C.) on two separate occasions 2 weeks apart. Interobserver reproducibility was determined by comparing paired readings of two observers (A.C., S.F.), who independently measured a set of six cardiac cycles. Pulmonary artery pressures were not determined because few patients had an adequate tricuspid regurgitation jet, and other methods poorly discriminate mild elevation. 9
Gated blood-pool cardiac scintigraphy. W i t h i n 1 h o u r from the completion of the echocardiography, equilibrium radionuclide angiography was performed at rest using red blood cells labeled in vivo with 15 to 20 mCi oftechnetium99m and a small field-of-view gamma camera (Starcam 300 A/M, GE) equipped with a low-energy all-purpose col-
1076
November 1995 American Heart Journal
Cittadini et CI].
Table II. Doppler-echocardiographic findings
Aortic root (cm) Left atrium (cm) Left ventricle-echo LV posterior wall thickness (cm) Septal thickness (cm) LV mass (gm) LV mass index (gm/m2) Relative wall thickness (cm) Right ventricle-echo RV long axis (cm) RV short axis (cm) RV free-wall thickness (cm) Left ventricle-Doppler Mitral E/A ratio LV PFR/EDV (EDV/sec) LV PFR/SV (SV/sec) Right ventricle-Doppler Tricuspid E/A ratio RV PFR/EDV (EDV/sec) RV PFR/SV (SV/sec)
Normals (n = 40)
Hypertensive (n = 41)
p value
3.5 _+0.5 3.6 ± 0.4
3.6 -~ 0.7 3.8 ± 0.5
n.s. <0.05
0.82 ± 0.1 0.86 ± 0.1 134 +_29 74 ± 15 0.34 ± 0.05
1.0 ± 0.1 1.17 +_0.18 193 -+ 41 107 -_ 24 0.40 ± 0.06
<0.001 <0.001 <0.001 <0.001 <0.01
4 +_0.5 t.8 ± 0.3 0.4 _+0.2
4.5 ± 0.5 1.9 ± 0.2 0.7 ± 0.2
n.s. n.s. <0.001
1.8 _+0.5 3.6 ± 0.6 4.9 _+0.8
0.9 +_0.4 2.9 _+1 3.7 ± 0.8
<0.001 <0.001 <0.001
1.7 _+0.3 2.9 ± 0.8 4.4 '± 0.6
1.0 _+0.2 2.3 ± 0.9 3.5 ± 0.7
<0.001 <0.001 <0.001
LV, Lei~ ventricle; RV, right ventricle; PFR/EDV, peak filling rate normalized to end-diastolic volume; PFR/SV, peak filling rate normalized to stroke volume.
limator. LV ejection fraction (%), peak ejection rate (EDV/ sec), time to peak ejection rate (msec), peak filling rate normalized to EDV (EDV/sec) and to SV (SV/sec), and time to peak filling rate were measured as previously described. 17 RV function indices were assessed in a modified left anterior oblique position, which provides the best separation between ventricles and atria, thus avoiding underestimation of RV ejection fraction.iS The average values in 50 normal subjects aged 17 to 71 years for LV and RV . • peak filhng rates corrected by SV were 5.5 + 0.9 and 5.1 -+ 0.9 SV/sec, respectively; thus we considered LV and RV peak filling rate <3.7 and <3,4 SV/sec (2 SD below these mean values), respectively, to be abnormal. Statistics. Data were handled and stored using the DBASE III plus software, and the statistical analysis was performed using the SPSS package. Independent sample t test for unpaired data, linear regression analyses, and Fisher's exact test (two-tailed) were used when appropriate. Multiple regression analysis with the backward elimination procedure was carried out. To evaluate determinants of each index of filling, the following clinical, echocardiographic, and scintigraphic characteristics of the patient groups were the 10 independent variables tested: age, sex, duration of hypertension, heart rate, LV mass index, RV free-wall thickness, systolic and diastolic blood pressures, peak filling rate of t h e opposite ventricle, and ejection fraction. A univariate p value <0.05 was required for each independent variable to enter the equation. Data are reported as mean +- SD.
RESULTS Clinical characteristics. (Table I) Age, weight, and h e i g h t did not significantly differ b e t w e e n the nor-
motensive a n d the h y p e r t e n s i v e groups. As expected, systolic a n d diastolic p r e s s u r e s were significantly h i g h e r in the h y p e r t e n s i v e t h a n in the control group. T w e n t y - o n e p a t i e n t s h a d mild h y p e r t e n s i o n , as defined by a diastolic blood p r e s s u r e v a l u e b e t w e e n 90 a n d 104 m m Hg, a n d 20 h a d m o d e r a t e hypertension, as defined b y a diastolic blood p r e s s u r e value bet w e e n 105 a n d 119 m m Hg. H e a r t r a t e was not significantly different b e t w e e n the two groups. Echocardiographic findings (Table II). Left atrial dimensions a n d i n t e r v e n t r i c u l a r septal a n d posterior wall thickne§ses were h i g h e r in the h y p e r t e n sives t h a n in t h e controls, as were relative wall thickness, LV mass, a n d LV m a s s index. LV hypertrophy, as defined by a LV m a s s >134 grn]m 2 for m e n a n d 110 gm]m 2 for w o m e n , 14 w a s p r e s e n t in 32% (13 of 41 patients). RV c h a m b e r size revealed no differences in long-axis a n d short-axis m e a s u r e m e n t s , w h e r e a s RV free-wall thickness was h i g h e r in the h y p e r t e n s i v e s as c o m p a r e d w i t h n o r m a l subjects. Doppler findings (Table II). Overall, the filling dyn a m i c s of both left a n d r i g h t ventricles in h y p e r t e n sion were c h a r a c t e r i z e d by a decrease in early p e a k filling r a t e s a n d the E/A ratio. Sixteen (39%) of 41 p a t i e n t s h a d a n a b n o r m a l RV p e a k filling r a t e normalized to SV, w h e r e a s 19 (46%) h a d a n a b n o r m a l LV p e a k filling rate. I n the s u b g r o u p of patients exa m i n e d for assessing the influence of s a m p l e v o l u m e positioning, m o v i n g the s a m p l e v o l u m e from the tricuspid a n n u l u s to the tip c h a n g e d RV E/A ratio from 1.35 +_ 0.4 to 1.71 _+ 0.6 (p < 0.001), w h e r e a s RV
Volume 130, Number 5 American Heart Journal
Cittadini et al.
1077
Table III. Scintigraphic findings Normals (n = 40) Left ventricle LV ejection fraction (%) LV p e a k ejection r a t e (EDV/sec) LV t i m e to p e a k filling r a t e (msec) LV p e a k filling r a t e (EDV/sec) LV p e a k filling r a t e (SV/sec) R i g h t ventricle RV ejection fraction (%) RV p e a k ejection r a t e (EDV/sec) RV t i m e to p e a k filling r a t e (msec) RV p e a k filling r a t e (EDV/sec) RV p e a k filling r a t e (SV/sec)
Hypertensives (n = 41)
p value
66 3.8 153 3.5 5.3
_+ 7 _+ 0.6 _+ 55 -+ 0.4 -+ 1.1
67 3.6 183 2.7 4.0
_+ 8 _+ 0.7 _+ 61 _+ 0.7 -+ 1.1
n.s. n.s. <0.02 <0.001 <0.001
58 2.5 133 2.9 5.0
_+ 8 _+ 0.8 _+ 45 +_ 0.4 _+ 0.8
57 2.4 160 2.2 3.8
_+ 7.8 _+ 0.6 _+ 70 _+ 0.7 _+ 1.0
n.s. n.s. <0.04 <0.001 <0.001
LV, Left ventricle; RV, right ventricle; EDV/sec, end-diastolic volumes/second; SV/sec, stroke volumes/second.
peak filling rate normalized to SV did not change (3.95 _+ 0.7 vs 3.91 ± 1, p -- n.s.) Reproducibility of echo-derived peak filling rate corrected by SV was as follows: intraobserver reproducibility 5% _+ 4% and 6% ± 4% for left and right ventricle, respectively; interobserver reproducibility was 7% -+ 5% and 7% _+ 6% for left and right ventricle, respectively. Gated blood-pool cardiac scintigraphy (Table III). Overall, indexes of LV and RV systolic function did not significantly differ between normotensive and hypertensive subgroups. Conversely, indexes of diastolic filling dynamics were impaired in both the left and right sides of the heart in t h e p a t i e n t s with systemic hypertension. In particular, average LV and RV peak filling rates were 2.7 and 2.2 EDV/sec as compared with 3.5 and 2.9 EDV/sec, respectively, in control subjects. LV and RV peak filling rates normalized to SVwere 4.0 - 1.1 and 3.8 ± 1.0 SV/sec in hypertension as compared with 5.3 _+ 1.1 and 5.0 ± 0.8 SV/sec in the controls, respectively (p < 0.01 for both). Fourteen (34%) of the 41 hypertensive subjects had an abnormal RV peak filling rate corrected by SV, whereas 17 (43%) had an abnormal LV peak filling rate. Univariate correlates of nuclear and Doppler-derived peak filling rates. The correlation between Doppler
and nuclear LV peak filling rates normalized to EDV was poor in the overall population (r = 0.42) (Fig. 1, A). Even worse was the correlation between echo and nuclear RV peak filling rate normalized to EDV (r = 0.34) (Fig. 1, B). On the contrary, there were excellent correlations between Doppler- and nuclearderived peak filling rate normalized to SV in the overall population (both hypertensives and normals) with r values of 0.92 (Fig. 1, C) and 0.87 (Fig. 1, D) for left and right ventricle, respectively. There were also excellent correlations in the hypertensive sub-
Table IV. Univariate correlations between RV and LV in-
dexes of filling Normals (n = 40) Echo p e a k filling r a t e (SV/sec) Echo p e a k filling r a t e (EDV/sec) N u c l e a r p e a k filling r a t e (SV/sec) N u c l e a r p e a k filling r a t e (EDV/sec)
r r r r
= = = =
0.68 0.51 0.65 0.61
Hypertensives (n = 41) r r r r
= = = =
0.66 0.49 0.63 0.57
RV, Right ventrieular; LV, left ventrieular; SV, stroke volume; EDV, enddiastolic volume.
jects alone, with r values of 0.89 for left ventricle and of 0.86 for right ventricle (p < 0.0001). Correlations between a variety of RV and LV indexes of filling. RV and LV peak filling rate indexes were
related (r = 0.49 to 0.68) (Table IV), suggesting that RV filling was often abnormal in patients with LV filling impairment (Fig. 2). These relations were found regardless of the technique or specific index used to measure filling or of the population in which they were examined. However, it should be noted that echo and peak filling rates corrected by SV produced closer relations between RV and LV filling in both normals and hypertensives than did peak filling rate corrected by EDV (r = 0.68 vs 0.51 in normals and r = 0.66 vs 0.49 in hypertensives) (Figure 3, A-D). Multivariate analysis of LV and RV diastolic filling.
Multivariate analyses of the determinants of peak filling rate by both echo and nuclear methods showed that, in both normals and hypertensives, normalization of peak filling rate to SVinstead of EDV removed the influence of ejection fraction on systolic function (Table V). In normals, assessment of both RV and LV filling dynamics by E/A ratio was affected by heart rate; this variable did not influence early peak filling rates normalized to SV. Filling dynamics in both
1078
November 1 9 9 5 American Heart Journal
Cittadini et al.
A
]B
I~. . . . . . . . . 9-°.'21
6y=O. 17x+2.44
+
4-
i 2
i 4
Echo-Doppler
C
~
4-
Z
2-
Echo-Doppler
LV-PFR (EDV/e)
D 1.18x-0.43
r=0.92
i
i
Echo-Doppler
•
i 6
8y=
]
r=0.34
LV-PFR (SV/s)
8-
~y
=1.22-0.41
RV-PFR (EDV/$)
r=0.87
i
I
I
I
I
2
4
6
8
Echo-Doppler
RV-PFR (SV s)
Fig, 1. Scatterplot of regression analysis of left ventricular estimated by echo-Doppler and nuclear techniques in both hypertensives (crosses, n = 41) and normal (dots, n = 40) populations. A and B, Left ventricular and right ventricular peak filling rates normalized to end-diastolic volume, respectively, are shown. C and D, Left ventricular (LV) and right ventricular (RV) peak filling rate (PFR) normalized to stroke volume (SV), respectively, are shown.
<3.5
ventricles in normals were affected by age (Fig. 4), regardless of method. In contrast, in hypertensive subjects, age and heart rate were not independently related to any filling parameter; their influence was superseded by t h a t of systolic blood pressure. Thus in the hypertensive population, the only independent determinant of the Doppler-derived RV peak filling rate normalized to SV was the LV filling rate.
8-
U) nu. D.
6-
•
+,o ¢,j
°
DISCUSSION
4-
o 0
.
.
.
.
.
.
.
.
.
.
.
6t o ul
2-
I
2
I I I I
I
I
I
4
6
8
Echo-Doppler
LV-PFR (SV/s)
Fig. 2. Scatterplot of left ventricular and right ventricular echo-detected peak filling rates values in hypertensives (crosses) and normal subjects (dots). D a s h e d lines indicate lower limits of left and right echo-detected peak filling rate normalized to stroke volume. Therefore, 19 and 16 hypertensives fell in abnormal range for left ventricular and right ventricular filling, respectively.
Although determination of the relative utility of any index of diastolic function is problematic because of the complexity of diastole and because of the lack of a gold standard, the results of this study suggest t h a t normalization of early peak filling rate to SV may be preferable to other methods in both right and left ventricles. In our study, this method eliminated the confounding influences on filling of systolic function, which confounds use of peak filling rate normalized to EDV, and of heart rate, which confounds use of the E/A ratio. Further, comparison between RV and LV filling rates was improved by normalization of Doppler-derived values to SV. The utility of this method in revealing pathophysiology was then demonstrated by application to a hypertensive pop-
Volume 130, Number 5 American
Cittadini et al.
HeartJournal
t
6y=O,65x+057 r=0.51
]
B
"6 g
•
~
O•
i
8-
1079
J~=o.,gx+2.o, r=9.~ i
64-
~ -
oo
~o u
Echo-Doppler nV-PFR (EDV/s)
C
6-
,,=
'=-
o.
1.0,~,07
.... 9 I
-t-
Echo-Doppler LV-PFR (SV/s)
D
6-
=.
+ 4-
+ +
.
,o
+ 2
Echo-Doppler LV-PFR (EDV/sv)
u 4
Echo-Doppler LV-PFR (SV~s)
Fig. 3. Scatterplot of regression analysis between right ventricular (RV) and left ventricular (LV) indexes of filling. A and B, Peak filling rate (PFR) normalized to end-diastolic volume (EDV) and stroke volume (SV), respectively, are considered in the normal population (dots). C and D, same indexes in hypertensive population (crosses) are shown.
ulation. Related abnormalities in both RV and LV filling were found, suggesting ventricular diastolic interdependence, and the well-described influence of age on filling was displaced by that of systolic blood pressure. Selection of the best index of filling. Doppler echocardiography and radionuclide ventriculography are the two most common noninvasive techniques that allow assessment of LV and RV diastolic filling. The correlation between these methods is rather poor when absolute or EDV-normalized peak filling rates are considered, 5, 6 perhaps because of the geometric assumptions necessary to estimate peak filling rates and their dependence on cavity size. When this technique is applied to the right ventricle, all these assumptions become even less reliable, as it is a structurally complex chamber, and poorly imaged. Thus it is not surprising that we found poor correlations between n u c l e a r - a n d echo-detected LV and especially RV peak filling rates normalized to EDV (r = 0.42 and r = 0.34, respectively). Therefore, a method of normalizing peak filling rate that does not depend on volume estimates would be highly desireable for both ventricles and would be particularly suited to RV peak filling rate assessment. Normalization to SV eliminates many of the sources of error of normalization to EDV. It is easily
calculated using both Doppler (E/total diastolic volume thickness index [VTI]) and nuclear (peak filling rate [PFR]/EDV + ejection fraction [EF])techniques, does not require visualization and measurement of annulus or chamber sizes, and does not depend on ejection fraction. 19, 20 The close correlations we found between Doppler and nuclear results in both left and right ventricles also help confirm the validity of our approach. Another index of filling, the E/A ratio, is also easy to obtain and less dependent on chamber size and systolic function. Nevertheless, this study suggests that there are several advantages of normalized peak filling rate relative to E/A ratio. These include independence of Doppler sample volume location, independence of H R influence in normal subjects, and values more comparable to those obtained by nuclear techniques. Although it is impossible definitively to determine the superiority of one technique over the other, our findings suggest that Doppler-derived peak early filling rates, when normalized to SV, are easy to obtain, accurate, and relatively independent of other physiologic influences, suggesting that this index is one of the best available. RV function in hypertension. RV filling dynamics have only recently been evaluated in normal subjects and in selected clinical settings, such as cardiac amyloidosis, acromegaly, and myocardial infarc-
November 1995 American Heart Journal
1080 Cittadini et al. Table V. Multivariate predictors of peak filling rates Dependent variable Normal subjects nLV-PFR (EDV/sec) nRV-PFR (EDV/sec) nLV-PFR (SV/sec) nRV-PFR (SV/sec) eLV-PFR (SV/sec) eRV-PFR (SV/sec) eLV-E/A eRV-E/A ' Hypertensive patients nLV-PFR (EDV/sec) nRV-PFR (EDV/sec) nLV-PFR (SV/sec) nRV-PFR (SV/sec) eLV-PFR (SV/sec) eRV-PFR (SV/sec) eLV-E/A eRV-E/A
Independent predictor(s)
Multiple r value
nLV-EF nRV-EF Age Age Age Age Age Age
nRV-PFR nLV-PFR nRV-PFR nLV-PFR eRV-PFR eLV-PFR H e a r t rate H e a r t rate
Age Age
nLV-EF nRV-EF SBP nLV-PFR SBP eLV-PFR SBP eLV-E/A
nRV-PFR nLV-PFR nRV-PFR
SBP
eRV-PFR eRV-E/A
eRV-E/A eLV-E/A
0.75 0.77 0.67 0.71 0.74 0.72 0.79 0.77 0.77 0.79 0.66 0.65 0.65 0.62 0.67 0.65
n, Nuclear; e, echocardiographic; LV, left ventricular; RV, right ventricular; PFR, peak filling rate; SV, stroke volume; EDV, end-diastolic volume; SBP, systolic blood pressure; EF, ejection fraction.
tion. 1-4 The complex anatomy of right ventricle, as well as the assumption that this chamber plays a minor role in overall cardiovascular function, have limited investigations about t h e effects of systemic pressure overload on RV function. Nevertheless, two recent studies have reported abnormal Doppler echocardiographic RV variables in hypertensive subjects as compared with age- and sex-matched healthy normals. 7,s The difficulty of obtaining reliable echocardiographic estimates of RV volumes led these investigators to exclude analysis of RV systolic function, This omission is significant, as it is reasonable to assume that RV systolic and diastolic function are related, as is the case with the LV. Moreover, no study thus far has constructed a multivariate analysis to find out which variables are truly independent in predicting RV peak filling rates. RV ejection fraction has been reported to be either normal or decreased in systemic hypertension. 9, 10 Our findings of normal resting values of RV ejection fraction in our patients are consistent with previous observations that, in the absence of LV dysfunction, RV ejection fraction is usually normal. 1° In spite of a normal ejection fraction, this study shows that impairment of RV diastolic filling is a common finding in systemic hypertension. In particular, as far as peak filling rate corrected by SV is concerned, 34% and 39% of the hypertensive patients had RV filling abnormalities by Doppler and nuclear techniques, respectively, as compared with an only slightly higher prevalence of LV filling impairment (46% and 43%, respectively). In spite of estimates of similar RV cavity size in
normals and hypertensives, RV hypertrophy was consistently present in our patients. This finding has been previously reported, 7 although the pathogenesis is still an open issue. It has been proposed that hemodynamic overload of the right ventricle might result from elevated LV diastolic pressure, or that unidentified local or circulating growth factors such as catecholamines, angiotensin, insulin-like growthfactor-l, and so on, may act on both ventricles. 21 Because we were unable to make an accurate noninvasive measurement of RV or LV pressures in our ambulatory patients (see Methods), we could provide no data correlating either RV hypertrophy or RV filling indexes to pulmonary or left heart pressures. Relations between RV and LV filling. An important finding of this study is that RV and LV filling indexes are closely interrelated in both the normal and hypertensive populations. Those patients with the most markedly impaired LV diastolic function also had the most striking RV abnormalities. Several lines of evidence support the concept of a mechanical and hemodynamic interaction between right and left ventricles, both in diastole and in systole. 22 From an anatomic standpoint, the two ventricles share common circular and spiral muscle fiber bundles, interventricular septum, and a common pericardium. Furthermore, the pulmonary circulation indirectly connects the two ventricles. Although our finding of close correlations between LV and RV indexes of filling by both indexes provides further in vivo evidence of this diastolic interplay, the results of the multivariate analysis are even more striking. For both LV and RV, and in both hypertensives and normals, ev-
Volume 130, Number 5 American Heart Journal
Cittc~ini et al.
n
U
7-
1081
6-
,,= a.
5r=0.70
r=0.74
o 6
~-+
:,
~ _ + j~H ~ r=0.40
, o
r=0.31
3-
4++ 2
2'0
20
go
8'0
Age
2
2'0
,'o
+ 8'0
Age
Fig. 4. Univariate correlation between age and left ventricular peak filling rate normalized to stroke volume (LV-PFR) in A; in B correlation between age and right ventricular peak filling rate normalized to stroke volume (RV-PFR) is shown. Crosses, Hypertensives; dots, normals.
ery index of filling had, as an independent determinant, the corresponding filling index of the opposite chamber. Thus RV filling was dependent on LV filling and vice versa. Determinants of LV and RV filling in hypertension. Although age is well described to influence LV filling in
normals, 23 its importance in disease states is unclear. Our findings are consistent with other studies that found that, in the presence of cardiomyopathy and hypertension, the effects of age on LV filling are no longer detected. 24 To our knowledge, no previous studies have examined the multiple influences of age and disease state on RV function. Our results demonstrate that, in the normal right ventricle, similar to the normal left ventricle, age produces intrinsic changes in myocardial stiffness or relaxation or both so as to impair RV filling 25, 26 However, we have also shown t h a t the effects of hypertension include abnormal RV as well as LV filling. This is of such a magnitude that the effects of hypertension on the RV can supersede those of normal aging. This is particularly remarkable because the primary hemodynamic abnormality is one of systemic pressure overload, whose effects should be more marked on the left ventricle. Study limitations. Because the validation of our method required comparison of several techniques in the same individual, m a n y of the patient-related factors t h a t affect diastole should have been constant (including HR, left atrial and LV pressures, PR interval, etc.) and not influenced our findings. In this study, we chose an ellipsoidal model for estimation of RV volumes; although the RV is clearly not a prolate ellipse, this formula does include the infundibular contribution to volume, is relatively simple, and has been demonstrated to yield an excellent correlation (r = 0.95) with h u m a n RV cast measurements. 16 Several different approaches have been proposed for calculation of RV volume, but none of them appears
to offer a superior estimate, and they usually require additional calculations. Further, most of our conclusions are based on the use of peak filling rate normalized to SV, a method that eliminates dependence on RV volume estimation. Although equilibrium radionuclide angiography does not suffer from the same geometric assumptions as does 2D echocardiography, a source of error has been described in difficulties in separating the right atrium from the right ventricle. We used a modified left anterior oblique position to avoid underestimation of RV ejection fraction; in addition, any underestimation should have been similarly present in both the study groups. Measurements of RV freewall thickness by echocardiography have been reported to have a high variability; however, previous data from our laboratory showed an acceptable reproducibility.4 Clinical implications. This study demonstrates that of the several indexes of RV filling that can be easily measured by clinically available techniques, RV peak filling rate normalized to SV might be preferred because of ease of calculation, repeatability, freedom from physiologic influences, and an excellent correlation between echo and radionuclide angiographic assessments. In applying this method, we showed that more t h a n one third of hypertensive subjects had RV filling abnormalities. Although altered RV filling has been described in several disease states, 1-4 the clinical relevance and the prognostic significance of RV diastolic impairment is unclear. In this respect, the availability of better indexes of RV filling, such as peak filling rate normalized to SV, should facilitate a detailed noninvasive assessment of RV diastolic function in various clinical settings and a better understanding of its importance. The fact that LV and RV filling are both dependent on diastolic performance in the other chamber strongly supports the concept ofventricular diastolic
1082
November 1995 American Heart Journal
Cittadini et al.
interdependence. Whether this interaction is the result of anatomic factors or extrinsic factors (i.e., hormonal influences acting on both ventricles) or both is unclear. However, it is readily apparent that RV abnormalities can be just as much a part of the complex of cardiac effects of systemic hypertension as are LV filling abnormalities, and that the evaluation of RV filling may provide additional insight into LV diastolic function. Also clinically relevant is the effect of age on RV filling in normal subjects; this finding suggests that determinations of right ventricular filling should include reference to age-related standards. However, the removal of such effects in the hypertensive population must also be taken into account when assessing RV filling in disease states; further research is necessary to determine the relative importance of age and disease in determining RV filling. Conclusions. In this study we examined the utility of echo-Doppler-derived peak filling rate normalized to tricuspid SV as an index of RV diastolic filling in normal and hypertensive populations. This index appears to be particularly suitable for the investigation of RV filling dynamics because of the excellent correlation with radionuclide angiographic techniques, the ease of calculation, and relative independence of geometric assumption, ventricular size, sample volume location, heart rate, and systolic function. Because RV filling abnormalities are common in hypertension and appear related to both LV abnormalities and the severity of systemic blood pressure elevation, RV diastolic abnormalities should be further investigated as an important component of hypertensive heart disease and a manifestation of ventricular interaction. REFERENCES
1. Nishimura RA, Abel MD, Hatle LK, Tajik AJ. Assessment of diastolic function of the heart: background and current application of Doppler echocardiography. Part II. Clinical studies. Mayo Clin Proc 1989;64:181204. 2. Klein AL, Hatle LK, Burstew DJ, Taliercio CP, Seward JB, Kyle RA, Bailey KR, Gertz MA, Tajik AJ. Comprehensive Doppler assessment of right ventricular diastolic function in cardiac amyloidosis. J Am Cell Cardiol 1990;15:99-108. 3. Fujii J, Yazaki Y, Sawada H, Aizawa T, Watanabe H, Kate K. Noninvasive assessment of left and right ventricular filling in myocardial infarction with a two-dimensional Doppler echocardiographic method. J Am Cell Cardiol 1985;5:1155-60. 4. Fazio S, Cittadini A, Sabatini D, Merola B, Colao AM, Biondi B, Lombardi G, Sacch L. Evidence for biventricular involvement in acromegaly: a Doppler echocardiographic study. Eur Heart J 1993;14:26-33. 5. Friedman BJ, Drinkovic N, Miles H, Shih WJ, Mazzoleni A, DeMaria
6.
7.
8.
9. 10.
11. 12.
13.
14.
15.
16.
17.
18.
19. 20.
21. 22. 23.
24.
25.
26.
AN. Assessment of left ventricular diastolic function: comparison of Doppler echocardiography and gated blood pool scintigraphy. J Am Cell Cardiol 1986;8:1348-54. Bowman LK, Lee FA, Jaffe CC, Mattera J, Wackers FJ, Zaret BL. Peak filling rate normalized to mitral stroke volume: a new Doppler echocardiographic filling index validated by radionuclide angiographic technique. J Am Cell Cardiol 1988;12:937-43. Chakko S, de Marchena E, Kessler KM, Materson BJ, Myerburg RJ. Right ventricular diastolic function in systemic hypertension. Am J Cardiol 1990;65:1117-20. Habib GB, Zoghbi WA. Doppler assessment of right ventricular filling dynamics in systemic hypertension: comparison with left ventricular filling. A~ HEARTJ 1992;124:1313:20. Ferlinz J. Right ventricular performance in essential hypertension. Circulation 1980;61:156-62. Nunez BDI Messerli FH, Amodeo C, Garavaglia GE, Frohlic ED. Biventricular cardiac hypertrophy in essential hypertension. AMHEART J 1987;111:813-18. Neubar T, ScwartzkopffB, Strauer BE. Right ventricular performance in arterial hypertension. Eur Heart J 1992;13(suppl D):33-8. Kirkendall WM, Burton HC, Epstein FH, Freis ED. Recommendation for human blood pressure determination by sphygmomanometer. Circulation 1967;36:980-8. Sahn DJ, DeMaria A, Kisslo J, Weyman A. The committee on M-mode standardization of the American Society of Echocardiography. Recommendations regarding quantification in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation 1978; 58:1072-83. Devereux RB. Detection of left ventricular hypertrophy by M-mode echocardiography. Anatomic validation, standardization and comparison to other methods. Hypertension 1987;9(suppl II):19-26. Tortoledo FA, Quinones MA, Fernandez GC, Waggoner AD, Winters WL Jr. Quantification of left ventricular volume by two-dimensional echocardiography: a simplified and accurate approach. Circulation 1983;67:579-84. Levine RA, Gibson TC, Aretz T, Gillam LD, Guyer DE, King ME, Weym a n AE. Echocardiographic measurement of right ventricular volume. Circulation 1984;69:497-505. Cuocol0 A, Frederick LS, Brush JE, Maron BJ, Bacharach SL, Bonow RO. Left ventricular hypertrophy and impaired diastolic filling in essential hypertension. Circulation 1990;81:978-86. Legrand V, Checigne M, Foulon J, Rigo P. Evaluation of right-ventricular function by gated blood-pool scintigraphy. J Nucl Med 1983;24:88693. Bonow RO. Radionuclide angiographic evaluation of left ventricular diastolic function. Circulation 1991;84(suppl I):I-208-I15. Smith VE, Schulman P, Karimeddini MK, White WB, Meeran MK, Katz AM. Rapid ventricular filling in left ventricular hypertrophy: II. Pathologic hypertrophy. J Am Cell Cardiol 1985;5:869-74. Morgan HE, Baker KM. Cardiac hypertrophy. Mechanical, neural, and endocrine dependence. Circulation 1991;83:13-25. Bove A, Santamore W. Ventricular interdependence. Prog Cardiovasc Dis 1981;23:365-87. Benjamin EJ, Levy D, Anderson KM, Wolf PA, Plehn JF, Evans JC, Comai K, Fuller DL, St. John Stutten M. Determinants of Doppler indexes of left ventricular diastolic function in normal subjects (the Framingham Heart Study). Am J Cardio] 1992;70:508-15. Sartory MP, Quinones MA, Kuo LC. Relation of Doppler-derived left ventricular filling parameters to age and radius/thickness ratio in normal and pathologic states. Am J Cardiol 1987;59:1179-82. Berman GO, Reichek N, Br°ws°n D, D°uglas PS" Effects °f sample v°lume location, imaging view, heart rate and age on tricuspid velocimetry in normal subjects. Am J Cardiol 1990;65:1026-30. Zoghbi WA, Habib GB, Qninones MA. Doppler assessment of right ventricular filling in a normal population: comparison with left ventricular filling dynamics. Circulation 1990;82:1316-24.