Background
Doppler echocardiography of mitral inflow or aortic outflow or both has been validated and advocated to guide biventricular (Biv) pacemaker optimization. A comprehensive and tailored Doppler echocardiographic evaluation may be required in patients with heart failure to assist with Biv pacemaker optimization. The third heart sound (S 3 ), an acoustic cardiographic parameter, has been demonstrated to be a highly specific finding for hemodynamic evaluation in patients with heart failure. The aims of this study were to evaluate the use of comprehensive Doppler echocardiography as a guide during Biv pacemaker optimization in patients after cardiac resynchronization therapy and to evaluate the feasibility of S 3 intensity to be a cost-efficient parameter for Biv pacemaker optimization compared with Doppler echocardiography.
Methods
Comprehensive Doppler echocardiographic evaluations were performed during Biv pacemaker optimization in 44 patients referred for pacemaker optimization (mean age, 71 ± 12 years; mean left ventricular ejection fraction, 34 ± 11%). Blinded assessment of S 3 intensity was performed simultaneously using acoustic cardiography. The correlation and improvement in cardiac hemodynamics were analyzed between the methods.
Results
Echocardiographically guided optimization resulted in significant improvements in the left ventricular outflow velocity-time integral (15.92 ± 4.77 to 18.51 ± 5.19 cm, P < .001), ejection time (278 ± 40 to 293 ± 40 ms, P < .001), myocardial performance index (0.57 ± 0.19 to 0.44 ± 0.14, P < .002), and peak pulmonary artery systolic pressure (42 ± 13 to 36 ± 11 mm Hg, P < .04) and decreased S 3 intensity from 4.81 ± 1.84 at baseline to 3.96 ± 1.22 after optimization ( P < .02) for the overall study group and from 6.63 ± 1.37 to 4.85 ± 1.13 ( P < .001) in the 18 patients with baseline S 3 intensity > 5.0. The correlation between echocardiographic and acoustic cardiographic S 3 intensity for optimal atrioventricular delay was 0.86 ( P < .001) and for optimal interventricular delay was 0.64 ( P < .001). Optimal atrioventricular delay was identical by echocardiographic and acoustic cardiographic S 3 intensity in 56%, and optimal interventricular delay was identical in 75% of patients. Pacemakers were permanently programmed on the basis of echocardiographic evaluation. In 35 patients available for follow up, the mean New York Heart Association class reduced from 2.55 ± 0.81 to 1.77 ± 0.90 ( P < .001) and the mean quality-of-life score as assessed by Minnesota Living With Heart Failure Questionnaire improved from 45 ± 28 to 32 ± 28 ( P = .08) at 2.5 ± 2.1 months.
Conclusion
Comprehensive echocardiographically guided Biv pacemaker optimization produces significant improvement in Doppler echocardiographic hemodynamics, a reduction in S 3 intensity, and an improvement in functional class in patients after cardiac resynchronization therapy.
Cardiac resynchronization therapy (CRT) has been effective and standard treatment in patients with drug-resistant heart failure (HF). Echocardiographically guided optimization of atrioventricular (AV) delay (AVD) improves left ventricular (LV) filling, and optimized interventricular delay (VVD) leads to further improvement in cardiac output in patients with biventricular (Biv) pacemaker. Advocated Doppler parameters include the aortic velocity-time integral (VTI), diastolic mitral inflow pattern, Doppler tissue imaging, and Doppler-derived dP/dt. Improved LV ejection fraction, New York Heart Association (NYHA) class, and cardiac output have been reported in follow-up studies. Despite these demonstrated benefits, echocardiographically guided Biv pacemaker optimization is not routinely performed, given that it is time, resource, and labor intensive. This suggests that methods that are simple and accurate and that shorten the time needed for echocardiographically guided optimization may allow a wider adoption of Biv pacemaker optimization.
Acoustic cardiography (Audicor; Inovise Medical, Inc, Portland, OR) has been demonstrated to be an accurate phonocardiographic method for cardiac hemodynamic assessment. The assessment of cardiac function is achieved by the detection and automated analysis of systolic and diastolic heart sounds and their temporal relationships to the electrocardiogram. The relevant parameters measured by acoustic cardiography in HF include (1) S 3 strength (expressed as 0-10 units), which correlates in a linear fashion to LV end-diastolic pressure and has been used to optimize preload in CRT patients ; (2) electromechanical activation time (EMAT), the interval from the onset of the QRS complex to the mitral component of S 1 , reflecting the time required for the left ventricle to generate sufficient force to close the mitral valve; and (3) LV systolic time (LVST), the time from the onset of QRS to closure of the aortic valve. EMAT has been shown to correlate with pulsed-wave (PW) Doppler echocardiography of mitral inflow or aortic outflow during Biv pacemaker optimization in patients with stable HF or those with conduction abnormalities with preserved heart function who had minimal S 3 intensity. In some studies, acoustic cardiography was not performed concurrently with echocardiographic evaluation.
In clinical practice, patients with symptomatic HF with Biv pacemakers have systolic HF, diastolic HF, elevated filling pressures, and S 3 intensity in varying combinations. In these patients, echocardiographically guided pacemaker optimization needs to take into consideration several other parameters besides LV filling and ejection, such as mitral regurgitation (MR) severity, the presence or absence of diastolic MR, tricuspid regurgitation, pulmonary artery systolic pressure (PAP), atrial filling by pulmonary veins, mechanical dyssynchrony, and the effect of respiration on cardiac hemodynamics. This tailored echocardiographic technique has been endorsed by recent guidelines from American Society of Echocardiography and results in significant improvement in HF.
We hypothesized that a comprehensive Doppler echocardiographic evaluation could guide pacemaker optimization in patients with varying degrees of HF after CRT. In addition, we compared S 3 intensity measured by phonocardiography against Doppler parameters in guiding pacemaker optimization in a consecutive series of patients referred for Biv pacemaker optimization.
Methods
Subjects
The study protocol was approved by the institutional review board, and all subjects provided written informed consent before enrollment. The study group was composed of 50 patients (mean age, 71 ± 12 years) referred to our center by their treating physicians for pacemaker optimization after CRT. All studies were performed as outpatients except in 3 patients who were hospitalized with HF at the time of optimization.
Study Protocol
All evaluation was performed after CRT. The pacemaker was interrogated using the algorithm shown in Table 1 . Underlying intrinsic heart disease, cardiac and valvular function, and cardiac filling pressures were evaluated by echocardiography in the left lateral decubitus position before pacemaker optimization.
| 1. Determine percentage Biv pacing. If <95%, determine mechanisms. |
| 2. Determine patient’s intrinsic AVD. This determines the range of AVDs available for programming. |
| 3. Determine backup atrial pacing rate, and turn it up to suppress ectopy and down to physiologic levels (60 beats/min) to allow atrial sensing. |
| 4. Determine heart rate variability, rate response, rate response sensitivity, and lead thresholds. |
| 5. Evaluate for the presence of sleep apnea. |
Doppler echocardiography and acoustic cardiography were performed simultaneously during pacemaker optimization. During pacemaker optimization, echocardiographic (T.Z.N.) and acoustic cardiographic evaluation (E.C. and E.G.) was carried out by independent experts for each method. Two blinded independent investigators (N.T. and J.Z.) performed offline analysis of all echocardiographic data. Investigators performing offline analysis were blinded to online assessments of optimal AVD and VVD settings. Two sonographers (M.C.M. and G.V.) performed imaging. Two investigators (N.T. and S.D.) collected follow-up data.
Conventional Doppler Echocardiographic Analysis
LV end-systolic and end-diastolic dimensions, septal and posterior wall thickness, and left atrial diameter in end-systole were obtained in the parasternal long-axis view. LV ejection fraction was measured using the biplane Simpson’s method. Mitral inflow and aortic ejection onset and termination were marked using PW Doppler signals. This allowed the assessment of ejection duration as well as mitral inflow filling time. Myocardial performance index time was measured as ( A − B )/ B , where A is the time from the end of the mitral inflow A wave to the beginning of the next E wave, and B is the ejection duration. The maximum right ventricular–right atrial gradient was obtained with continuous-wave Doppler using a standard method. Mitral inflow peak E and A velocities and E-wave deceleration time were measured using PW Doppler with the sample volume placed at the tip of the mitral leaflets. The PW Doppler sample volume was placed 0.5 to 1 cm below the aortic valve to obtain the LV ejection duration and aortic VTI. The frame rate was kept above 100 frames/s by using a single-focus, narrow imaging sector and appropriate depth and frame rate. Parallel Doppler beam alignment to myocardial segments and color Doppler was used for all Doppler data acquisition. An electrocardiogram was displayed on the ultrasound system, and 5 cardiac cycles were used for each data acquisition for patients in sinus rhythm. For patients in atrial fibrillation or those with atrial or ventricular ectopic beats, 10 beats were acquired in each view. Raw data were stored digitally as Digital Imaging and Communications in Medicine cine loops of ≥5 cardiac cycles and transferred for offline analysis to a customized dedicated workstation equipped with custom-built software (EchoPAC PC Dimension version 6.0.1; GE Vingmed Ultrasound AS, Horten, Norway) via the Internet.
Echocardiographically Guided Optimization
For all patients, AVD optimization was performed first. Broad principles of PW Doppler echocardiography during pacemaker optimization were followed on the basis of published recommendations, as well as using a previously described algorithm. Care was taken to ensure that echocardiographic sample volume and transducer were in the same position for each testing delay to reduce sampling error. AVDs of 30 to 300 ms were tested if possible, with the majority in an available range of 50 to 250 ms. In patients with native AV conduction (n = 21), echocardiographic and acoustic cardiographic data were obtained during native rhythm. AVD was lowered until complete Biv capture was attained to determine the upper limit of AVD. Ritter’s method was used if feasible to determine the optimal AVD. In patients in whom Ritter’s method was not feasible, an iterative method was used whereby AVD was changed in increments of 10 to 20 ms depending on native AVD and mitral inflow pattern. PW Doppler of the LV outflow tract was performed and LV ejection duration and peak velocity were measured at each AVD and VVD. Optimal AVD was selected on the basis of the “best diastolic LV filling pattern” and highest LV ejection duration and peak velocity, but in patients who had diastolic MR or tricuspid regurgitation, significant systolic MR, restrictive pulmonary vein filling patterns, prominent pulmonary vein atrial reversal, and measureable and elevated PAP, these parameters were reevaluated at optimal AVD to ensure maximum improvement in diastolic MR or tricuspid regurgitation, least restrictive pulmonary vein filling pattern (least S:D reversal, highest D-wave deceleration time), minimum pulmonary vein atrial reversal, least PAP, and minimum systolic MR along with optimum LV VTI and diastolic filling pattern. In some patients, mitral filling time had to be compromised to minimize pulmonary vein flow reversal, while in others, the mitral inflow A wave had to be truncated to avoid diastolic MR. MR was graded semiquantitatively as MR jet area in relation to left atrial area and averaged in apical 4-chamber, 2-chamber, and 3-chamber views. LV VTI, myocardial performance index, and mitral inflow VTI were measured offline.
The optimal AVD was programmed and then VVD was adjusted at the optimal AVD. Values of LV preexcitation tested ranged from 4 to 30 ms, and values of RV preexcitation tested ranged from 10 to 20 ms. Mechanical dyssynchrony was assessed by visual assessment of color-coded tissue velocity imaging in apical 4-chamber, 2-chamber, and 3-chamber views. Progressively increasing LV or right ventricular preexcitation was tested until maximum LV ejection duration was obtained. In patients with significant posterolateral wall delay on tissue velocity imaging or greater than mild MR and measurable PAP, these were reevaluated to ensure minimum values at optimal VVD. Mitral inflow was reevaluated after optimal VVD programming, and AVD was readjusted if required.
Acoustic Cardiography
Proprietary dual-purpose sensors were placed in the V 3 or V 4 positions on the patient’s chest wall to record digital electrocardiographic and heart-sound data. During echocardiographic evaluation, a computerized acoustic cardiographic measurement was performed simultaneously at each selected pacemaker setting. To avoid interference in data, the sonographer removed the transducer from the chest wall during acoustic cardiographic data recording The computerized acoustic cardiography representing 12 cardiac cycles was produced immediately within 10 seconds of collection, including analysis of S 3 strength, EMAT, and LVST and a timely trend map against the pacemaker testing.
At the end of the optimization procedure, the pacemaker was programmed on the basis of the results of the online echocardiographic evaluation. NYHA class and quality of life were assessed using the Minnesota Living With Heart Failure Questionnaire at baseline and by verbal interrogation and via a questionnaire sent out 1 month after optimization.
Statistical Analysis
Data are presented as mean ± SD. All parameters at baseline and after optimization were compared using paired 2-tailed Student’s t test. Pearson’s correlation analysis was performed between optimized AVD and VVD by echocardiography and acoustic cardiography. P values < .05 were considered statistically significant.
Results
Fifty patients were evaluated. All had >85% Biv pacing at baseline. The mean time since CRT was 13.8 months. Six patients in whom adequate-quality acoustic cardiographic data could not be obtained were excluded. Twenty patients had Medtronic devices (Medtronic Inc, Minneapolis, MN), 13 had Boston Scientific devices (Boston Scientific Corporation, Natick, MA), and 11 had St Jude devices (St Jude Medical, St Paul, MN). Five patients had atrial fibrillation at the time of optimization, and 1 had atrial flutter. AVD could therefore be optimized in 38 patients. Nine patients had prior mitral valve surgery with annuloplasty rings or valve replacements. Twenty-four patients had atrial pacing, and 14 had atrial sensing. Because of device characteristics, VVD could be optimized in 41 patients. The baseline clinical characteristic of the study population at the time of pacemaker optimization are shown in Table 2 . Patients had significant LV enlargement and impaired LV ejection fractions. Fifty percent of patients were in NYHA class III or IV at baseline. The majority were on diuretics, and as many as 43% were on amiodarone, thus indicating significant cardiac dysfunction despite CRT.
| Variable | Value |
|---|---|
| Age (y) | 71 ± 12 |
| Men/women | 33(75%)/11(25%) |
| NYHA class | 2.55 ± 0.81 |
| I | 2 (5%) |
| II | 20 (45%) |
| III | 14 (32%) |
| IV | 8 (18%) |
| Rhythm | |
| Sinus | 39 (89%) |
| Atrial fibrillation/flutter | 5 (11%) |
| Cardiomyopathy | |
| Ischemic | 32 (73%) |
| Nonischemic | 12 (27%) |
| Other clinical conditions | |
| Hypertension | 30 (68%) |
| Diabetes mellitus | 14 (32%) |
| Medications | |
| Diuretics | 34 (77%) |
| β-blockers | 35 (80%) |
| ACE inhibitors | 24 (55%) |
| Amiodarone | 19 (43%) |
| Echocardiographic findings | |
| LV diastolic diameter (cm) | 6.05 ± 1.14 |
| LV systolic diameter (cm) | 5.08 ± 1.18 |
| LV ejection fraction (%) | 33.77 ± 11.12 |
| LA diameter, superior to inferior (cm) | 6.13 ± 1.19 |
| LA diameter, anterior to posterior (cm) | 4.98 ± 0.88 |
The mean short AVD tested was 70 ± 35 ms, and the mean longest AVD tested was 225 ± 65 ms. In 21 patients, the intrinsic AVD was tested at 253 ± 74 ms. Optimization led to a significant improvement in systolic and diastolic Doppler echocardiographic as well as acoustic cardiographic parameters, as shown in Table 3 . The effects of optimization on LV VTI, S 3 intensity, EMAT, and LVST at baseline, shortest AVD, longest AVD, optimal AVD, and optimal VVD in 33 patients in whom paired data were available are shown in Figure 1 . Figure 2 shows the correlation between baseline and optimal S 3 intensity and LV VTI and between baseline and optimal EMAT and LV VTI. S 3 intensity also correlated well with Doppler echocardiographic LV ejection duration both at baseline ( r = −0.46, P < .001) and at optimal AVD ( r = −0.43, P < .001). Besides systolic parameters, S 3 intensity also correlated with Doppler echocardiographic filling. Figure 3 shows the correlation between S 3 intensity and diastolic filling parameters at baseline and after optimization in 35 patients without prior mitral valve surgery.
| Method | Parameter | Baseline | After echocardiographically guided optimization | Change ‡ |
|---|---|---|---|---|
| Echocardiography | LVOT VTI (cm) | 15.92 ± 4.77 | 18.51 ± 5.19 † | 17% |
| LV ET (ms) | 278 ± 40 | 293 ± 40 | 5.4% | |
| MI VTI (cm) | 20.08 ± 8.78 | 21.09 ± 9.7 ∗ | 4.46% | |
| ET (ms) | 277.75 ± 41.01 | 292.44 ± 40.12 † | 5.5% | |
| FT (ms) | 421.71 ± 110.81 | 432.01 ± 108.91 | 3.35% | |
| MPI | 0.57 ± 0.19 | 0.44 ± 0.14 † | −21% | |
| MR | 2.1 ± 4.7 | 1.18 ± 0.9 † | −15% | |
| PAP (mm Hg) | 41.88 ± 12.72 | 36.2 ± 10.9 † | −14% | |
| E wave (cm/s) | 0.93 ± 0.32 | 0.82 ± 0.29 | −11% | |
| DT (ms) | 203 ± 76 | 228 ± 85 | 17% | |
| Acoustic cardiography | S 3 intensity | 4.81 ± 1.84 | 3.96 ± 1.23 † | −12% |
| EMAT (ms) | 159.99 ± 32.09 | 146.67 ± 27.07 ∗ | −6% | |
| LVST (ms) | 330.14 ± 34.95 | 337.6 ± 38.55 | 2.47% |
Figure 4 shows the effect of pacemaker optimization on acoustic cardiographic S 3 intensity in a representative patient. In this patient, echocardiographic and acoustic cardiographic techniques determined identical optimal AVD and VVD pacemaker setting. S 3 became undetectable, and this was associated with improvements in LV filling and ejection.
The mean baseline AVD was 150 ± 43 ms, and the mean LV offset was 8 ± 13 ms; the mean final AVD was 164 ± 55 ms, and the mean LV offset was 4.5 ± 6.7 ms ( P = NS vs baseline AVD and VVD). The mean change in optimal AVD was 15 ms. With respect to VVD, 5 patients required RV preexcitation (4 at 10 ms and 1 at 15 ms), and 16 required LV preexcitation (3 at <10 ms, 8 at 10 ms, 1 at 15 ms, and 4 at 20 ms). The mean baseline and optimal AVDs in A-sensed patients were 131 ± 46 and 114 ± 35 ms and in A-paced patients were 165 ± 38 ms ( P < .03 vs A-sensed patients) and 192 ± 44 ms ( P < .001 vs A-sensed patients), respectively.
There was no significant change in LV ejection time, LV VTI, diastolic filling time, PAP, or S 3 intensity at optimal AVD and VVD compared with optimal AVD alone in the overall study group. In patients with atrial fibrillation, a 14% improvement in LV VTI occurred with VVD optimization compared with baseline.
Optimal AVD and VVD by Echocardiographic and Acoustic Cardiographic Methods
There was no significant difference in optimal AVD and VVD between these two techniques. Of the 36 patients with paired echocardiographic and acoustic cardiographic data, there was complete agreement (ie, a difference of 0 ms in AVD between the two methods) in 58% (n = 21) for optimal AVD. In 6 patients (17%), there was a discordance of <20 ms, and in 9 patients (25%), there was a discordance of >20 ms. Optimal VVD was identical by echocardiography and S 3 intensity in 30 patients (75%); in 5 patients (12.5%), there was a discordance of <10 ms, and in 5 patients (12.5%), the discordance was >10 ms. Figure 5 shows the correlation between optimal AVD by echocardiography and the AVD at which acoustic cardiography showed minimum S 3 intensity. As noted in Figure 5 , a wide range of optimal AVDs was present both by echocardiography (50-300 ms) and acoustic cardiography (50-300 ms).
