Cardiac resynchronization therapy (CRT) is a well-accepted and proven therapy for the treatment of patients with advanced heart failure (HF), significant left ventricular (LV) systolic dysfunction, and a wide QRS complex on electrocardiogram (ECG). In this patient population, large multicenter studies have demonstrated that CRT improves symptoms of HF, exercise capacity, LV size and systolic function, hospitalization rates, and mortality.1–4 Most studies of this therapy have used the presence of electrical dyssynchrony (QRS ≥120–130ms) as an entrance criterion for enrollment. This is because CRT was thought to provide its benefit by resynchronizing the electrical activation of the LV in patients with marked disparities in timing of electrical activation. In addition, QRS duration is an easily quantified variable that can be readily and cheaply measured in all patients. Current research has focused on the role of mechanical dyssynchrony for a number of reasons. It is now well recognized that electrical and mechanical dyssynchrony are quite different despite a modest correlation between the two.
In one study of advanced HF patients using tissue velocity imaging (TVI), 27% of patients with a narrow QRS (<120ms) had mechanical dyssynchrony by TVI and 30% of patients with a very wide QRS (>150ms) had no significant mechanical dyssynchrony.5 In a 3D echocardiography (ECHO) study of 62 HF patients with ejection fraction (EF) <40%, 37% of the patients with narrow QRS had mechanical dyssynchrony and 38% of those with wide QRS did not have mechanical dyssynchrony.6 There are a number of reasons why mechanical dyssynchrony may be an important variable to measure in addition to electrical dyssynchrony. First, the QRS duration is only a surface ECG summation of the time required for all ventricular depolarization; this includes electrical activity in the right ventricle (RV) and in general does not provide detail on the timing of activation of different regions of the LV. Some areas of electrical conduction in the LV may not show up as electrical activity on the surface ECG. Second, although electrical activity is essential for systolic heart function, it is an early step in the process. Co-ordinated myocyte contraction is what moves blood across the aortic valve and there can be regional and variable delays in electromechanical coupling. Third, QRS duration has not uniformly predicted response to CRT. Fourth, measurement of mechanical dyssynchrony may have great value in helping to better understand the mechanism of action of CRT. Finally, measurement of mechanical dyssynchrony can be of value in clinical decisions related to the treatment of patients with HF and/or CRT such as selection of patients for CRT, lead placement, assessing response to CRT, and optimizing pacemaker settings or physiological tailoring of CRT to the individual patient.
This review focuses on echocardiographic measurement of mechanical dyssynchrony. We will review various methodologies for measuring mechanical dyssynchrony, discuss single-center and multicenter studies that have addressed the utility of measuring mechanical dyssynchrony, and provide recommendations on the clinical use of echocardiographic measurement of dyssynchrony in managing patients with HF and CRT.
Physiology of Mechanical Dyssynchrony
Left bundle branch block (LBBB) is seen in about 30% of HF patients, and is characterized by a wide prolonged QRS on ECG. The normal activation of the LV is disturbed in this condition, resulting in a delayed activation of the LV lateral wall. When the early-activated inter-ventricular septum contracts, very little mechanical work is produced since blood is simply pushed toward the relaxed lateral wall. The late-activated LV lateral wall is pre-stretched and performs a higher workload than normal when the myocytes ultimately contract. This disparity results in many differences in structure and function between the early- and late-activated regions of myocardium, including different expression of multiple proteins, heterogeneous calcium-handling properties, changes in blood flow and wall thickness, and variations in electrical conduction speeds within the LV.7–11 In patients with previous myocardial infarctions and electrical dyssynchrony, the scenario is more variable and complex since the functional line of electrical block and the latest site of electrical and mechanical activation can be quite varied.12
CRT is a relatively new treatment for patients with severe HF and wide QRS on optimal medical therapy. With this therapy, both the RV and LV are paced via an implanted device. An RV lead is placed in the apex or septum and an LV lead in a lateral or postero-lateral position within the coronary sinus. Pacing the heart in this way acutely and chronically improves the efficiency of the LV by improving myocardial contraction timing. Despite the impressive success of CRT, approximately 20–35% of patients meeting current guidelines for CRT do not improve. Furthermore, there is speculation that some patients not currently meeting standard CRT indications may improve with this treatment.13 As a result, there is considerable interest in the exact mechanisms of improvement with CRT. In order to better understand and apply this new therapy to the large numbers of potential HF candidates, intensive efforts have been made to measure mechanical dyssynchrony. We will discuss the types of mechanical dyssynchrony present in patients with HF and/or CRT and then describe the main echocardiographic methodologies that have been used to quantify this pathology.
Types of Mechanical Dyssynchrony
Atrioventricular Dyssynchrony
Abnormal timing of contraction of the left atrium (LA) with respect to the LV impairs cardiac function. Atrial contraction should occur after passive filling of the LV and should be completed before LV contraction. Atrioventricular (AV) dyssynchrony can be seen on echo Doppler mitral inflow velocities as either a fusion of early (E) and late (A) waves (AV delay too long) or a truncation of the A wave (AV delay too short). AV dyssynchrony can result in impaired filling or preload of the LV, resulting in reduced myocyte stretch and reduced LV stroke volume via the Frank–Starling mechanism. Alternatively, AV dyssynchrony can result in late diastolic mitral regurgitation if filling of the LV from active atrial contraction has finished well before LV contraction. Finally, simultaneous contraction of the LA and the LV will result in retrograde blood flow into the pulmonary veins, increased pulmonary venous pressure, and, possibly, pulmonary edema.
Inter-ventricular Dyssynchrony
Inter-ventricular dyssynchrony is characterized as a prolonged delay between mechanical activations of the RV and the LV. This can be easily measured by echo Doppler acquisition of aortic and pulmonic outflow velocities. A difference in time of onset between the two velocities of >40ms is considered abnormal.2 Although some studies have shown this variable to be predictive of response to CRT, others have not found it to be of great value.14–16
Intra-ventricular Dyssynchrony
Intra-ventricular mechanical dyssynchrony is a term used to describe the presence of abnormal timing of motion within the LV. In the normal heart, most—if not all—myocardial segments demonstrate onset of motion and peak motion that is highly synchronized in longitudinal, radial, and circumferential planes. This motion can be measured as velocity, displacement, or strain. Current research suggests that intra- ventricular dyssynchrony is a very important pathophysiological abnormality that leads to the development and progression of HF. The following paragraphs will describe many of the echocardiographic techniques currently used to measure intra-ventricular mechanical dyssynchrony and the relative advantages and disadvantages of each.
Measurements of Intra-ventricular Mechanical Dyssynchrony
A number of techniques have been proposed for quantifying mechanical dyssynchrony. Echocardiography is a practical and cost- effective tool that has become widely used for this purpose. In particular, tissue Doppler imaging (TDI) and, more recently, speckle tracking echocardiography (STE) have emerged as robust techniques for quantifying mechanical LV dyssynchrony. However, no single measure has been agreed upon as the most useful for describing intra-ventricular dyssynchrony or predicting response to CRT.
M-mode
M-mode images of the parasternal long-axis view can be used to measure mechanical dyssynchrony using standard ultrasound equipment without more advanced motion quantification software. A difference of more than 130ms between the maximal inward movements of the basal septal and posterior walls has been used to identify dyssynchrony.17 This method has a number of limitations: the movement of only two regions of myocardium can be measured, only motion along the line of incidence of the ultrasound beam can be assessed, accurate identification of peak excursion is difficult in many patients due to broad peaks, and defining the endocardium can be difficult.18
Tissue Doppler Imaging
TDI utilizes the Doppler effect to determine the velocity of reflective material in the path of the ultrasound beam. The frequency of the reflected ultrasound beam is proportional to the component of velocity in the direction of the transducer. If the measurements of high-velocity red blood cells are filtered out, the result is a signal containing the velocities of myocardial tissue in relation to the transducer. Regions of interest can be placed on the echocardiographic image so that motion is captured within a specific area of the myocardium and curves representing the motion of this area over time can be generated. Motion can be depicted as velocity, displacement, or strain. To assess motion in the longitudinal direction, sample volumes are typically placed at the base and mid-ventricle on each pair of opposing walls in apical four-chamber, two-chamber, and long-axis views. In the normal heart, each of the LV wall segments moves toward the transducer during systole, and the motion of each segment peaks at the same time. In a dyssynchronous heart, peak regional motion curves do not overlap. A multitude of dyssynchrony indices have been proposed but, in general, differences in the times at which regional motion curves reach their peak value are measured. The standard deviation of these times or the maximum difference between two peaks are the most common methods used to quantify dyssynchrony.
Tissue Velocity Imaging
Tissue velocity imaging (TVI) is the most commonly used technique for quantifying intra-ventricular dyssynchrony. Regional velocities have been shown to agree with phantoms, sonomicrometry, and conventional M-mode echocardiography.19–21 Typical TVI images from a normal subject and a patient with LV dyssynchrony before and after CRT are shown in Figure 1A. In the normal heart, velocity curves from all regions of the LV peak together during early systole. In the patient with HF, the TVI curves reach peak velocity at differing times during systole. After CRT, the times to peak velocity within systole move closer together. Yu et al. proposed one commonly used index of dyssynchrony using TVI in which the standard deviation of the time to peak systolic velocity for 12 basal and mid-level segments is used to describe LV dyssynchrony. A cut-off value of 32ms is most frequently used to predict response to CRT.22,23 Another commonly used index is the time difference between the earliest and latest peaks in the regional systolic velocity curves. A cut-off value of 65ms has been proposed.24,25
The largest body of literature supports the use of TVI to measure dyssynchrony, and many ultrasound machines are capable of accurately measuring myocardial velocity.26 However, in some patients peak velocity is difficult to determine due to multiple, jagged, or flat peaks in the velocity curves. In addition, this technique cannot distinguish between active contraction and passive motion due to tethering or translational movement of the heart. Velocity measurements are also limited to one dimension, along the line of the ultrasonic beam, so the 3D motion of the heart may not be adequately represented. Furthermore, velocity measurements and timing of peak velocities can vary significantly with small changes in the placement or size of the region of interest, and can be affected by whether or not the region of interest remains stationary during the cardiac cycle or moves with the myocardium to remain in the mid-wall location.27 Finally, there are serious concerns regarding the intra-observer, inter-observer, and inter-laboratory reproducibility of these measurements.28
Tissue Tracking
Tissue tracking (TT) is a technique that describes the displacement, rather than velocity, of myocardium relative to the ultrasound transducer. Displacement can be derived from TVI data by integration of the velocity profiles over time. Displacement curves represent the distance that the regions of interest move from their locations at a reference time during the cardiac cycle (usually the onset of the QRS or the closing of the mitral valve). All regions of the normal heart demonstrate peak systolic excursion at about the time of aortic valve closure. Figure 1B shows examples of TT tracings from a normal subject and a patient with HF, pre- and post-CRT. In the normal subject all TT curves begin upward movement together, peak together at the time of aortic valve closure, and return to their original position together. In the patient with HF and intra-ventricular dyssynchrony, the septal wall initially moves away from the transducer in the wrong direction and then peaks in the middle of diastole. Following CRT, there is more normal motion of the septal wall without the initial negative deflection and with the peak motion occurring closer to the time of aortic valve closure. Also of note, the mean systolic basal and mid-ventricular wall movement toward the apex before CRT was 3.0mm, as seen in Figure 1B. After CRT this improved to 9.0mm, demonstrating a 300% improvement in the longitudinal systolic function of these walls. Dyssynchrony has been defined as the standard deviation of the times to peak displacement of 12 basal and mid-LV regions of interest, or as the number of wall segments with peak systolic excursion after aortic valve closure.22,29–31
The displacement curves generated in TT are often smoother, with less noise than velocity curves from TVI. The identification of times to peak wall displacement may be less variable with this technique. Inter- observer and intra-observer reproducibility are higher for TT-derived peak wall motion timings than TVI-derived timings.29,32 However, the software for acquiring and displaying TT curves is not available on all ultrasound machines. Like TVI, TT also can only be used to detect displacement in one direction, and may not optimally describe 3D heart motion. Finally, the ability of this technique to predict responders to CRT has not been studied as extensively as TVI, and direct comparisons between methods have yielded mixed results.22,32
Doppler-derived Strain and Strain Rate
Myocardial strain, defined as a percentage change in length, can be measured to quantify regional myocardial deformation or contraction. Myocardial strain rate may be derived from TDI by differentiating velocity over the distance from the transducer. Like velocity, strain rate can be integrated over time to yield strain. The standard deviation of times to peak strain rate or strain have been suggested to be sensitive and effective measures of myocardial synchrony.33–35 However, Doppler-based strain measurements have not consistently predicted response to CRT.15,22
Since strain and strain rate depend on velocity differences between nearby regions of myocardium, this method has the advantage of differentiating motion due to active contraction from tethering, or translational motion of the heart. However, inter-observer and intra- observer variability is relatively high, and temporal resolution is less than that of TVI or TT.33
Speckle Tracking Echocardiography
Another ultrasound technique has recently been introduced that does not rely on the Doppler effect to determine regional velocity, displacement, or strain information. STE involves the identification and tracking of stable speckle patterns in the 2D ultrasound image, which are assumed to correspond to fixed points within the myocardial tissue. The relative motion of these speckles represents the motion of the myocardium. Figure 1C shows STE curves generated from six segments of a mid-ventricular short-axis view of the LV. In the normal subject all six myocardial regions demonstrate strain (myocardial thickening) curves that move together and peak at end-systole. In the HF patient there is marked variation in regional strain, which improves following CRT.
STE measurements of strain in both the longitudinal and radial planes of the LV, as well as LV rotation, agree well with sonomicrometry and tagged magnetic resonance imaging (MRI).36 In a study by Suffoletto et al., STE measures were also found to correlate well with Doppler- derived indices of dyssynchrony.24 In addition, these investigators found that a time difference of >130ms between peak septal wall and posterior wall radial strain predicted long-term response to CRT. The standard deviation of times to peak strain can also be used to describe LV dyssynchrony in longitudinal, radial, circumferential, or rotational LV motion. Furthermore, the difference between apical and basal rotations of the LV is defined as LV twist or torsion, and has been suggested to be a sensitive indicator of global LV function. Dyssynchrony between the time of peak apical rotation and basal rotation has been shown to occur in HF, and is improved with CRT.37
STE imaging is less dependent on the angle of incidence of the ultrasound beam than TDI. Furthermore, not only is longitudinal motion measured by STE, but motion in radial or circumferential directions, as well as rotational motion, can also be assessed. Temporal resolution using STE can be reduced compared with TDI, however. In addition, out-of-plane motion is not detected by 2D STE.
3D Echocardiography
Echo images have been used to render 3D images of the heart for nearly 40 years; however, only recently has 3D echo been used to quantify mechanical dyssynchrony. In 3D techniques for measuring dyssynchrony, the LV volume is divided into 16 pyramid-shaped segments that correspond to 16 regions across the LV wall. Dyssynchrony is quantified as the standard deviation of time to minimal LV volume for the 16 regional volumes measured. Using this echo technique, Kapetanakis et al. reported that, independent of QRS duration, the magnitude of LV mechanical dyssynchrony correlated well with the extent of LV systolic dysfunction.6
3D echocardiography is not as widely used to measure mechanical dyssynchrony as more traditional echocardiography techniques such as TVI or TT. However, recent studies have shown that measurements correlate well with gold standards such as MRI. Also, advances in both imaging equipment and computational power for realtime 3D echo analysis have made it possible to measure, quantify, and interpret LV dyssynchrony in a timely and accurate manner, with the combined acquisition time and analysis reported to be five to seven minutes.38
Clinical Utilization of Mechanical Dyssynchrony in Heart Failure and Cardiac Resynchronization Therapy
Prognosis
Electrical dyssynchrony as measured by QRS duration is an independent predictor of high morbidity and mortality in patients with HF.39,40 Mechanical dyssynchrony may also be a marker of poor prognosis in patients with advanced HF. Bader et al. followed over 100 patients with HF and a mean EF of 31% for one year.41 Event-free survival was approximately 4% in patients without significant mechanical dyssynchrony and approximately 47% in those without mechanical dyssynchrony. This predictive value was independent of QRS duration or LVEF. By contrast, the presence of mechanical dyssynchrony is associated with a better prognosis once a patient receives CRT.16 Patients with mechanical dyssynchrony prior to CRT as measured by TVI had a 6% event rate (hospitalization for HF or death) in the year following the implant versus a 50% event rate for patients without significant mechanical dyssynchrony.
Selection of Patients for Cardiac Resynchronization Therapy
Retrospective Studies in Patients with Wide QRS
A number of retrospective, mostly single-center studies have assessed mechanical dyssynchrony as a predictor of response to CRT. A few of these studies will be mentioned to highlight various techniques and measurements that have been found useful in patient selection, without attempting to be comprehensive. Sogaard et al. found that the number of basal myocardial segments with delayed longitudinal contraction correlated with the extent of LV systolic function improvement and beneficial reverse remodeling.31 Yu and colleagues have published many papers evaluating longitudinal TVI and various time to peak velocity measurements as predictors of response to CRT. In their laboratory, TVI measures of longitudinal mechanical dyssynchrony prior to CRT are powerful predictors of the LV remodeling response.42 Bax et al. studied 85 patients receiving CRT for standard indications and found that the only significant predictor of response was intra-ventricular dyssynchrony as measured by TVI.16 Using STE, Suffoletto et al. found that a time difference of >130ms between peak septal and posterior wall radial strain predicted long-term response to CRT (defined as an increased EF of >15%) with 89% specificity and 83% sensitivity.24
Some investigators have assessed mechanical dyssynchrony in two or more planes in order to better quantify the magnitude and extent of this abnormality. Gorcsan et al. retrospectively analyzed data from 190 HF patients receiving CRT at two institutions. Dyssynchrony was measured in both longitudinal (TVI) and radial (STE) planes.43 Either baseline longitudinal or radial dyssynchrony (but not both) predicted an improvement in EF of ≥15% in 58% of patients. However, the presence of mechanical dyssynchrony in both planes had a positive predictive value (PPV) for EF response of 94%. The EF response rate in patients without dyssynchrony in either plane (negative predictive value [NPV]) was 10–21% depending on the TVI measure used. Similarly, our laboratory retrospectively analyzed mechanical dyssynchrony in longitudinal, radial, and circumferential planes in 70 consecutive HF patients receiving CRT for standard indications.44 A decrease in LV end- systolic volume (LVESV) ≥15% was found in 57% of patients. The PPV for this remodeling response was approximately 70% for patients with either longitudinal or radial dyssynchrony. The PPV for dyssynchrony in both of these planes was 90% and the NPV was approximately 30%. Data from these two studies are consistent and suggest that the presence of mechanical dyssynchrony in both longitudinal and radial planes can provide increased predictive value compared with measurements of dyssynchrony in a single plane.
Multicenter Prospective Studies in Patients with Wide QRS
Most of the large prospective multicenter studies demonstrating the clinical value of CRT in HF patients have not incorporated any measures of mechanical dyssynchrony in their protocols. An exception is the CArdiac REsynchronization in Heart Failure (CARE-HF) study, which required some echocardiographic measure of mechanical dyssynchrony to be present in patients with QRS duration 120–149ms.2
The largest prospective multicenter study of mechanical dyssynchrony variables as predictors of response to CRT is the Predictors of Response to CRT (PROSPECT) trial.28 In this trial, multiple echocardiographic indices of dyssynchrony were measured at baseline in almost 500 HF patients meeting standard criteria for CRT at 53 centers in the US, Europe, and Hong Kong. The 12 measured echocardiographic dyssynchrony variables predicted LVESV response to CRT with variable sensitivity (9–77%) and specificity (31–93%). No single measure of dyssynchrony demonstrated enough predictive value to be recommended for use in general clinical practice to improve patient selection for CRT. Based on this study, some have argued that mechanical dyssynchrony measurements are not of significant value in helping to select patients for CRT. However, there are a number of methodological and practical issues that should be mentioned. The study was performed across three continents with echocardiographic data acquired using machines from several different manufacturers. There were three separate core laboratories analyzing the data. There was high intra-observer, inter-observer, and inter-laboratory variability in the measurement of the dyssynchrony variables. In addition, only approximately 57% of the patients in the study had analyzable LV volume and longitudinal dyssynchrony pre- and post-CRT data for assessing predictive response. Finally, the study was designed fairly early in the evolution of echocardiographic methodology for measuring mechanical dyssynchrony and did not incorporate newer technologies such as STE. Rather than proving that measurement of mechanical dyssynchrony is not important, this study demonstrated the need for better standardization and consistency of data acquisition and analysis, and better and less variable measures of mechanical dyssynchrony.
A sub-analysis of data from 286 PROSPECT patients was performed by grouping patients into super-responders, responders, non-responders, and negative responders based on the change in LVESV at six months after CRT.45 This study demonstrated that measures of electrical dyssynchrony (QRS width), inter-ventricular mechanical dyssynchrony, and intra- ventricular mechanical dyssynchrony (time delay between peak lateral and septal longitudinal velocity), in addition to other clinical variables, were statistically significant predictors of LV remodeling response to CRT.
A second, smaller, prospective multicenter study of mechanical dyssynchrony echocardiographic measures as predictors of response to CRT was the PROspective Minnesota study of ECHO/TDI in cardiac resynchronization therapy (PROMISE-CRT) study.29 This study enrolled 71 patients with standard indications for CRT at nine sites in the Minneapolis-St Paul and surrounding area. Patients had multiple echocardiographic dyssynchrony variables measured before CRT and at one week, three months, and six months following device implantation. All echocardiographic data were acquired using the same equipment and a single core laboratory analyzed all data. Availability of complete echocardiographic data and variability of the mechanical dyssynchrony measurements were in general better than reported in PROSPECT. Three baseline measures of mechanical dyssynchrony individually explained 12–30% of the individual variation in end-systolic volume response to CRT. Using a cut-off of 55ms for the standard deviation of time to peak radial strain as measured by STE, PPV for an improvement in LV end-systolic volume of ≥15ml was 75% and NPV was 69%.
Narrow QRS
The role of CRT in patients with QRS duration <120ms is controversial. Achilli et al. studied 52 patients with mechanical dyssynchrony.46 The 14 patients with narrow QRS had similar improvements in LV size and function, mitral regurgitation, and six-minute-walk test (6MWT) distance compared with the wide-QRS patients. Yu et al. studied 102 patients with advanced HF.47 The 51 patients with narrow QRS and mechanical dyssynchrony by TVI improved New York Heart Association (NYHA) class, exercise capacity, 6MWT, EF, LVESV, and mitral regurgitation, similar to results in patients with wide QRS. In addition, there was a very good correlation between baseline dyssynchrony and remodeling response, with regression lines of similar slope in both groups. Bleeker et al. studied 66 patients with mechanical dyssynchrony, half with a narrow QRS.48 Improvements in LV remodeling and clinical response were significant and similar in both groups.
The only multicenter prospective study of CRT in patients with narrow QRS was the ReThinQ study.49 In this study 172 patients with advanced HF, EF ≤35%, QRS <130ms, and mechanical dyssynchrony by TVI were randomized to CRT or medical therapy for six months. The groups did not differ in the primary end-point (percentage of patients with an increase in peak oxygen consumption on cardiopulmonary exercise test of >1ml/kg/min), quality of life score, 6MWT, or echocardiographic measures of LV size and function. Although the authors concluded that CRT may not benefit patients with moderate to severe HF and narrow QRS, there are a number of limitations to this study that should be recognized. The methodologies used to measure mechanical dyssynchrony and meet enrollment criteria were TVI and M-mode performed and analyzed at each investigative site. Sites had different echo machines and varying levels of expertise at acquiring and interpreting TVI data. Additionally, no data were reported on dyssynchrony measurements, including reproducibility of the measures or values of dyssynchrony at baseline and after treatment.
Assessment of Pacing-induced Dyssynchrony
Mechanical dyssynchrony caused by RV pacing has been extensively covered in a recent review, but will be briefly discussed here.50 A series of animal studies by Prinzen and colleagues has demonstrated deleterious effects of RV pacing, including alterations in regional blood flow, fiber shortening, wall thickness, and impaired regional and global LV function.9–11 Human studies have shown similar detrimental effects of RV pacing in subjects with normal and reduced LV systolic function.51–53 The Dual Chamber And VVI Implantable Defibrillator (DAVID) trial was a randomized study assessing dual-chamber pacing versus ventricular back-up pacing in HF patients with EF ≤40% receiving a defibrillator.54 The AV delay in the paced group was set fairly long in most patients, resulting in nearly 60% of the ventricular beats in this group being RV-paced. The paced group had an increased incidence of the primary end-point: time to death or first hospitalization for HF.
Approximately 180,000 patients a year in the US receive pacemakers. Many of these patients are paced for treatment of high-grade AV block. In these patients, RV pacing is mandatory in light of their poor or absent electrical conduction from the atria to the ventricles. A significant but uncertain percentage of these patients develop pacing-induced dyssynchrony, LV systolic dysfunction, and/or HF.55–58 These patients remodel in a unique fashion, with less LV dilatation and a less spherical ventricle despite similar reductions in EF compared with patients with HF from other causes.59 In addition, these patients develop a type of dyssynchrony that is unique. Pacing from the RV apex results in an LBBB pattern on ECG and a very abnormal LV contraction pattern.
Figure 2 shows TT longitudinal displacement curves for eight segments from base to apex within the inter-ventricular septum in a normal subject and in a patient with RV apical pacing before and after CRT. In the normal subject all septal myocardial segments move together and all displacement curves demonstrate motion directed from the base toward the apex or chest wall. In the RV-paced patient the distal six walls (which are closest to the pacing lead) demonstrate abnormal motion away from the chest wall and toward the base of the LV. Many of the walls peak after aortic valve closure. Following upgrade to CRT, there is an improvement in both septal regional dyssynchrony and wall motion. None of the walls moves in the wrong direction and the average movement of the septal wall segments during systole has increased from 1.6 to 4.9mm. We have called this type of dyssynchrony intramural longitudinal dyssynchrony as it characterizes dysynchronous longitudinal motion within a single wall of the LV, in contrast to most measures of intra-ventricular dyssynchrony, which compare motion between walls of the LV (e.g. lateral versus septal). This methodology for assessing intramural dyssynchrony has been helpful clinically in managing patients with RV pacing who develop decreased LV systolic function or HF and are being considered for possible upgrade to CRT.
Left Ventricular Lead Placement
Electrical activation patterns of the LV in HF patients are highly variable. It has been hypothesized that CRT will be most effective when the LV lead is placed in the last activated myocardium. Ansalone et al. demonstrated that the last wall activated in HF patients prior to receiving CRT was the lateral wall in 35%, the anterior wall in 26%, the posterior wall in 23%, and the inferior/septal wall in 16%.60 In this retrospective study, the best clinical response was noted in those patients with an LV lead positioned in or near the latest activated wall. Becker et al. also found (using 3D echocardiography) that the patients with optimal lead location responded better to CRT, with greater improvements at one year in EF, peak myocardial oxygen consumption on cardiopulmonary exercise testing, and LV volumes.61 Despite these data, it is generally accepted that the ‘sweet spot’ for LV lead location encompasses a fairly large area of the lateral or postero-lateral wall of the LV.62 Information based on measurement of mechanical dyssynchrony could potentially be used to guide LV lead location. However, this approach is challenging, since lead location is often limited by venous anatomy and in many patients there are only a few technically feasible lead locations from which to choose.
Determination of Response to Cardiac Resynchronization Therapy
A potential clinical use of mechanical dyssynchrony measurements is to determine whether a patient has resynchronized following placement of a CRT device. Bleeker et al. evaluated 100 consecutive patients who met standard criteria for CRT and had mechanical longitudinal LV dyssynchrony at baseline.63 LV dyssynchrony improved immediately after CRT in responders (>10% reduction in LVESV), but did not change significantly in non-responders. Patients who did not acutely reduce the amount of dyssynchrony by ≥20% did not respond significantly to CRT. In the PROMISE-CRT study, acute (one week post- CRT) improvement in radial dyssynchrony as measured by STE was the single best predictor of six-month reduction in LVESV and explained 72% of the variance in individual remodeling response to CRT.29 These data support the argument that improvement of mechanical dyssynchrony is an important mechanism of CRT benefit. In patients who do not respond to CRT, assessment of mechanical dyssynchrony can assist in management. Occasionally, LV lead position can be revised or patients can undergo epicardial lead placement. More frequently, changes can be made in pacemaker settings to optimize ventricular diastolic function, systolic function, or mechanical dyssynchrony.
Optimization of Pacemaker Settings
Programmed variables that relate to mechanical dyssynchrony and may have important implications regarding cardiac function include atrial pacing lower rate, AV delay, ventricular-to-ventricular (VV) timing, and pacing lead configuration. Left-sided AV dyssynchrony can be readily assessed by echo Doppler measurement of mitral inflow velocities. Ideal AV synchrony results in a long diastolic filling period, minimal fusion of mitral inflow E and A waves, and no truncation of the A wave. While a number of methodologies have been proposed and used to optimize AV delay, an iterative method based on visualization of mitral inflow velocities is relatively simple and straightforward.
Optimization of VV timing is much more controversial. Some studies have shown hemodynamic improvements with optimization of VV timing, but clarifying the methodology and defining the magnitude of improvement and long-term benefits of these approaches still require further study.64,65
Pacing lead configuration is another parameter that can be optimized post-CRT. Some studies have shown that LV-only pacing produces a short- or mid-term hemodynamic benefit similar to that of biventricular pacing.66,67 In the Device Evaluation of CONTAK RENEWAL® 2 and EASYTRAK® 2 Assessment of Safety and Effectiveness in Heart Failure (DECREASE-HF) study, biventricular pacing tended to have a greater effect on LV remodeling than LV-only pacing, although patients in both groups improved.68 However, a physiological approach to patients with HF, where LV-only pacing is used only in selected patients with demonstrated acute hemodynamic benefit, has not been studied.
We have retrospectively reviewed data from 50 consecutive patients who received a CRT device and were either non-responders or weak responders to therapy.69 In about two-thirds of these patients we were able to make a change in pacemaker programming that resulted in an acute hemodynamic benefit as measured by echo. Measurements of systolic function, diastolic function, and mechanical dyssynchrony acutely and significantly improved with the changes made. The pacemaker changes varied among the patients but included, either alone or in combination, changes in atrial lower rate limit, AV delay, VV timing, or pacing configuration (e.g. turning off the RV lead and pacing only from the LV). In nine of these patients, turning off the RV lead and pacing only from the LV resulted in acutely improved hemodynamics as measured by echo. Figure 3 shows TT TDI curves from an HF patient with severe LV systolic dysfunction who did not improve with CRT. The septal (shown in the figure) and inferior wall (not shown) demonstrated markedly abnormal motion, with movement away from the echo transducer initially and peak movement in the middle of diastole. These walls are the closest to the RV apical pacing lead. Turning off the RV lead and pacing only from the LV resulted in return of septal and inferior wall motion toward normal and improvement in dyssynchrony, as shown in panel B, with LVEF increasing acutely from 35 to 45% in this patient.
Mechanical Dyssynchrony and Clinical Strategies in Heart Failure and Cardiac Resynchronization Therapy
The use of mechanical dyssynchrony echo measurements to assist in the management of patients with HF and CRT is evolving. Figure 4 depicts an algorithm utilized in our clinic to help guide decisions related to device therapy in systolic HF patients. The top third of the algorithm briefly summarizes a standard medical approach to systolic HF. The cause of HF is determined, reversible causes are addressed, and risk factors for coronary disease and other associated problems are treated. Patients are educated about HF and titrated to maximal tolerated doses of standard medications such as beta-blockers, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, aldosterone receptor antagonists, and digoxin with diuretics used as needed to control volume status. Approximately three months after medical therapy is maximized, a repeat echo is performed and mechanical dyssynchrony is measured. If the EF is >35%, in most cases the patient does not need a device. If the EF is ≤35%, the patient is likely a candidate for a defibrillator with or without a CRT device. The presence or absence of mechanical dyssynchrony, in conjunction with other clinical and laboratory data, including QRS duration, is then used to help determine whether a patient is likely to benefit from CRT. If a patient receives a CRT device, measurement of mechanical dyssynchrony by echo can help in determining response to CRT. In addition, knowledge of the type and extent of mechanical dyssynchrony in a given patient can help in optimizing pacemaker settings for that patient. This approach is favored by our clinic as an individualized and physiological strategy in which mechanical dyssynchrony measurements provide valuable information for managing systolic HF patients.
Summary
CRT is now standard therapy for patients with advanced HF and wide QRS. Indications for this therapy are likely to expand in light of recent studies showing benefit in patients with less severe HF.70,71 Studies utilizing echo measurements of mechanical dyssynchrony have been instrumental in better understanding the mechanism of CRT. Echocardiographic measurement of mechanical dyssynchrony is currently used at multiple institutions to help make important clinical decisions in patients with HF and CRT, such as selection of patients for CRT, determination of response to CRT, and optimizing pacemaker programming. An understanding of the various echo methodologies available for measuring mechanical dyssynchrony and the relative strengths and weaknesses of each is important for individual cardiologists, programs, and institutions in determining how to apply this technology to their patients. Improvements in standardization of techniques, reductions in time and complexity of measurements, and further research into the role of these mechanical dyssynchrony measurements in clinical care are under way and should lead to reduced costs and better quality of care for patients with HF.