Genesis of the restrictive filling pattern: Pericardial constraint or myocardial restraint

Genesis of the Restrictive Filling Pattern: Pericardial Constraint or Myocardial Restraint Steven J. Lavine, MD, FACC, Jacksonville, Florida

Background: Restrictive filling pattern has been predictive of heart failure in patients with cardiomyopathy and after myocardial infarction, and is similar to the filling pattern in constrictive pericarditis and amyloid heart disease. The purpose of this study was to determine the role of both myocardial restraint and pericardial constraint in a chronic left ventricular dysfunction model with restrictive filling. Methods: After instrumentation, a flat balloon containing a high-fidelity pressure catheter was inserted through a pericardial incision in 12 dogs with chronic left ventricular dysfunction. Intracardiac volume (ICV) was manipulated by inferior venal caval balloon occlusion and volume loading while hemodynamics, echo-assessed chamber size, and transmitral Doppler were obtained at the same atrial paced rate with an intact pericardium and after pericardiectomy. Results: With an intact pericardium, deceleration time increased with reduced ICV (130 ⴞ 35 vs 153 ⴞ

Restrictive filling pattern (RFP) has been associated

with increased heart failure and mortality in patients postmyocardial infarction and with dilated cardiomyopathy.1-3 RFP has also been demonstrated in both amyloid heart disease and constrictive pericarditis.4,5 However, the genesis for this diastolic filling pattern has not been elucidated. Previous clinical and experimental data have suggested that the pericardium exerts restraining properties on left ventricular (LV) size, influencing the diastolic filling pattern both acutely and chronically.5-10 Using a model of acute dilated ischemic LV dysfunction,10 the extent of early diastolic filling increased but was redistributed to later in diastole after pericardiectomy.10 Using this same model, nitroglycerin reduced LV From Wayne State University and University of Florida/Jacksonville. Supported by a Grant-In-Aid from the American Heart Association of Michigan. Reprint requests: Steven J. Lavine, MD, Cardiovascular Center, 655 W Eighth St, Jacksonville, FL 32209 (E-mail: [email protected]). 0894-7317/$30.00 Copyright 2004 by the American Society of Echocardiography. doi:10.1067/j.echo.2003.10.025

152

47 milliseconds, P < .05) and shortened with increased ICV (107 ⴞ 45 milliseconds, P < .05). The filling fraction at one-third of diastole decreased with reduced ICV (45.6 ⴞ 29.3 vs 24.2 ⴞ 15.8%, P < .01) and increased with increased ICV (60.1 ⴞ 14.8%, P < .05). Deceleration time could be predicted from intrapericardial pressure, the transmural left ventricular chamber stiffness constant, and filling fraction at one-third of diastole. After pericardiectomy, deceleration time also shortened with increased ICV (141 ⴞ 26 vs 112 ⴞ 38 milliseconds, P < .01). However, filling fraction at one-third of diastole was markedly reduced at paced baseline (19.9 ⴞ 14.4%, P < .01) and with increased ICV (15.5 ⴞ 11.8%, P < .001) as compared with an intact pericardium. Conclusions: Pericardial constraint and myocardial restraint play a role in restrictive filling pattern. Pericardial constraint becomes evident with redistribution of diastolic filling to later in diastole after pericardiectomy. (J Am Soc Echocardiogr 2004;17: 152-60.)

size and altered the diastolic filling pattern similarly.11 For patients with extensive myocardial amyloid deposition, RFP was observed suggesting a role for myocardial restraint.4 Finally, for patients with heart failure, nitrate reduction in LV filling pressures resulted in parallel downward displacement of the LV pressure-volume plot suggesting relief of pericardial restraint.8 It is our hypothesis that the RFP is a result of external constraint by the pericardium and, to a lesser extent, by myocardial restraint produced by a dilated and scarred LV myocardium. Therefore, the purpose of this study was to determine the role of both myocardial restraint and pericardial constraint in a model of chronic LV dysfunction characterized by LV dilatation, myocardial fibrosis, and a RFP that was similar to ischemic cardiomyopathy.12

METHODS The animals used in this study were maintained in accordance with the guidelines of the committee on animal studies at Wayne State University School of Medicine, Detroit, Mich, and with the position of the American Heart

Journal of the American Society of Echocardiography Volume 17 Number 2

Association on research animal use. Anesthesia was induced in 12 conditioned mongrel dogs (21-27 kg) with intramuscular morphine sulfate (1.5 mg/kg) and acepromazine (1.1 mg/kg) followed in 15 minutes by 30 mg/kg of intravenous ketamine hydrochloride. Maintenance anesthesia was produced by intravenous morphine sulfate (1.5 mg/kg/h) and pentobarbital (3 mg/kg/h). The dogs were intubated and artificially ventilated with a Harvard respirator using room air. Using fluoroscopic guidance, 2 7F high-fidelity catheters (Millar Instruments, Houston, Tex) were introduced by the right carotid artery and advanced to the LV and ascending aorta. A No. 8 multipurpose Judkins catheter was introduced through a sheath (Cordis, Miasi Lakes, Fla) into the right femoral artery and advanced into the left coronary ostium. A No. 7F thermodilution pulmonary artery catheter was advanced to the pulmonary artery. The proximal port was used to measure right atrial pressure. Continuous electrocardiographic monitoring was performed using lead II. At held end-expiration, electrocardiography, LV pressures, dP/dt, central aortic pressures, and mean right atrial pressure were obtained at 100 mm/s using an 8-channel physiologic recorder (Gould, Eastlake, Ohio). Simultaneous 2-dimensional echocardiograms and Doppler were obtained with the use of a phased-array echocardiograph (Aloka 880, Tokyo, Japan). Transesophageal 4-chamber view was obtained from a 5-MHz probe placed in the midesophagus and transmitral pulsed Doppler recordings were obtained from beyond the tips of the mitral leaflets in the LV at 100 mm/s. The left atrium was examined with color flow Doppler for mitral regurgitation. Induction of LV Dysfunction Ischemic LV dysfunction was induced by left main coronary artery plastic microsphere injections (58 ⫾ 2 ␮m) (3M, Minneapolis, Minn) injected in boluses of 17,500 microspheres every 5 to 10 minutes until the peak positive dP/dt was reduced by ⬎25% and the LV end-diastolic pressure (LVEDP) was ⬎12 mm Hg. Moderate LV dysfunction (ejection fraction ⬍ 40%) was produced in 45 to 60 minutes with only mild mitral regurgitation as determined by a maximal jet area/left atrial area ⬍ 20%. This approach led to a model of chronic moderate LV dysfunction characterized by patchy interstitial and replacement fibrosis throughout the myocardium.12 The right carotid and femoral arteries were repaired and the dogs were allowed to recover without any dog dying. At 8 weeks postcoronary microsphere embolization, the animals were intubated, ventilated, and anesthesia was induced as above. Using fluoroscopy, high-fidelity catheters (Millar Instruments) were advanced to the LV, ascending aorta, and right ventricle (RV). A No. 7F thermodilution pulmonary artery catheter and No. 5 bipolar pacing wire were advanced from the internal jugular to the pulmonary artery and right atrium, respectively. Atrial pacing commenced 5 beats above baseline rate with a PR interval ⬍ 160 milliseconds. The pacing catheter was repositioned if needed to insure that the PR interval remained ⬍160 milliseconds. Transthoracic and trans-

Lavine 153

esophageal echocardiography, and Doppler, was obtained as above. Phonocardiogram recordings of heart sounds were obtained at 100 mm/s. To reduce intracardiac volume (ICV), a balloon catheter (Fogarty, Bacchus Vascular, Santa Clara, Calif) was inserted through the left femoral vein and positioned in the inferior vena cava. The balloon was expanded with air to demonstrate inferior vena cava occlusion confirmed both fluoroscopically and by a decrease in right atrial and RV end-diastolic pressure (RVEDP). Continuous electrocardiographic monitoring was performed using lead II. Intrapericardial Pressure Measurements and Pericardiectomy A thoracotomy was performed after an additional dose of pentobarbital (15 mg/kg). A flat intrapericardial latex balloon containing 1.5 mL of saline and a high-fidelity pressure catheter (Millar Instruments) was inserted through an incision in the pericardium over the LV anterior wall and 0 referenced to fluid pressure.13,14 The thoracotomy was closed and the lungs were re-expanded using a chest tube. The dogs were allowed to become hemodynamically stable for at least 60 minutes. Atrial pacing commenced at 5 beats above the new basal rate. The above hemodynamic, echocardiographic, and Doppler parameters were obtained at held end-expiration (to account for the effects of ventilation and intrapleural pressure). Care was taken to avoid merging of the transmitral peak rapid filling velocity (E) and peak atrial filling velocity (A) wave by positioning the pacing wire to maintain a PR interval ⬍ 160 milliseconds. Total ICV was manipulated first by expanding the inferior vena cava balloon for at least 3 minutes. Hemodynamics, echocardiography, and transmitral Doppler were obtained when RVEDP had declined ⱖ 2 mm Hg but the mean aortic pressure had declined ⬍ 10 mm Hg. If the mean arterial pressure decreased ⬎ 10 mm Hg before 3 minutes, the balloon was minimally deflated to maintain a mean arterial pressure decrease ⬍ 10 mm hg. After acquisition of the above parameters, the inferior vena cava balloon was deflated and hemodynamics were allowed to return to baseline for at least 15 minutes before the above parameters were obtained. ICV was increased by infusion of 10 mL/kg of normal saline over 15 minutes. Hemodynamics, echocardiography, and transmitral Doppler were then obtained. The pericardium was then widely excised. The chest was closed, and the lungs were re-expanded. The dogs were observed until hemodynamics returned to baseline for at least 60 minutes. Total ICV was again manipulated as above, and hemodynamic, echocardiographic, and Doppler parameters were obtained similarly. The dogs were then killed. Hemodynamic, Echocardiographic, and Transmitral Doppler Measurements For all stages, LV pressures, dP/dt, aortic pressures, mean right atrial pressure, and RV pressures were measured

Journal of the American Society of Echocardiography February 2004

154 Lavine

from the average of 3 consecutive cycles at held endexpiration. As intrapericardial pressure (IPP) is a contact pressure, it varies throughout the cardiac cycle and was measured at end-diastole. There are regional variations in IPPs measured with a flat balloon, with pressures highest around the lateral LV wall and intermediate around the anterior and posterior LV wall. The anterior wall was chosen because of stability of positioning. Transmural LVEDP was determined by subtracting the IPP from the intracardiac LVEDP (IPP ⫽ 0 after pericardiectomy). The time constant of LV pressure decline was calculated using the method of Weiss et al.15 LV volumes at end-diastole and systole were obtained using the modified method of disks. Left atrial, right atrial, and RV volumes were obtained using the area-length formula from the transesophageal 4-chamber view. Frameby-frame analysis of LV endocardial areas (30 frames/s time resolution) in the parasternal short-axis and apical 4-chamber views were performed using an offline analysis system (Imagevue, Microsonics, Bothell, Wash). LV enddiastole was the largest endocardial area and end-systole was the smallest endocardial area. Ejection fraction was calculated in the standard fashion. ICV was calculated as: ICV ⫽ LV end-diastolic epicardial volume ⫹ left atrial volume ⫹ right atrial volume ⫹ RV end-diastolic volume. LV epicardial volume at end-diastole was calculated using the modified method of disks applied to the enddiastolic epicardial area. For transmitral Doppler indices, E and A were measured. The E/A ratio was calculated. The rapid filling deceleration time (DCT) was calculated as the time interval from E to the time mitral flow decelerated to the 0 baseline. The tracing was extrapolated to the 0 baseline if atrial filling commenced before mitral flow fully decelerating to 0. Diastolic, rapid filling, and atrial filling velocity integrals were determined. The RFP has been defined as a rapid filling DCT ⬍ 150 milliseconds, an isovolumic relaxation period ⬍ 60 milliseconds, and E/A ratio ⬎ 2.16 We chose to use a DCT ⬍ 150 milliseconds as the single criterion for the existence of restrictive filling. LV pressure short-axis area composite plots were constructed for each stage from mean LV pressures and short-axis endocardial areas obtained at LV end-systole, LV pressure minimum, LV pressure before the A wave, and at end-diastole. LV short-axis areas were obtained from frame-by-frame outline of endocardial borders from 3 consecutive cycles (30 frames/s). The time interval from end-systole (aortic notch pressure) to LV pressure minimum and to the onset of the LV pressure A wave was determined. As each videoframe is 33 milliseconds apart, the time interval divided by 33 milliseconds determined the number of frames from end-systole. The associated LV endocardial area was determined using linear interpolation between videoframes. As a measure of LV chamber stiffness, an estimate of the chamber stiffness constant

was calculated using the approach of Marino et al.17 Essentially, the difference between LV pressure minimum and LVEDP is divided by the change in endocardial area from the time of LV pressure minimum to end-diastole. This calculation (using LV volume) has correlated moderately well with DCT.17 The filling fraction at one-third of diastole (FF1/3D), a measure of early diastolic filling, was also determined as the LV endocardial area at one-third of diastolic time minus end-systolic area divided by the stroke area (enddiastolic area minus end-systolic area). Linear interpolation was used if one-third diastolic time was between videoframes. The diastolic time (12 ⫾ 2 frames) was equal to the number of frames between end-systole and the next end-diastole. Intraobserver and interobserver variability for LV cavity area was determined by randomly choosing 50 frames spanning the cardiac cycle from 10 previously studied dogs. Each frame was analyzed at least 3 weeks apart by the 2 observers (see “Acknowledgement”). For intraobserver variability, the average difference for a given frame was 0.23 cm2 (or 2% of LV cavity area). For interobserver variability, the average difference for a given frame was 0.29 cm2 (or 2.5% of LV cavity area). Intraobserver and interobserver variability was also determined for total ICV by randomly calculating this parameter in the above 10 previously studied dogs with chronic LV dysfunction 3 weeks apart by 2 observers. For intraobserver and interobserver variability, the average difference between determinations was, respectively, 11 ⫾ 7 mL or 7.5 ⫾ 2.6% and 13 ⫾ 8 mL or 8.4 ⫾ 3.9% of total ICV. Statistics All data were expressed as mean ⫾ SD. Differences between a variable among stages was assessed using analysis of variance for repeated measures. If the F statistic indicated a significant difference existed (P ⬍ .05), then the Tukey test was used to determine where the significant differences existed. A P value ⬍ .05 was considered to be significant. The relationship between variables was determined using least squares linear regression. Forward stepwise multiple linear regression was used to determine the predictors of DCT. All variables for which the relationship with DCT had a value of P ⬍ .10 were included in the modeling.

RESULTS Table 1 summarizes the hemodynamics, echocardiographic chamber volumes, and diastolic indices at baseline and after induction of chronic LV dysfunction. As expected, with coronary microsphere embolization, LV dilatation, systolic dysfunction, elevated LVEDP, and RFP (shortened DCT) was noted. Table 2 summarizes the hemodynamics, echocardiographic chamber volumes, and diastolic indices with changes in loading conditions that altered total ICV after the placement of the intrapericardial bal-

Journal of the American Society of Echocardiography Volume 17 Number 2

Table 1 Left ventricular volume, hemodynamics, and transmitral Doppler at baseline and 8 weeks postembolization Base

EDV ESV EF RR LVEDP LVSP Peak ⫹ dP/dt Peak - dP/dt Tau E/A Decel time IRP

mL cc mL Percentage msec mm Hg mm Hg mm Hg/s mm Hg/s msec msec msec

56 ⫾ 15 20 ⫾ 7 64 ⫾ 6 849 ⫾ 190 4⫾2 113 ⫾ 14 1898 ⫾ 330 1613 ⫾ 228 23 ⫾ 6 1.6 ⫾ 0.4 191 ⫾ 22 52 ⫾ 23

Chronic LV dysfunction

68 ⫾ 25* 42 ⫾ 16‡ 38 ⫾ 12‡ 663 ⫾ 107‡ 12 ⫾ 3† 115 ⫾ 11 1422 ⫾ 266† 1395 ⫾ 187* 41 ⫾ 8‡ 1.8 ⫾ 0.5 133 ⫾ 37‡ 46 ⫾ 31

A, Peak atrial filling velocity; decel time, mitral valve deceleration time; E, transmitral peak rapid filling velocity; EDV, end-diastolic volume; EF, ejection fraction; ESV, end-systolic volume; IRP, isovolumic relaxation period; LV, left ventricular; LVEDP, LV end-diastolic pressure; LVSP, LV systolic pressure. *P ⬍ .05; †P ⬍ .01; ‡P ⬍ .001.

loon. As expected, total ICV and LV volume, LVEDP, transmural LVEDP, RVEDP, and IPP declined with inferior vena cava obstruction and increased with volume loading. In a given dog, the PR interval varied ⱕ20 milliseconds (average difference was 12 ⫾ 3 milliseconds) and varied from 130 to 160 milliseconds. With decreased ICV, there was DCT prolongation and reduction of FF1/3D. With volume loading, the DCT shortened and FF1/3D increased despite prolongation of the isovolumic relaxation period and ␶. The transmural LV chamber stiffness constant increased with volume loading. Figure 1 depicts the DCT at baseline, with decreased ICV and increased ICV. In 10 of 12 dogs, there was both an increase in DCT with a reduction in ICV and a decrease in DCT with an increase in ICV. This diastolic filling pattern is suggestive of pericardial constraint and is associated with increased IPP and ICV. DCT was predicted (R ⫽ 0.773 or R2 ⫽ 0.595) from the IPP (P ⫽ .0013), transmural LV chamber stiffness constant volume (P ⫽ .0016), and FF1/3D (P ⫽ .044). The relation of LV transmural chamber stiffness constant to DCT suggests an element of myocardial restraint. The influence of changes in loading conditions that alter total ICV after pericardiectomy are summarized in Table 3. DCT declined with volume loading and was associated with an increase in ␶, the isovolumic relaxation period, and the transmural LV chamber stiffness constant. However, FF1/3D remained low as compared with similar stages with an intact pericardium (Table 2). Figure 2 examines the DCTs with alterations in ICV after pericardiectomy. Of 12 dogs, 6 demonstrated both an increase in DCT with reduced ICV and a decrease with volume

Lavine 155

loading suggesting myocardial restraint. Multiple linear regression revealed that DCT was best predicted from LVEDP (R ⫽ ⫺0.52, P ⬍ .01) although transmural LV chamber stiffness constant correlated with DCT (R ⫽ 0.44, P ⬍ .01). In a given dog, the PR interval varied ⱕ 20 milliseconds (average difference was 12 ⫾ 4 milliseconds) and varied from 120 to 160 milliseconds. Table 4 summarizes hemodynamics, cardiac volumes, and diastolic indices at baseline and with increased ICV in dogs with and without an intact pericardium. At baseline and with volume loading, the LV end-diastolic volume and ICV were increased after pericardiectomy and were associated with a decline in FF1/3D as compared with an intact pericardium. Volume loading further accentuated differences in the FF1/3D, which was lower in the pericardiectomized dogs (Figure 3). The transmural LV chamber stiffness constant was greater at baseline and with volume loading with an intact pericardium. Figure 4 plots composite LV pressure shortaxis area for dogs with and without an intact pericardium and after increased ICV. Both LV pressure-area plot curves demonstrate a somewhat parallel upward displacement (range: 1.1-1.5 mm Hg and 1.5-2 mm Hg) at paced baseline (P ⫽ .08) and with volume loading (P ⬍ .05) in the dogs with an intact pericardium. Corresponding LV areas during diastole at the same LV pressure were 1.0 ⫾ 0.3 cm2 larger after pericardiectomy (P ⫽ .06) at paced baseline and 2.8 ⫾ 1.1 cm2 larger with volume loading postpericardiectomy (P ⬍ .001). At the same LV short-axis area, LV diastolic pressures were 2.6 ⫾ 1.1 mm Hg lower at paced baseline (P ⬍ .01) and 2.6 ⫾ 1.0 mm Hg lower with volume loading (P ⬍ .01) after pericardiectomy. Larger diastolic LV areas at lower diastolic pressures after pericardiectomy demonstrates the constraining effects of the pericardium. IPPs were higher than would be calculated from the difference in LVEDP from before and after pericardiectomy as a result of the increased LV size after pericardiectomy.

DISCUSSION We hypothesized the RFP was a result of both pericardial constraint because of dilated cardiac chambers and myocardial restraint from scarred myocardium. To test this hypothesis, we chose a model of chronic LV dysfunction characterized by interstitial and replacement fibrosis, LV dilatation, and systolic dysfunction12,18 that histopathologically resembled that seen with ischemic dilated cardiomyopathy. We induced only moderate LV dysfunction with no more than mild mitral regurgitation by limiting the number of injections to produce sufficient necrosis and ultimate remodeling resulting in a

Journal of the American Society of Echocardiography February 2004

156 Lavine

Table 2 Measures of cardiac volume, function, and diastolic function with intracardiac volume changes EDV ESV EF ICV LVEDP LV sys pressure Mean art pressure RA pressure RVEDP RV Sys pressure IPP E/A RR Decel time FF1/3D IRP Tm LVEDP Tm LV stiffness Tau

mL mL Percentage mL mm Hg mm Hg mm Hg mm Hg mm Hg mm Hg mm Hg msec msec Percentage msec msec mm Hg/mL msec

IP paced I

IP decreased volume

IP paced II

IP increased volume

63 ⫾ 15 37 ⫾ 16 42 ⫾ 14 148 ⫾ 24 10.2 ⫾ 2.9 100 ⫾ 9 91 ⫾ 8 6.2 ⫾ 3.1 6.9 ⫾ 3.2 28 ⫾ 4 4.3 ⫾ 2.9 1.8 ⫾ 0.4 618 ⫾ 47 130 ⫾ 35 45.6 ⫾ 29.3 53 ⫾ 24 4.6 ⫾ 2.5 0.32 ⫾ 0.12 44 ⫾ 14

49 ⫾ 13* 26 ⫾ 9* 47 ⫾ 12 115 ⫾ 27* 5.7 ⫾ 3.1† 90 ⫾ 8* 78 ⫾ 11† 4.0 ⫾ 2.4* 4.2 ⫾ 2.3* 24 ⫾ 4† 2.3 ⫾ 3.0† 1.7 ⫾ 0.7 618 ⫾ 47 153 ⫾ 47* 24.2 ⫾ 15.8† 43 ⫾ 23 3.4 ⫾ 3.3* 0.12 ⫾ 0.07† 42 ⫾ 10

61 ⫾ 16 35 ⫾ 14 43 ⫾ 16 152 ⫾ 26 9.8 ⫾ 2.4 103 ⫾ 7 93 ⫾ 10 6.0 ⫾ 2.9 6.6 ⫾ 2.7 27 ⫾ 3 4.1 ⫾ 2.5 1.7 ⫾ 0.3 618 ⫾ 47 133 ⫾ 31 43.1 ⫾ 27.9 50 ⫾ 26 4.4 ⫾ 2.3 0.32 ⫾ 0.11 45 ⫾ 11

81 ⫾ 26*¶ 47 ⫾ 24*㛳 42 ⫾ 16 192 ⫾ 32†¶ 14.9 ⫾ 3.2†¶ 104 ⫾ 13¶ 94 ⫾ 15¶ 10.1 ⫾ 4.5‡¶ 9.9 ⫾ 4.2‡¶ 29 ⫾ 4㛳 8.3 ⫾ 3.5‡¶ 2.2 ⫾ 1.0*§ 618 ⫾ 47 107 ⫾ 45*㛳 60.1 ⫾ 14.8*¶ 69 ⫾ 27*㛳 6.6 ⫾ 2.1*㛳 0.41 ⫾ 019*¶ 52 ⫾ 12†¶

A, Peak atrial filling velocity; art, arterial; Decel time, deceleration time; E, transmitral peak rapidfilling velocity; EF, ejection fraction; EDV, end-diastolic volume; ESV, end-systolic volume; FF1/3D, filling fraction at one third of diastole; ICV, intracardiac volume; IP, intact pericardium; IPP, intrapericardial pressure; IRP, isovolumic relaxation period; LV, left ventricular; RA, right atrial; RV, right ventricular; RVEDP, RV and diastolic pressure; Sys, systolic; Tm, transmural. *P ⬍ .05; †P ⬍ .01; ‡P ⬍ .001 vs IPP paced I or II; §P ⬍ .05; 㛳P ⬍ .01; ¶P ⬍ .001 vs IPP decreased volume.

Figure 1 Deceleration time (DCT) is shown for each dog with pacing, after decreased intracardiac volume (ICV), and with volume loading with intact pericardium (IP Incr Vol). Only 1 paced baseline left ventricular dysfunction is shown, as data for both paced baselines were nearly identical. Of 12 dogs, 10 demonstrated prolongation of DCT with inferior vena cava occlusion and shortening with volume loading. IP, Pacing with intact pericardium; IP Dec Vol, inferior vena cava occlusion reducing ICV with intact pericardium; IP Incr Vol.

moderately dilated LV12,18 that would not decompensate with the 2 thoracotomy procedures. Moreover, we selected this level of LV dysfunction and

dilatation on the basis of our previous experience with this model in which volume loading or nitroglycerin infusion resulted in marked changes in LV volumes and pressure. The influence of the pericardium is evident from Table 4. Pericardiectomy demonstrated increased LV volume and ICV with a reduction in FF1/3D. The diastolic influence of the pericardium was more evident with volume loading as evidenced by the lower FF1/3D and E/A ratio after pericardiectomy as compared with the intact pericardium. With an intact pericardium, an increase in ICV was associated with an even shorter DCT, greater early diastolic filling (FF1/3D), and a higher intrapericardial contact pressure strongly supporting the role of pericardial constraint. Pericardial constraint was further demonstrated by alterations in DCT with alterations in ICV produced in 10 of 12 dogs. DCT, a noninvasive indicator of LV chamber stiffness,19,20 was predicted by IPP. However, LV transmural chamber stiffness constant was also an important determinant suggesting an element of myocardial restraint as evidenced by DCT shortening after pericardiectomy in 6 of 12 dogs. However, despite DCT shortening with volume loading after pericardiectomy, the loss of pericardial constraint was evident, as FF1/3D was markedly lower early in diastole after pericardiectomy permitting redistribution of diastolic filling to later in diastole. Previous Literature The RFP resembles the inflow pattern associated with constrictive pericarditis whether characterized

Journal of the American Society of Echocardiography Volume 17 Number 2

Lavine 157

Table 3 Measures of hemodynamics cardiac volume, and function, and diastolic parameters after pericardiectomy with intracardiac volume changes EDV ESV EF ICV LVEDP LV Sys pressure Tau Mean art pressure RA pressure RVEDP RV Sys pressure RR E/A Decel time FF1/3D IRP Tm LV stiffness

mL mL Percentage mL mm Hg mm Hg msec mm Hg msec mm Hg mm Hg msec msec Percentage msec mm Hg/mL

PECT paced I

PECT decreased volume

PECT paced II

PECT increased volume

75 ⫾ 28 43 ⫾ 23 43 ⫾ 12 178 ⫾ 39 9.2 ⫾ 3.1 97 ⫾ 9 41 ⫾ 9 87 ⫾ 8 5.3 ⫾ 3.7 5.5 ⫾ 3.5 24 ⫾ 9 618 ⫾ 47 1.9 ⫾ 0.5 141 ⫾ 26 19.9 ⫾ 14.4 61 ⫾ 25 0.19 ⫾ 0.07

62 ⫾ 23* 35 ⫾ 15 44 ⫾ 11 158 ⫾ 31* 5.0 ⫾ 3.1† 85 ⫾ 9† 44 ⫾ 16 72 ⫾ 12† 3.0 ⫾ 2.4* 3.3 ⫾ 3.7* 22 ⫾ 7* 618 ⫾ 47 1.9 ⫾ 0.5 162 ⫾ 57 21.2 ⫾ 9.2 39 ⫾ 37* 0.09 ⫾ 0.06†

73 ⫾ 31 40 ⫾ 21 44 ⫾ 10 175 ⫾ 35 9.4 ⫾ 3.5 99 ⫾ 11 43 ⫾ 7 86 ⫾ 10 5.1 ⫾ 3.5 5.2 ⫾ 3.3 24 ⫾ 10 618 ⫾ 47 1.9 ⫾ 0.4 142 ⫾ 24 20.2 ⫾ 12.5 63 ⫾ 21 0.20 ⫾ 0.09

91 ⫾ 20*§ 50 ⫾ 18* 45 ⫾ 11 214 ⫾ 32*§ 14.5 ⫾ 5.2‡¶ 113 ⫾ 14‡¶ 55 ⫾ 12‡㛳 102 ⫾ 13‡¶ 8.7 ⫾ 3.9†㛳 9.1 ⫾ 4.0†㛳 30 ⫾ 5*㛳 618 ⫾ 47 1.7 ⫾ 0.6 112 ⫾ 38㛳 15.5 ⫾ 11.8§ 86 ⫾ 40*¶ 0.29 ⫾ 0.12†㛳

A, Peak atrial filling velocity; art, arterial; Decel, deceleration; E, transmitral peak rapid filling velocity; EF, ejection fraction; EDV, end-diastolic volume; ESV, end-systolic volume; FF1/3D, filling fraction at one third of diastole; ICV, intracardiac volume; IRP, isovolumic relaxation period; LV, left ventricular; LVEDP, LV end-diastolic pressure; PECT, pericardiectomy; RA, right atrial; RV, right ventricular; RVEDP, RV end-diastolic pressure; Sys, systolic; Tm, transmural. *P ⬍ .05; †P ⬍ .01; ‡P ⬍ .001 vs PECT paced I or II; §P ⬍ .05; 㛳P ⬍ .01; ¶P ⬍ .001 vs PECT decreased volume.

Figure 2 Deceleration time (DCT) is shown for each dog with pacing, after decreased intracardiac volume (ICV), and with volume loading after pericardiectomy (Pect Incr Vol). Of 12 dogs, 6 demonstrated prolongation of DCT with inferior vena cava occlusion and shortening with volume loading. Only 1 paced baseline left ventricular (LV) dysfunction is shown, as data for both paced baselines were nearly identical. Pect, Paced LV dysfunction with pericardiectomy; Pect Dec Vol, inferior vena cava occlusion reducing ICV after pericardiectomy.

by transmitral pulsed Doppler or by volume curve assessment using contrast ventriculography.20,21 As DCT is directly related to LV chamber compliance (inversely related to LV chamber stiffness) and the

net left atrial-ventricular compliance,19,22 it follows that restrictive filling may be related to restraining forces in diastole produced either by a constraining pericardium or a restraining myocardium. Further support for this hypothesis can be derived from the increase in left atrial conduit function, reduction in left atrial stiffness constant, and increase in peak mitral flow velocity after pericardiectomy in open chest dogs.23 The RFP may be the result of being on the ascending limb of the LV pressure-volume curve where small volume changes result in marked diastolic pressure changes.24 Previous data have hinted at the importance of diastolic ventricular interaction mediated in part by external constraint as characterized by changes in LV end-diastolic volume with changes in lower body negative pressure25 and with RV infarction.26 Experimentally, the role of pericardial constraint has also been demonstrated with increases in LV volumes and reduction in LV filling pressures after pericardiectomy in normal and failing ventricles.27,28 The use of contact pericardial pressure as IPP has provided insight into how much of intracardiac right-sided filling pressures is attributable to pericardial contact pressure that begins to increase disproportionately when LVEDP and RVEDP exceed 9 and 4 mm Hg, respectively.29 Pericardial constraint influences LV diastolic filling pattern in that with pericardiectomy, E/A increases,28,30 and the E-left atrial pressure relation is altered.31 However, there have been no mechanistic studies addressing the relationship of pericardial restraint with LV dysfunction to transmitral filling indices. The results of this study provide an initial

Journal of the American Society of Echocardiography February 2004

158 Lavine

Table 4 Measures of cardiac volume, function, hemodynamics, and diastolic function before and after pericardiectomy EDV ESV EF ICV LVEDP LV Sys pressure Tau Mean art pressure RR E/A Decel time FF1/3D IRP Tm LV stiffness

mL mL Percentage mL mm Hg mm Hg msec mm Hg msec msec Percentage msec mm Hg/mL

IP paced

PECT paced

IP increased volume

PECT increased volume

63 ⫾ 15 37 ⫾ 16 42 ⫾ 14 148 ⫾ 24 10.2 ⫾ 2.9 100 ⫾ 9 44 ⫾ 14 91 ⫾ 8 618 ⫾ 47 1.8 ⫾ 0.4 130 ⫾ 35 45.6 ⫾ 29.3 53 ⫾ 24 0.32 ⫾ 0.12

75 ⫾ 28* 43 ⫾ 23 43 ⫾ 12 178 ⫾ 39* 9.2 ⫾ 3.1 97 ⫾ 9 41 ⫾ 9 87 ⫾ 8 618 ⫾ 47 1.9 ⫾ 0.5 141 ⫾ 26 19.9 ⫾ 14.4† 61 ⫾ 25 0.19 ⫾ 0.07†

81 ⫾ 26 47 ⫾ 24 42 ⫾ 16 192 ⫾ 32 14.9 ⫾ 3.2 104 ⫾ 13 52 ⫾ 12 94 ⫾ 15 618 ⫾ 47 2.2 ⫾ 1.0 107 ⫾ 45 60.1 ⫾ 14.8 69 ⫾ 27 0.41 ⫾ 0.19

91 ⫾ 20 50 ⫾ 18 45 ⫾ 11 214 ⫾ 32‡ 14.5 ⫾ 5.2 113 ⫾ 14‡ 55 ⫾ 12 102 ⫾ 13‡ 618 ⫾ 47 1.7 ⫾ 0.6‡ 112 ⫾ 38 15.5 ⫾ 11.8§ 86 ⫾ 40† 0.29 ⫾ 0.12

A, Peak atrial filling velocity; art, arterial; Decel, deceleration; E, transmitral peak rapid filling velocity; EF, ejection fraction; EDV, end-diastolic volume; ESV, end-systolic volume; FF1/3D, filling fraction at one third of diastole; ICV, intrapericardial volume; IP, intact pericardium; IRP, isovolumic relaxation period; LV, left ventricular; LVEDP, LV end-diastolic pressure; PECT ⫽ pericardiectomy; Sys, systolic; Tm, transmural. *P ⬍ .05; †P ⬍ .01, vs IP paced; ‡P ⬍ .05; §P ⬍ .001 vs IP increased volume.

Figure 3 Individual filling fractions at one-third of diastole (FF 1/3 D) are plotted with pacing and volume loading with intact pericardium (IPP Vol) and after pericardiectomy (PECT Volume). IPP, Intact pericardium paced baseline; PECT, paced baseline after pericardiectomy.

step toward the understanding of the genesis of the RFP. Limitations One important limitation of this study was that this was an acute study in a chronic model of LV dysfunction. Measurements were made after surgical procedures for IPP measurement and after pericardiectomy. Both the order of the stages and sequence of ICV alterations were fixed (pericardiectomy), which may introduce bias. As IPP is a contact

pressure, there is always the concern that the fluidfilled flat balloon may increase the IPP. Transmitral Doppler rapid and atrial filling fractions did not demonstrate alterations with changes in ICV (data not shown). Alternatively, FF1/3D, which includes the isovolumic relaxation period, did demonstrate significant differences. Its calculation was on the basis of short-axis area on a manual frame-by-frame measurement that is subject to errors in measurement. This measurement is on the basis of a strict time-based measurement and differs from rapid or atrial filling fractions where the lengths and timing of these filling periods may vary with changes in ICV. Although frame-by-frame measurement of LV short-axis area does not have the same time resolution that Doppler tracings have, each frame represents 33 milliseconds with the number of diastolic frames averaging approximately 12. Despite the time resolution issue, the difference in FF1/3D after pericardiectomy was substantial, making the time resolution differences with Doppler less important. Full pressure-area loops were not constructed because short-axis LV area determinations had a time resolution of 33 milliseconds. Significant scatter may be seen because manual outlining of borders (as a result of intraobserver and interobserver variability) renders parallel movement of curves more difficult to discern. Pressure-volume loops are easier to construct using volume conductance catheters rather than echocardiography because of greater time resolution and less variability. Other limitations included the lack of left atrial pressure measurements and pulmonary venous inflow recordings. Doppler tissue recordings of the mitral annulus might have been instructive but were not available on the echocardiograph. The acute effects of pericardiectomy might have been avoided by performing pericardiectomy or

Journal of the American Society of Echocardiography Volume 17 Number 2

Lavine 159

Figure 4 Expanded composite diastolic left ventricular (LV) pressure short-axis area plots are shown with paced LV dysfunction with intact pericardium (IP) and after pericardiectomy (PECT) on left at end-systole (point not shown), LV pressure minimum, A wave, and LV end-diastolic pressure. Error bars are shown for both LV pressure and endocardial short-axis areas for each point (SEM). Only 1 paced baseline LV dysfunction is shown, as data were identical for both. On right, LV pressure short-axis area plots are shown with volume loading with intact pericardium (IP Incr Vol) and after pericardiectomy with volume loading (PECT Incr Vol). Plots with intact pericardium are displaced parallel and upward.

sham pericardiectomy before coronary microsphere embolization. Unfortunately, our own experience with pericardiectomy before coronary microsphere embolization resulted in a dilated spherical LV with significant mitral regurgitation further reinforcing the role of the pericardium in preventing overdistension and maintaining LV shape (unpublished observations). Total ICV was calculated from the addition of cardiac chamber volumes and LV wall volume, but did not include other chamber wall volumes that were thought to be negligible. This measurement has not had prior validation. It serves as a crude index of volume to substantiate an increase in ICV. Changes in total ICV after volume loading, pericardiectomy, or both may be underestimated or overestimated on the basis of changing geometry of cardiac chambers. Clinical Implications Although it is difficult to separate out the role of both myocardial and pericardial influences on the genesis of the RFP, both chamber dilatation and myocardial interstitial and replacement fibrosis may be potential targets for intervention. As both myocardial restraint and pericardial constraint are associated with chamber dilatation, efforts to reduce chamber size with therapy designed to reduce ICV (angiotensin-converting enzyme inhibitors) may produce clinical improvement. For example, reversible RFP is associated with a better prognosis than irreversible RFP.32 Myocardial restraint may in part be related to marked interstitial and replacement fibrosis associated with a dilated LV cavity. Similar

therapies used may be beneficial as angiotensinconverting enzyme inhibition and aldosterone antagonism reduces fibrosis.33,34 We conclude that the genesis of the RFP as characterized by a shortened DCT may be in part related to pericardial constraint. Myocardial restraint is also a prominent feature in this canine model of LV dysfunction characterized by LV dilatation, dysfunction, and myocardial interstitial and replacement fibrosis. The coexistence of both factors is likely to occur for patients with advanced heart disease with LV dysfunction. Removal of the pericardial external constraint results not only in lengthening of the DCT, but redistribution of diastolic filling to later in diastole. We would like to thank Petar Prcevski, DVM, and Vicki Johnson for their invaluable assistance in both the technical performance and the echocardiographic analysis of this investigation. REFERENCES 1. Xie GY, Berk MB, Smith MD, Gurley JC, DeMaria AN. Prognostic value of Doppler transmitral flow patterns in patients with congestive heart failure. J Am Coll Cardiol 1994; 24:132-9. 2. Faris R, Coats AJ, Enein MY. Echocardiography-derived variables predict outcome in patients with nonischemic dilated cardiomyopathy with and without restrictive filling pattern. Am Heart J 2002;144:343-50. 3. Nijland F, Kamp O, Karreman AJ, van Eenige MJ, Visser CA. Prognostic implications of restrictive left ventricular filling in acute myocardial infarction: a serial Doppler echocardiographic study. J Am Coll Cardiol 1997;30:1618-24.

160 Lavine

4. Klein AL, Hatle LK, Taliercio CP, Oh JK, Kyle RA, Gertz MA, et al. Prognostic significance of Doppler measures of diastolic filling in cardiac amyloidosis: a Doppler echocardiographic study. Circulation 1991;83:808-16. 5. Appleton CP, Hatle LK, Popp RL. Demonstration of restrictive ventricular physiology by Doppler echocardiography. J Am Coll Cardiol 1988;11:57-68. 6. Eusebio J, Louie EK, Edwards LC III, Loeb HS, Scanlon PJ. Alterations in transmitral flow dynamics in patients with early mitral valve closure and aortic regurgitation. Am Heart J 1994;128:941-7. 7. Louie EK, Rich S, Levitsky S, Brundage BH. Doppler echocardiographic demonstration of differential effect of right ventricular pressure and volume overload on left ventricular geometry and filling. J Am Coll Cardiol 1992;19:84-90. 8. Smith ER, Smiseth OA, Kingma I, Manyari D, Belenkie I, Tyberg JV. Mechanism of action of nitrates: role of changes in venous capacitance and in left ventricular diastolic pressurevolume relation. Am J Med 1984;76:14-21. 9. Glantz SA, Misbach GA, Moores WY, Mathey DG, Lekven J, Stowe DF, et al. The pericardium substantially effects the left ventricular diastolic pressure-volume relation in the dog. Circ Res 1978;42:433-41. 10. Lavine SJ, Campbell CA, Kloner RA, Gunther SJ. Diastolic filling in acute left ventricular dysfunction: role of pericardium. J Am Coll Cardiol 1988;12:1326-33. 11. Lavine SJ, Campbell CA, Held AC, Johnson V. Effect of nitroglycerin induced reduction of left ventricular filling pressure on diastolic filling in acute dilated heart failure. J Am Coll Cardiol 1989;14:233-41. 12. Lavine SJ, Prcevski P, Held AC, Johnson V. Experimental model of chronic global left ventricular dysfunction secondary to left coronary microembolization. J Am Coll Cardiol 1991; 18:1794-803. 13. Smiseth OA, Frais MA, Kingma I, Smith ER, Tyberg JV. Assessment of pericardial constraint in dogs. Circulation 1985;71:158-64. 14. Hoit BD, Lew WY, Lewinter M. Regional variations in pericardial pressure in the canine ventricle. Am J Physiol 1988; 255:H1370-7. 15. Weiss JL, Fredrickson JW, Weisfeldt ML. Hemodynamic determinants of the time course of fall in canine left ventricular pressure. J Clin Invest 1976;58:751-60. 16. Cohen GI, Pietrolungo JF, Thomas JD, Klein AL. A practical assessment of ventricular diastolic function using Doppler echocardiography. J Am Coll Cardiol 1996;27:1753-60. 17. Marino P, Little WC, Rossi A, Barbieri E, Anselmi M, Destro G, et al. Can left ventricular diastolic stiffness be measured noninvasively? J Am Soc Echocardiogr 2002;15:935-43. 18. Sabbah HN, Stein PD, Kono T, Gheorghiade M, Levin TB, Jafri S, et al. A canine model of chronic heart failure produced by multiple sequential coronary microembolizations. Am J Physiol 1991;260:H1379-84.

Journal of the American Society of Echocardiography February 2004

19. Ohno M, Cheng CP, Little WC. Mechanism of altered patterns of left ventricular filling during the development of congestive heart failure. Circulation 1994;89:2241-50. 20. Hatle LK, Appleton CP, Popp RL. Differentiation of constrictive pericarditis and restrictive cardiomyopathy by Doppler echocardiography. Circulation 1989;79:357-70. 21. Tyberg TL, Goodyer AVN, Hurst VW III, Alexander J, Langou RA. Left ventricular filling in differentiating restrictive amyloid cardiomyopathy and constrictive pericarditis. Am J Cardiol 1981;47:791-6. 22. Thomas JD, Newell JB, Choong CY, Weyman AE. Physical and physiological determinants of transmitral velocity: numerical analysis. Am J Physiol 1991;260:H1718-31. 23. Hoit BD, Shao Y, Gabel M, Walsh RA. Influence of the pericardium on left atrial compliance and pulmonary venous flow. Am J Physiol 1993;264:H1781-7. 24. Solomon SB, Nikolic SD, Glantz SA, Yellin EL. Left ventricular diastolic function of remodeled myocardium in dogs with pacing induced heart failure. Am J Physiol 1998;274:H945-54. 25. Atherton JJ, Moore TD, Thomson HL, Frenneaux MP. Restrictive left ventricular filling patterns are predictive of diastolic ventricular interaction. J Am Coll Cardiol 1998;31:413-8. 26. Calvin SE. Optimal right ventricular filling pressures and the role of pericardial constraint in right ventricular infarction in dogs. Circulation 1991;84:852-61. 27. Horne SG, Belenkie I, Tyberg JV, Smith ER. Pericardial pressure in experimental heart failure. Can J Cardiol 2000;16: 607-13. 28. Yamamoto K, Masuyama T, Tanouchi J, Uematsu M, Doi Y, Naito J, et al. Decreased and abnormal left ventricular filling in acute heart failure: role of pericardial constraint and its mechanism. J Am Soc Echocardiogr 1992;5:504-14. 29. Applegate RJ, Johnston WE, Vinten-Johansen J, Klopfenstein HS, Little WS. Restraining effects of intact pericardium during acute volume loading. Am J Physiol 1992;262:H1725-33. 30. Hoit BD, Dalton N, Bhargava V, Shabetai R. Pericardial influences on right and left ventricular filling dynamics. Circ Res 1991;68:197-208. 31. Nakatani S, Beppu S, Miyatake K, Nimura Y. Effect of pericardium on left ventricular early filling assessed by pulsed Doppler echocardiography. J Am Soc Echocardiogr 1991;4: 29-34. 32. Pinamonti B, Zecchin M, Di Lenarda A, Gregori D, Sinagra G, Camerini F. Persistence of restrictive left ventricular filling pattern in dilated cardiomyopathy: an ominous prognostic sign. J Am Coll Cardiol 1997;29:604-12. 33. Goldstein S, Sharov VG, Cook JM, Sabbah HN. Ventricular remodeling: insights from pharmacologic intervention with angiotensin converting enzyme inhibitors. Mol Cell Biochem 1995;147:51-5. 34. Tsutamoto T, Wada A, Maeda K, Mabushi N, Hayashi M, Tsutsui T, et al. Effect of spironolactone on brain natriuretic peptide and left ventricular remodeling in patients with congestive heart failure. J Am Coll Cardiol 2001;37:1228-33.