Background
Risk factors for ventricular arrhythmias after cardiac resynchronization defibrillator therapy (CRT-D) for severely symptomatic heart failure are of clinical importance but are not clearly defined. The objective of this study was to test the hypothesis that mechanical dyssynchrony after CRT-D is a risk factor for ventricular arrhythmias.
Methods
A total of 266 consecutive CRT-D patients with class III or IV heart failure, QRS duration ≥120 msec, and ejection fractions ≤ 35% were prospectively studied. Dyssynchrony was assessed before and 6 months after CRT-D using speckle-tracking radial strain anteroseptal–to–posterior wall delay, predefined as ≥130 msec. Ventricular arrhythmias were predefined as appropriate antitachycardia pacing or shock, and the combined end point of ventricular arrhythmias, death, transplantation, or left ventricular assist device implantation was followed over 2 years.
Results
Of the initial 266 patients, 11 died, five underwent transplantation, three received left ventricular assist devices before their 6-month echocardiographic examinations, 19 (7%) had inadequate speckle-tracking at 6-month follow-up, and 27 (10%) were lost to follow-up. Accordingly, the study group consisted of 201 patients. Dyssynchrony after CRT-D was observed in 79 (39%) and was associated with a significantly higher ventricular arrhythmic event rate: 21% ( P < .001) with persistent dyssynchrony and 35% ( P < .001) with new dyssynchrony, compared with 8% with no dyssynchrony after CRT-D. The combined end point of ventricular arrhythmias, death, transplantation, or left ventricular assist device implantation was significantly associated with dyssynchrony after CRT-D (hazard ratio, 2.53; 95% confidence interval, 1.49–4.28; P = .001). Dyssynchrony after CRT-D was associated with ventricular arrhythmias or death in patient subgroups by cardiomyopathy type, QRS width, and morphology ( P < .05 for all).
Conclusions
Persistent or new radial dyssynchrony after CRT-D in severely symptomatic patients with heart failure with widened QRS complexes and reduced ejection fractions was associated with an increased rate of ventricular arrhythmias or death and appears to be a marker for a less favorable prognosis.
Cardiac resynchronization defibrillator therapy (CRT-D) is beneficial to many patients with severely symptomatic heart failure (HF) with wide QRS complexes and low ejection fractions (EFs) receiving optimal pharmacological therapy. Although the positive effects on cardiac function and remodeling are well known, reported effects of CRT-D on ventricular arrhythmias have been conflicting. Some previous studies have reported a protective effect of CRT-D on ventricular arrhythmias, whereas others have reported concerns of proarrhythmic effects. Baseline mechanical dyssynchrony has been shown to be a marker for more favorable outcomes in routine candidates for CRT-D with QRS widening. Specifically, patients with HF with widened QRS complexes who lack baseline dyssynchrony have less favorable survival after CRT-D. More recently, CRT-D was associated with increased mortality in severely symptomatic patients with HF with narrow QRS complexes selected by echocardiographic dyssynchrony. Factors such as scar and/or ischemia may play a role in the genesis of ventricular arrhythmias in patients with coronary disease, independent of the effects of CRT-D. Furthermore, suboptimal left ventricular (LV) positioning or the relationship of pacing lead positioning to scar may be associated with worsening of dyssynchrony and be a marker for worse outcome. The relationships of mechanical dyssynchrony before or after CRT-D to ventricular arrhythmic events appear complex and less clear. Accordingly, our objective was to test the hypothesis that persistent or worsened mechanical dyssynchrony as assessed by speckle-tracking radial strain after CRT-D is associated with increased risk for ventricular arrhythmic events and death.
Methods
Patient Population and Device Implantation
We prospectively studied 266 consecutive patients with severely symptomatic New York Heart Association (NYHA) class III or IV HF with LV EFs ≤ 35% and QRS durations ≥ 120 msec receiving optimal pharmacologic therapy, with baseline echocardiography, who underwent successful CRT-D device implantation. We prospectively included all cardiac resynchronization therapy (CRT) patients in the study with the intention of collecting ventricular arrhythmia and other outcome data. The protocol was approved by the institutional review board on biomedical research, and all patients gave informed consent consistent with the protocol. Prespecified patient subgroups were by etiology of HF, with ischemic cardiomyopathy defined as the presence of significant coronary artery disease (70% stenosis in one major coronary artery by coronary angiography) and/or a documented history of myocardial infarction or prior revascularization. All other etiologies of HF were defined as nonischemic cardiomyopathy. Subgroup analysis by QRS width < 150 or ≥ 150 msec and QRS morphology left bundle branch block (LBBB) or non-LBBB were also pre-defined on the basis of recently updated guidelines. All patients received CRT-D devices with right atrial leads, right ventricular apical leads, and LV pacing leads implanted through the coronary sinus in an epicardial vein, targeting the posterolateral or lateral LV free wall. Patients received CRT with device programming including rate cutoffs by the implanting physicians as deemed appropriate for clinical circumstances. Outcome events were defined over 2 years, with the primary end point pre-defined as ventricular arrhythmias requiring appropriate therapy, including appropriate antitachycardia pacing and appropriate shock obtained from CRT-D device interrogation at 3-month intervals after implantation. At the end of follow-up, the device interrogation records for every patient were retrospectively cross-checked for arrhythmic events adjudicated by an electrophysiologist blinded to all other study data. For patients who had both antitachycardia pacing and shock, the first event was used as the end point. The combined end point of ventricular arrhythmic event and death was also assessed, recording all-cause mortality because this study had no means to adjudicate cause of death. We combined the hard end points of cardiac transplantation or LV assist device (LVAD) implantation with deaths because only end-stage patients with limited anticipated survival would undergo these procedures at our institution.
Echocardiography
Echocardiographic studies were performed with GE Vivid 7 system (GE Vingmed Ultrasound AS, Horton, Norway) before and approximately 6 months after CRT-D device implantation. LV end-diastolic volume, end-systolic volume (ESV), and EF were assessed using the biplane Simpson’s rule. All quantitative analyses were performed on digitally stored images (EchoPAC BTO8-BTO12; GE Vingmed Ultrasound AS). Digital grayscale cine loops from three consecutive beats were obtained from standard mid-LV short-axis views at depths of 12 to 20 cm. Speckle-tracking radial dyssynchrony was evaluated with frame rates of 60 to 90 Hz. Dyssynchrony was evaluated by anteroseptal–to–posterior wall delay, as previously described. Briefly, regions of interest were manually placed on endocardial and epicardial borders, with time to peak strain evaluated from the onset of the QRS complex to the peak of the strain curve in six different segments from the mid left ventricle ( Figure 1 ). Significant radial dyssynchrony before or after CRT-D was defined as a time difference of ≥130 msec between the anterior septum and posterior wall peak strain.
Statistical Analyses
Continuous data are presented as mean ± SD. Proportions were compared using χ 2 tests. Comparisons of means were analyzed using unpaired t tests for independent variables and paired t tests for dependent variables (SPSS version 20; IBM, Armonk, NY). Kaplan-Meier survival analyses with follow-up censored at 2 years were performed for patients with mechanical dyssynchrony ≥ 130 and < 130 msec. The log-rank test was used to detect differences in event rate between groups. Cox regression analysis was performed to identify predictors of arrhythmic events. The multivariate regression was performed by including significant variables from the univariate models ( P < .05). Agreement and variability of radial strain dyssynchrony were assessed using Cohen’s κ and the coefficient of variation (MedCalc version 12.1; MedCalc Software, Ostend, Belgium). P values < .05 were considered significant.
Results
Association of Dyssynchrony before CRT with Ventricular Arrhythmias and Outcomes
There were 266 consecutive patients with HF aged 65 ± 11 years with successfully implanted CRT-D devices; 26% were women, with a mean EF of 25 ± 6%, a mean QRS duration of 159 ± 27 msec, and a mean NYHA class of 3.1 ± 0.5. Ventricular arrhythmias occurred in 39 of 266 patients (15%) during 2 years of follow-up: 24 with antitachycardia pacing and 15 with appropriate shock therapy. Baseline characteristics before CRT-D for those with arrhythmic events did not differ from those without arrhythmic events regarding age, gender, ischemic disease, NYHA class, QRS duration, EF, or baseline dyssynchrony ( Table 1 ). Defibrillator use for secondary prevention was higher in those with arrhythmic events (12%) than those without arrhythmias (1%), as expected, and LV volumes were greater in those with arrhythmias. There were 57 hard end points (21%) in 2 years. When hard end points were combined with arrhythmic events, there were 96 events in 85 unique patients, and the combined end point of death or ventricular arrhythmia was therefore reached in 85 patients (32%). Patients who died or had ventricular arrhythmias compared with those with arrhythmia-free survival were more frequently men ( P = .043), were in higher NYHA classes ( P = .004), had shorter QRS widths ( P = .006), and had less baseline dyssynchrony ( P = .025), with a trend toward more frequent ischemic disease ( P = .097). Furthermore, they had higher baseline end-diastolic volumes and ESVs and lower EFs ( P = .002, P = .001, and P = .017, respectively).
| Patients without arrhythmic events ( n = 227) | Patients with arrhythmic events ( n = 39) | P | |
|---|---|---|---|
| Age (y) | 65 ± 12 | 66 ± 9 | .76 |
| Men | 165 (73%) | 33 (85%) | .12 |
| NYHA class | 3.09 ± 0.55 | 3.08 ± 0.54 | .89 |
| Ischemic disease | 133 (59%) | 26 (67%) | .34 |
| QRS duration (msec) | 160 ± 28 | 155 ± 26 | .29 |
| β-blocker therapy | 193 (85%) | 35 (90%) | .59 |
| ACE inhibitor/ARB therapy | 194 (85%) | 32 (82%) | .76 |
| Baseline echocardiography ( n = 266) | |||
| LV EDV (mL) | 188 ± 66 | 224 ± 90 | .003 |
| LV ESV (mL) | 143 ± 58 | 174 ± 77 | .001 |
| LV EF (%) | 25 ± 6 | 23 ± 6 | .21 |
| Baseline dyssynchrony ∗ (msec) | 221 ± 121 | 209 ± 137 | .54 |
| Follow-up echocardiography ( n = 201) | n = 172 | n = 29 | |
| LV EDV (mL) | 161 ± 59 | 195 ± 81 | .01 |
| LV ESV (mL) | 109 ± 54 | 147 ± 59 | <.01 |
| LV EF (%) | 35 ± 11 | 27 ± 9 | <.01 |
| Δ LV EF (%) | 9 ± 9 | 4 ± 7 | <.01 |
| Follow-up dyssynchrony ∗ (msec) | 111 ± 93 | 188 ± 125 | <.01 |
| Δ Dyssynchrony ∗ (msec) | −111 ± 44 | −12 ± 192 | <.01 |
Associations of Dyssynchrony after CRT with Ventricular Arrhythmias and Outcomes
Of the initial 266 consecutive CRT-D patients, 19 (7%) had technically inadequate echocardiograms for speckle-tracking analysis at 6-month follow-up. There were 11 deaths, five transplantations, and three LVAD placements before the 6-month echocardiographic examinations, and 27 patients (10%) were unavailable because they underwent follow-up echocardiography elsewhere. Accordingly, the study group with paired baseline and 6-month echocardiography for dyssynchrony analysis consisted of 201 patients. In this group, ventricular arrhythmias requiring appropriate antitachycardia pacing or shock occurred in 29 of 201 (14%) during 2 years after CRT-D ( Table 1 ). Baseline characteristics in the 201 patients with follow-up echocardiography did not differ in significance between patients with and without arrhythmic events from the total population. Kaplan-Meier analyses showed a significantly higher incidence of arrhythmic events in patients with mechanical dyssynchrony after CRT-D. Dyssynchrony after CRT-D, either persistent or newly acquired, was a strong predictor of events (hazard ratio [HR], 3.76; 95% confidence interval [CI], 1.71–8.27; P = .001; Table 2 ). When adjusted for frequency of secondary prevention and changes in both EF and ESV, radial dyssynchrony after CRT-D remained significantly associated with ventricular arrhythmias ( P = .012). Furthermore, dyssynchrony after CRT-D remained associated with arrhythmias after correcting for ESV (HR, 2.87; 95% CI, 1.16–7.08; P = .02) and correcting for EF (HR, 2.91; 95% CI, 1.21–6.94; P = .02) 6 months after CRT-D. Subsequent to their 6-month follow-up echocardiographic examinations in this group of 201 patients, there were 25 deaths, seven transplantations, and four LVAD implantations. The combined end point as primary event of ventricular arrhythmia, death, transplantation, or LVAD implantation was reached in 57 of these patients (28%). Mechanical dyssynchrony observed after CRT was significantly associated with this combined end point of ventricular arrhythmias or death (HR, 2.53; 95% CI, 1.49–4.28; P = .001; Figure 2 ) and remained significantly associated after separately adjusting for changes in EF and ESV ( P = .005). The frequency of ventricular arrhythmias overall was similar in patients with nonischemic disease (32 of 80 [40%]) and in those with ischemic disease (47 of 121 [39%]). When examining subgroup analysis, patients with nonischemic etiology and no dyssynchrony had the most favorable prognosis. Dyssynchrony after CRT-D was also associated with the combined end point of ventricular arrhythmias, death, transplantation, or LVAD placement in both ischemic and nonischemic cardiomyopathy groups ( Figure 3 ). When grouped by QRS width and morphology consistent with current guidelines, dyssynchrony after CRT remained significantly associated with ventricular arrhythmic events in patients with QRS duration 120 to 149 msec (HR, 4.65; 95% CI, 1.61–13.42; P = .004), QRS duration ≥ 150 msec (HR, 3.43; 95% CI, 1.05–11.15; P = .04), LBBB (HR, 4.20; 95% CI, 1.37–14.06; P = .012), and non-LBBB (HR, 3.17; 95% CI, 1.08–9.29; P = .035).
| Parameters | Univariate | Multivariate | ||
|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | P | |
| Age (per year) | 0.99 (0.96–1.02) | .64 | ||
| Male gender | 3.17 (0.96–10.48) | .06 | ||
| Ischemic vs nonischemic disease | 1.27 (0.60–2.73) | .32 | ||
| ESV relative change after CRT-D | 7.31 (1.15–46.36) | .04 | 0.62 (0.05–8.53) | .73 |
| LV EF before CRT-D | 0.96 (0.91–1.01) | .15 | ||
| LV EF change after CRT-D | 0.91 (0.85–0.97) | <.01 | 0.91 (0.85–0.99) | .03 |
| Dyssynchrony before CRT-D | 0.54 (0.26–1.13) | .101 | ||
| Dyssynchrony after CRT-D | 3.76 (1.71–8.26) | .001 | 2.97 (1.25–7.07) | .01 |
