Author + information
- Received December 20, 2004
- Revision received March 31, 2005
- Accepted April 5, 2005
- Published online August 2, 2005.
- Alison M. Duncan, MRCP, PhD⁎,⁎,† (, )
- Eric Lim, MSc, MRCS⁎,†,
- Derek G. Gibson, FRCP⁎,† and
- Michael Y. Henein, MD, PhD, FACC⁎,†
- ↵⁎Reprint requests and correspondence:
Dr. Alison Duncan, Echocardiography Department, The Royal Brompton Hospital, Sydney Street, London, SW3 6NP, United Kingdom.
Objectives The purpose of this research was to study the effect of dobutamine on left ventricular (LV) filling in ischemic cardiomyopathy (ICM) and to determine whether restrictive filling pattern (RFP) at peak stress has prognostic value.
Background The prognostic value of RFP at peak stress in ICM is unknown.
Methods A total of 69 patients with ICM were studied by Doppler echocardiography at rest and stress; RFP was defined as transmitral E:A ratio ≥1.0, isovolumic relaxation time (IVRT) <80 ms, and E-wave deceleration time (EDT) <120 ms.
Results A total of 42 of 69 had RFP at rest, which reverted to non-RFP at stress in 24 (EA), but persisted in 18 (EE); 27 of 69 had non-RFP at rest and peak stress (AA). In EA, IVRT and EDT lengthened (by 43 ms and 46 ms), and tricuspid regurgitation (TR) decreased (by 26 mm Hg, p < 0.01), suggesting a fall in left atrial (LA) pressure. The stress response in AA was similar to EA. In EE, IVRT and EDT shortened (by 21 ms) and TR increased (by 13 mm Hg, p < 0.01), suggesting a rise in LA pressure. Peak aortic acceleration (LV inotropy) increased by 0.8 g in EA but only by 0.2 g in EE (difference p < 0.001). Median follow-up (interquartile range) was 34 (20 to 57) months. Three-year survival for EE, EA, and AA was 49%, 79%, and 89%, respectively (p < 0.001). Compared with AA, the hazard ratio for EE was 9.5 (p < 0.001) and for EA was 1.9 (p = 0.30).
Conclusions In ischemic cardiomyopathy, persistence of restrictive filling during stress implies a striking rise in LA pressure, greatly attenuated LV inotropic response, and markedly reduced survival. Stress echocardiography uniquely identifies these high-risk patients.
Patients with dilated cardiomyopathy secondary to coronary artery disease (CAD) (ischemic cardiomyopathy [ICM]) have a wide spectrum of left ventricular (LV) filling patterns (1–3). A “restrictive” LV filling pattern, characterized on Doppler echocardiography by a shortened transmitral E-wave deceleration time (EDT) (2), is frequently found in patients with dilated cardiomyopathy (4). It is associated with severe impairment of LV hemodynamics (5,6) and worse clinical functional class (7,8), and is a powerful indicator of increased mortality and transplantation rate (9,10). However, the majority of studies of restrictive filling in patients with ICM have been performed at rest. Those that have investigated changes in LV filling pattern during pharmacological stress in ICM have either not addressed the prognostic value of changes in LV filling pattern (11), or have studied patients with idiopathic cardiomyopathy (12). The primary objective of this study was to assess whether dobutamine stress alters LV filling pattern in patients with ICM. Secondary aims were to determine whether possible underlying mechanisms for change in LV filling pattern with stress could be elucidated, and whether LV filling pattern at peak stress has any prognostic value.
We studied 69 patients with heart failure, age 62 ± 8 years (62 men), referred for assessment after symptomatic deterioration. Entry criteria were: 1) the presence of a uniformly dilated LV, with an LV end-diastolic dimension >56 mm, and an end-systolic dimension >40 mm; 2) significant CAD, demonstrated by at least two-vessel disease (>70% stenosis) at coronary angiography; and 3) New York Heart Association functional class III to IV. All three entry criteria were fulfilled in each patient. Predetermined exclusion criteria were atrial fibrillation, structural valve disease, and severe mitral regurgitation (MR). All patients underwent dobutamine stress echocardiography at entry into the study. Treatment was recorded at entry, and was closely monitored and optimized as clinically indicated during regular follow-up visits in a nurse-run heart failure clinic throughout the study. Information on survival was collected by telephone contact with patients or their relatives and physicians. Primary end points for follow-up were all-cause mortality or the end of the study. The Royal Brompton and Harefield Ethics Committee approved the study protocol. All subjects gave written informed consent, and there were no complications related to the investigation.
Dobutamine stress echocardiography
Dobutamine stress echocardiography was performed using a Philips Sonos 5500 echocardiograph and a multifrequency transducer (Andover, Massachusetts). A 12-lead electrocardiogram and arterial blood pressure were recorded at each stage of stress (13). Predetermined stress end points were 85% predicted target heart rate (220 minus age in years), or the development of symptoms, ventricular ectopics, 20 mm Hg drop in systolic arterial pressure, ST-segment shift >2 mm, or T-wave inversion.
Two-dimensional echocardiography was performed from the parasternal long- and short-axis views and apical four- and two-chamber views. Cross-sectional two-dimensional-guided M mode recordings were performed using the left parasternal long-axis view with the cursor at the tips of the mitral valve leaflets. Left ventricular minor axis dimensions were measured at end diastole (the onset of the QRS complex) and at end systole (the first high-frequency vibration of the aortic component of the second heart sound on the phonocardiogram, A2, confirmed as synchronous with the onset of the closure artifact on the aortic Doppler record). Transmitral flow velocities were recorded from the apical four-chamber view using pulsed Doppler, with the sample volume between the tips of the mitral valve leaflets and 5 cm/s baseline filter. Doppler isovolumic relaxation time ([IVRT] A2 to the onset of mitral flow), peak E (early diastolic), and A (atrial) velocities, EDT, and the ratio of early transmitral flow velocity to atrial flow velocity (E:A ratio) were measured in all patients. A-wave amplitude was measured from the baseline, rather than from the declining E-wave. Mitral regurgitation was graded as mild, moderate, or severe according to the distance from the valve orifice that the regurgitant jet remained detectable on the color-flow Doppler recording. Transaortic flow velocity was obtained by pulsed-wave Doppler from the apical five-chamber view. Stroke distance was calculated as the time integral of aortic velocity, and stroke volume (SV) as the product of stroke distance and subaortic area. Mitral and aortic velocity traces were digitized off-line (100 Hz), and peak mitral E-wave acceleration rate (PMEA) and peak aortic acceleration rate (PAA), both expressed in g (1 g = 9.81 m/s2), were derived by differentiating the velocity traces with respect to time (14). Tricuspid regurgitation (TR) was assessed by color-flow and continuous-wave Doppler from the apical four-chamber view.
Wall motion score index was analyzed according to American Society of Echocardiography criteria (15); LV long-axis M-mode recordings were obtained with the cursor positioned at the lateral, septal, and posterior angles of the mitral ring. The average measurements of the three sites are presented. Left ventricular total long-axis amplitude was defined as maximum displacement of the ring between the onset of QRS and peak inward movement at or after A2. Post-ejection shortening (PES) was measured as the amplitude of shortening after A2, and systolic amplitude (SA) as total amplitude less PES. Early diastolic lengthening velocities were measured from the digitized long-axis traces. Right ventricular long-axis M-mode recordings were obtained with the cursor positioned at the right angle of the tricuspid ring, and right ventricular systolic long-axis amplitude measured as the amplitude of tricuspid ring motion occurring between the end of the QRS and the peak innermost point at or after the pulmonary valve closure (P2). All recordings were acquired at rest and peak stress at a paper speed of 100 mm/s, with an ECG (lead II) and a phonocardiogram superimposed.
Classification of patients
Restrictive filling was defined at entry by means of a short IVRT (<80 ms), increased E:A ratio (≥1.0), and short EDT (≤120 ms, the lower 95% confidence interval [CI]) (2,5,6). All three criteria were required to be present. A nonrestrictive filling pattern was defined as IVRT >80 ms, an E:A ratio <1.0, and an EDT >120 ms, and all three criteria were required to be present. Restrictive filling at peak stress was defined as an E:A ratio ≥1.0, whereas nonrestrictive filling at peak stress was defined as an E:A ratio <1.0. Patients were classified into three groups according to the Doppler LV filling pattern at rest and at peak stress: 1) group AA had nonrestrictive filling at rest that remained nonrestrictive at peak stress; 2) group EA had restrictive LV filling at rest that became nonrestrictive at peak stress; and 3) group EE had restrictive LV filling at rest that persisted at peak stress.
Statistical analyses were performed using Statview 4.5 (Abacus Concepts, Berkeley, California) and S Plus 6.2 (Insightful, Seattle, Washington). Normally distributed continuous variables were expressed as mean ± SD, whereas categorical variables were expressed as number (percentage). The between group (AA, EA, and EE) differences in baseline characteristics were tested using the unpaired ttest (continuous variables) or the chi-square or Fisher exact test as appropriate (when expected cell frequencies were <10 for categorical variables). Within each group, the change with stress was compared using the paired ttest. Analysis of covariance was applied to quantify and test the significance of baseline-adjusted differences between rest and stress values between patient groups. Kaplan-Meier analysis was performed to evaluate survival in the three groups, and difference was tested using the log-rank test. Prognostic variables were entered into a Cox proportional hazards regression model to identify independent predictors of mortality. Backward stepwise selection was used with the criterion for variable retention as p < 0.10, and a significance level of p < 0.05 was used in the final multivariate model.
Two investigators, both unaware of the original measurements, analyzed duplicate echocardiographic measurements of EDT, right ventricular long-axis SA, and PAA in 20 randomly selected patients from the same original records. Reproducibility was expressed as the root mean square difference between duplicate values. Intraobserver variability for EDT, right ventricular amplitude, and PAA was 9 ms, 0.6 mm, and 0.1 g, respectively, and interobserver variability was 13 ms, 0.7 mm, and 0.1 g, respectively.
Of 69 patients studied, resting LV filling pattern was restrictive in 42, and non-restrictive in 27. Distinct E- and A-wave velocities could be distinguished at peak stress in all individuals. All 27 patients with non-restrictive filling at rest remained non-restrictive at peak stress (group AA). In 24 of 42 patients with restrictive filling at rest, the E:A ratio fell to <1.0 with stress, and these patients were classified as group EA (Fig. 1);E:A ratio remained ≥1.0 in the remaining 18 of 42 patients, classified as group EE (Fig. 2).Isovolumic relaxation time and EDT lengthened with stress in all 24 patients in group EA, and shortened with stress in all 18 patients in group EE (Fig. 3,top).
At rest, there was no significant difference between patients in group AA and group EA in terms of age, mean arterial blood pressure, hemoglobin concentration, renal function, serum cholesterol, or medical therapy (Table 1).Furthermore, there was no difference in heart rate, SV, LV cavity size, long-axis SA, or wall motion score index between the two groups at rest, though PAA was 0.5 g lower in group EA (p < 0.001) (Table 2).By design, resting IVRT and EDT were shorter in group EA compared to those in group AA (in the event by 67 ms and 52 ms, respectively), and peak E-wave velocity and E:A ratio were greater (by 0.3 m/s and 1.4) (Table 2). In addition, PMEA was greater (p < 0.01) in group EA compared to group AA, right ventricular long-axis SA was lower (p < 0.001), and TR pressure drop was greater (p < 0.001). The majority of echocardiographic variables in group EE were not significantly different from those in group EA, other than E:A ratio, right ventricular SA, and TR pressure drop (all p < 0.01).
Effect of dobutamine stress
Group AA: Non-restrictive filling at rest that remained non-restrictive at peak stress
Peak aortic acceleration rate (p < 0.001) and SV (p < 0.01) increased significantly with stress, and LV cavity size decreased (p < 0.001) (Table 3);IVRT (p < 0.05), EDT, and peak A-wave velocity (p < 0.001) all increased, whereas PMEA decreased (p < 0.001). Left ventricular long-axis SA did not change, but PES increased (p < 0.001). Right ventricular long-axis amplitude increased (p < 0.01), and TR pressure drop decreased (p < 0.001).
Group EA: Restrictive filling at rest that became non-restrictive at peak stress
Peak dobutamine infusion rate was not different in group EA (27 ± 7 μg/kg/min) compared to group AA (30 ± 7 μg/kg/min, p = 0.12). Moreover, the stress response was similar (Table 3). Peak aortic acceleration rate and SV both increased (p < 0.01), end-diastolic dimension was unchanged, and end-systolic dimension fell (p < 0.05). Isovolumic relaxation time, EDT, and peak mitral A-wave all increased with stress (Fig. 2), and PMEA decreased (all p < 0.001). Left ventricular long-axis SA did not change, though PES increased (p < 0.001). Right ventricular long-axis amplitude increased, and TR pressure drop decreased (both p < 0.001; Fig. 1, right panel).
Group EE: Restrictive filling at rest that remained restrictive at peak stress
There was no difference in peak dobutamine infusion rate between group EE (31 ± 9 μg/kg/min) and group EA (27 ± 7 μg/kg/min, p = 0.15). Left ventricular cavity size fell with stress (p < 0.05, Table 3), but the rest of the stress response differed significantly from that in group EA. In group EE, PAA increased (p < 0.001), but SV did not change; IVRT and EDT both shortened (p < 0.001), peak E- and A-wave velocities and E:A ratio did not change, and PMEA increased (p < 0.001) (Fig. 3). Neither LV long-axis SA nor PES changed with stress, right ventricular long-axis amplitude failed to increase, and there was a significant increase in TR pressure drop (p < 0.01) (Fig. 2, right panel).
Differences in stress response
There was no difference in the baseline-adjusted stress response between group EA and group AA in terms of PAA (stress-induced change [95% CI]) (−0.3 g [95% CI = 0.0 to 0.6], p = 0.07), SV (+2 ml [95% CI −3 to +4], p = 0.32), IVRT (+8 ms [95% CI −12 to +28], p = 0.43), EDT (−6 ms [95% CI −24 to −11], p = 0.47), mitral A-wave velocity (+0.2 m/s [95% CI 0.0 to 0.3], p = 0.05), PMEA (+0.1 g [95% CI −0.1 to +0.3], p = 0.27), right ventricular long-axis amplitude (+1.0 mm [95% CI −0.8 to +2.7], p = 0.28), or TR pressure drop (−1 mm Hg [95% CI −8 to +6], p = 0.70).
However, there were significant differences in the baseline-adjusted stress response between groups EE and EA. In group EE, PAA was 0.6 g (95% CI 0.3 to 0.9) lower (Fig. 3, top), IVRT and EDT were 64 ms (95% CI 52 to 76) and 61 ms (95% CI 48 to 74) shorter, peak E-wave velocity, E:A ratio, PMEA, and TR pressure drop were all higher (Fig. 3, bottom) (by 0.5 m/s [95% CI 0.4 to 0.6], 2.5 [95% CI = 1.7 to 3.4], 0.7 g [95% CI 0.5 to 0.9], and 44 mm Hg [95% CI 37 to 51]), and right ventricular long-axis amplitude was 3.2 mm (95% CI 1.1 to 5.3) lower (all p < 0.001 compared with group EA). The difference in the increment in SV was not significant (5 ml [95% CI 2 to 7], p = 0.06).
Effect of stress on MR
All 69 patients had functional MR at rest, but its severity did not increase in any patient during stress. In group AA, moderate MR became mild in 10 of 27 patients, mild MR regressed completely in 15 of 27, and MR severity did not change in 2 of 27. In group EA, moderate MR became mild in 8 of 24 patients, mild MR regressed completely in 12 of 24, and did not change in 4 of 24. In group EE, moderate MR became mild in 4 of 18 patients, mild MR regressed completely in 10 of 18, and MR severity did not change in 4 of 18. These differences were not significantly different.
Effect of stress on wall motion score index
The overall follow-up was a median of 34 months (interquartile range 20 to 57). During the follow-up period, 13 of 18 patients in group EE died, compared with 7 of 24 patients in group EA, and 4 of 27 patients in group AA (Fig. 4).Three-year survival for patients by group (EE, EA, AA) was 49%, 79%, and 89%, respectively (log-rank p < 0.001). Univariate predictors of mortality are presented in Table 4.Measures of systolic function at rest demonstrated a trend toward univariate predictors of mortality (PAA: p = 0.05, fractional shortening: p = 0.07), and although stress-induced changes in PMEA and TR pressure drop were univariate predictors of mortality, both were subsumed by group allocation in Cox proportional hazard analysis. There were no independent predictors on multivariate analyses for measures at rest apart from group variable. Compared with group AA, the overall hazard ratio for group EE was 9.5 (95% CI 4.3 to 14.7, p < 0.001) and for group EA was 1.9 (95% CI −4.2 to 8.1, p = 0.30).
Our primary aim was to investigate the effect of dobutamine stress on LV filling pattern in a heterogeneous group of patients with ICM. Our secondary aims were to determine physiological mechanisms that might underlie differences in LV filling pattern at peak stress, and to determine whether stress-induced LV filling pattern had prognostic significance.
Stress-induced physiological findings
At rest, there were only minor differences between patients with restrictive filling who subsequently remained restrictive at peak stress and those with restrictive filling who became non-restrictive with stress. However, the difference in the stress response between the two groups was striking (Fig. 3). In patients who remained restrictive at peak stress, both IVRT and EDT shortened significantly with stress, suggesting an increase in left atrial pressure. This was confirmed by a significant increase in peak mitral E-wave acceleration, representing a higher early diastolic atrioventricular pressure gradient with stress, and a simultaneous rise in right ventricular systolic pressure. The unlikely possibility of diastolic suction was excluded by the correspondingly low values of E/Ea. In contrast, the stress response in patients who converted to a non-restrictive filling pattern was very similar to that in patients who had been non-restrictive at rest: IVRT and EDT both lengthened with stress, suggesting a fall in left atrial pressure, which was confirmed by a significant reduction in peak mitral E-wave acceleration, and a reduction in right ventricular systolic pressure. The factor underlying the difference between the two groups, therefore, appeared to be an increase or reduction in left atrial pressure during stress (5,6).
A fall in left atrial pressure at constant or increased SV is likely to depend on the LV being able to mount a significant positive inotropic response to dobutamine. Patients in group EA demonstrated the combination of a fall in left atrial pressure, a fall in LV cavity size, a rise in SV, and a significant rise in peak aortic acceleration, previously described as a sensitive measure of ventricular inotropy (14). In contrast, patients in group EE demonstrated physiological features of a further considerable rise in left atrial pressure at peak stress at constant SV. This difference in response could not be explained by differences in dobutamine infusion rate, LV cavity size, a reduction in venous return as reflected in SV, or worsening MR. However, patients in group EE had a much lower increment in peak aortic acceleration with stress compared with groups EA and AA, demonstrating that the positive inotropic response to stress was considerably impaired in group EE.
The extent of separation in LV filling pattern in individual patients between group EE and group EA was striking—IVRT and EDT both always fell in group EE and both always rose in group EA (Fig. 3, top), and that between changes in TR pressure drop and peak aortic acceleration was only slightly less so (Fig. 3, bottom). This biphasic response suggests a unifactorial rather than a multifactorial basis for the difference in response, but we have no evidence as to its nature. Ischemia may impair the inotropic response (16), but its extent, judged by the changes in QRS duration, PES, and wall motion score (17), was not more severe in group EE compared with group EA. Alternatively, the difference might have been the result of down-regulation of beta-receptors in group EE, although we have no direct evidence of its extent in the two groups.
Restrictive filling at peak stress proved a highly significant predictor of reduced survival (hazard ratio >9). Furthermore, it subsumed all other rest and stress-induced univariate predictors of prognosis. This increased risk of death in group EE was so great that, if confirmed, it might be allowed to influence clinical management in individual patients. Thus, early intervention, such as with an implantable cardioverter-defibrillator, would seem particularly warranted in these individuals. Conversely, patients who convert to a non-restrictive filling pattern at peak stress appear to have a much lower risk of death. Mechanisms underlying the impaired inotropic response to dobutamine would seem to merit further investigation, in particular determining whether they could be modified pharmacologically, with the aim of mitigating the very poor outcome in these patients.
A possible limitation of this study might be the possible effect of selection criteria, because group EA and AA were clearly differentiated at rest in that three criteria of restrictive filling were all present in the former and all were absent in the latter. Despite this, however, the response to stress in groups EA and AA was identical. Moreover, groups EA and EE had identical entry criteria, and yet their response to stress was fundamentally different. The study was noninvasive by design, and thus direct measures of absolute left atrial pressure were unavailable. However, our conclusions were based on change in left atrial pressure in individual patients rather than absolute values, for which echo-Doppler methods are more suitable (18). Patient numbers were quite adequate to determine the physiological mechanisms underlying changes in LV filling pattern with stress, but were smaller than those usually necessary to achieve statistical power in survival studies. However, the difference in survival between groups EE and EA was so great that it was readily demonstrable with small numbers. There may be lesser differences in survival between groups EA and AA that were not apparent with the numbers we used. Plasma brain natriuretic peptide has been reported as an independent prognostic marker in patients with systolic heart failure (19) although not to the same degree that persistence of restrictive filling proved to be in our study. Unfortunately, the results of plasma natriuretic peptides were not available when we embarked on the study.
A restrictive LV filling pattern at rest is recognized as being associated with increased mortality in ICM. We have demonstrated that further risk stratification may be available by assessing the response of LV filling pattern during pharmacological stress. Conversion of a restrictive filling pattern to a nonrestrictive filling pattern during stress is not only associated with evidence of a fall in left atrial pressure as a result of a preserved inotropic response, but also with increased survival compared with patients in whom restrictive filling persists at peak stress. Retention of a restrictive filling pattern at peak stress is associated with further rise in left atrial pressure, impairment of the inotropic response to stress, and, particularly, poor survival. Patients with end-stage systolic heart failure and restrictive filling at rest should, therefore, not be viewed as a homogeneous group. Pharmacological reversal of the impaired inotropic response to stress might represent a new approach to the treatment of this common and high-risk group of patients.
Dr. Duncan is supported by The Garfield Weston Trust.
- Abbreviations and Acronyms
- coronary artery disease
- confidence interval
- E-wave deceleration time
- E:A ratio
- ratio of early transmitral flow velocity to atrial flow velocity
- group AA
- non-restrictive at rest, remained non-restrictive at stress
- group EA
- restrictive at rest, became non-restrictive at peak stress
- group EE
- restrictive at rest, remained restrictive at peak stress
- ischemic cardiomyopathy
- isovolumic relaxation time
- left ventricle/ventricular
- mitral regurgitation
- peak aortic acceleration rate
- post-ejection shortening
- peak mitral E-wave acceleration rate
- systolic amplitude
- stroke volume
- tricuspid regurgitation
- Received December 20, 2004.
- Revision received March 31, 2005.
- Accepted April 5, 2005.
- American College of Cardiology Foundation
- Pinamonti B.,
- Di Lenarda A.,
- Sinagra G.,
- et al.,
- Heart Muscle Disease Study Group
- Mattheos M.,
- Shapiro E.,
- Oldershaw P.J.,
- Sacchetti R.,
- Gibson D.G.
- Appleton C.P.,
- Hatle L.K.,
- Popp R.L.
- Vanoverschelde J.L.,
- Raphael D.A.,
- Robert A.R.,
- Cosyns J.R.
- Rihal C.S.,
- Nishimura R.A.,
- Hatle L.K.,
- Bailey K.R.,
- Tajik A.J.
- Pinamonti B.,
- Zecchin M.,
- Di Lenarda A.,
- Gregori D.,
- Sinagra G.,
- Camerini F.
- Hansen A.,
- Haass M.,
- Zugck C.,
- et al.
- Duncan A.M.,
- Francis D.P.,
- Henein M.Y.,
- Gibson D.G.
- Noble M.I.M.,
- Trenchard D.,
- Guz A.
- Schiller N.B.,
- Shah P.M.,
- Crawford M.,
- et al.
- Steendijk P.,
- Baan J. Jr..,
- Van der Velde E.T.,
- Baan J.
- Duncan A.M.,
- Francis D.P.,
- Gibson D.G.,
- Henein M.Y.
- Ommen S.R.,
- Nishimura R.A.,
- Appleton C.P.,
- et al.
- Tsutamoto T.,
- Wada A.,
- Maeda K.,
- et al.