- Email: [email protected]
Ventricular premature depolarization QRS duration as a new marker of risk for the development of ventricular premature depolarization–induced cardiomyopathy
Ventricular premature depolarization QRS duration as a new marker of risk for the development of ventricular premature depolarization–induced cardiomyopathy Lidia Carballeira Pol, MD,* Marc W. Deyell, MD, MSc, FHRS,* David S. Frankel, MD, FHRS,* Daniel Benhayon, MD,* Fabien Squara, MD,*† William Chik, MD,* Maria Kohari, MD,* Rajat Deo, MD,* Francis E. Marchlinski, MD, FHRS* From the *Electrophysiology Section, Cardiovascular Division, Department of Medicine, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, and †Cardiology Department, Pasteur University Hospital, Nice, France. BACKGROUND Frequent ventricular premature depolarizations (VPDs) can cause cardiomyopathy (CMP). The mechanisms underlying its development remain unclear, with VPD burden being only a weak predictor of risk. OBJECTIVE To determine whether VPD QRS duration at the time of initial presentation could predict risk for the subsequent development of CMP in patients with normal left ventricular ejection fraction (LVEF). METHODS From consecutive patients referred for ablation between January 1, 2006, and April 2, 2013, with Z10% VPDs on 24-hour Holter monitoring, we identified 45 patients with normal LVEF and an electrocardiogram of the targeted VPD, who were then followed for at least 6 months (median 14 months; interquartile range [IQR] 8–32 months) before intervention. We excluded patients with structural or genetic heart disease. RESULTS Of the 45 patients, 28 (62%) maintained normal LVEF and 17(38%) developed VPD-induced CMP. VPD burden was similar (26.5% [IQR 19.3%–39.5%] vs 26.0% [IQR 16.4%–41.0%]; P ¼ 0.4) between the 2 groups. Patients who developed VPD-induced CMP had significantly longer VPD QRS duration (159 ms vs 142 ms; P o .001) and a
Introduction Ventricular premature depolarizations (VPDs) are usually regarded as benign in the absence of structural heart disease. However, a subset of patients with frequent VPDs develop a cardiomyopathy (CMP) that usually reverses after successful This work was supported in part by the Mark S. Marchlinski Research Fund at the University of Pennsylvania. Dr Carballeira Pol has received a research grant from the Sociedad Española de Cardiología. Dr Squara is supported by a research grant from the Fédération Française de Cardiologie. Address reprint requests and correspondence: Dr Francis E. Marchlinski, Electrophysiology Section, Cardiovascular Division, Department of Medicine, Hospital of the University of Pennsylvania, 9 Founders Pavilion – Cardiology, 3400 Spruce St, Philadelphia, PA 19104. E-mail address: francis. [email protected].
1547-5271/$-see front matter B 2014 Heart Rhythm Society. All rights reserved.
longer sinus QRS duration (97 ms vs 89 ms; P ¼ .04). A VPD QRS duration of Z153 ms best predicted development of VPD CMP (82% sensitivity and 75% specificity). Longer VPD QRS duration and a non– outflow tract site of VPD origin were independent risk factors for left ventricular dysfunction after multivariate analysis. CONCLUSION VPD QRS duration longer than 153 ms and a non– outflow tract site of origin might be useful predictors of the subsequent development of VPD-induced CMP. KEYWORDS Ventricular premature depolarizations; Cardiomyopathy; Ventricular outflow tract; Ventricular arrhythmia; Electrocardiogram ABBREVIATIONS CMP ¼ cardiomyopathy; ECG ¼ electrocardiogram; EF ¼ ejection fraction; IQR ¼ interquartile range; LV ¼ left ventricle/ventricular; LVEDD ¼ left ventricular end-diastolic diameter; LVEF ¼ left ventricular ejection fraction; MRI ¼ magnetic resonance imaging; VPD ¼ ventricular premature depolarization (Heart Rhythm 2014;11:299–306) I 2014 Heart Rhythm Society. All rights reserved.
ablation.1–6 The spectrum of dysfunction that can occur with frequent VPDs is wide, ranging from severe CMP to more frequent and subtle forms of biventricular impairment.7 Currently, the mechanisms underlying the development of VPD-induced CMP remain unclear. Although it seems that a critical VPD burden is required, VPD burden remains a weak predictor of the development of CMP.2,3,8–11 A longer VPD QRS duration has been associated with the presence of CMP in previous cross-sectional studies.11,12 We have previously shown that shorter VPD QRS duration is associated with the recovery of left ventricular (LV) function after successful ablation, but the temporal relationship of VPD QRS duration to CMP is unknown.13 We therefore sought to determine whether VPD QRS duration measured before the onset of http://dx.doi.org/10.1016/j.hrthm.2013.10.055
300 CMP would be a useful predictor of the subsequent development of CMP.
Methods Study population We identified all patients who underwent VPD ablation at the Hospital of the University of Pennsylvania between January 1, 2006, and April 2, 2013. As previous data suggested that a significant VPD burden is required for the development of VPD-induced CMP, we restricted our study to patients with Z10% VPDs on preprocedural 24-hour Holter monitoring.9 We further identified for inclusion those patients with an electrocardiogram (ECG) of the targeted VPD obtained at least 6 months before ablation. In addition, the echocardiographic documentation of normal left ventricular ejection fraction (LVEF) at the time of initial VPD measurement as well as subsequent echocardiographic monitoring of LV function before or immediately after ablation were required. We excluded the following patients: (1) those with structural or genetic heart disease and (2) those with a history of sustained ventricular tachycardia or sudden cardiac death.
Data collection Baseline demographic and clinical characteristics were collected through a detailed review of the electronic medical records from the beginning of follow-up in the outpatient arrhythmia center. For patients not followed at the University of Pennsylvania, the referring cardiologist was contacted and medical records reviewed. The presence of structural or inherited heart disease was excluded by clinical history, physical examination, echocardiographic and electrocardiographic assessment, and genetic testing when appropriate. In selected patients with coronary risk factors or symptoms, the presence of significant ischemic heart disease was excluded by stress test or coronary angiography. All electrocardiographic measurements were performed initially by using digital calipers by a blinded investigator. An averaged VPD coupling interval was calculated by measuring all the coupling intervals on the initial ECG. VPD QRS durations were reassessed by the same and an additional blinded independent observer in 20 randomly selected patients. In addition, the electrocardiographic measurements were repeated at the ablation procedure recordings by using digital calipers at 100 mm/s on Cardiolab (version 6.5.4.1858, GE medical Systems, Waukesha, WI). We averaged the QRS duration on 5 separate VPDs before the introduction of the catheters. In the case of multiple VPD morphologies, the dominant VPD was measured or, if 41 dominant VPD was present, the measurements were averaged between morphologies. Transthoracic echocardiograms were formally interpreted by echocardiographers. When a range of LVEF was reported, the lower value was analyzed. Furthermore, the validity of echocardiographic data was reassessed in 20 randomly selected studies by 2 independent investigators.
Heart Rhythm, Vol 11, No 2, February 2014 Informed consent was obtained from all patients undergoing ablation in accordance with the institutional guidelines of the University of Pennsylvania Health System. Acute procedural success was defined as the absence of spontaneous or inducible VPDs after isoproterenol infusion and burst pacing from the right ventricular apex for at least 30 minutes after ablation. Partial acute procedural success was defined as 80% reduction in VPD frequency or persistence of infrequent nondominant VPDs from a nontargeted morphology. Long-term success was defined as Z80% reduction in VPD burden in follow-up.14
Patient classification VPD-induced CMP was defined as decrease in LVEF to o50% with Z10% decrease from baseline on echocardiogram recorded before or immediately after ablation. A Z10% LVEF reduction was chosen to account for interobserver variability of noncontrast echocardiography (reported to range from 9% to 14%).15 The site of VPD origin was defined during ablation as the location of earliest activation and/or best pace map. VPD sites of origin were categorized as follows: (1) septal RV outflow tract/pulmonary artery/right coronary cusp, (2) RV outflow tract free wall, (3) para-Hisian, (4) left coronary cusp/left-right coronary cusp junction/aortomitral continuity/anterior interventricular vein, (5) other LV sites (great cardiac vein and anterolateral basal LV sites), (6) papillary muscle, and (7) multiple sites of origin. For logistic regression analysis, sites of origin were further dichotomized as follows: (1) LV vs RV, (2) midline (septal locations, coronary cusps, LV apex, posteromedial papillary muscle, and basal anterior LV) vs non–midline, and (3) outflow tract vs non–outflow tract. On the basis of the response of the CMP to the VPD suppression (either with ablation or antiarrhythmic drugs if ablation was unsuccessful), the CMP was classified as reversible (LVEF improvement of Z10% for a final LVEF Z50%) or irreversible (LVEF improvement of o10%).
Statistical Analyses Continuous variables were expressed as mean ⫾ SD and for nonparametric variables as median with interquartile range (IQR). Categorical variables were expressed as percentages. The Student t test and the Mann-Whitney test were used to compare normally and nonnormally distributed continuous variables, respectively. Categorical variables were compared by using the Fisher exact test or the χ2 test. Pearson and Spearman product-moment correlation coefficients were calculated to evaluate intra- and interobserver variability and linear relationships. To determine independent predictors of the development of VPD-induced CMP, multivariate logistic regression analysis was performed. Covariates were assessed initially in a univariate fashion. Those with a P value of r.1 were entered into the multivariate model. All tests were 2-tailed, and a P value of r.05 was considered as statistically significant. Analyses
Carballeira Pol et al
VPD QRS Duration as a Marker of Risk for Cardiomyopathy
301
Figure 1 Flowchart of patient selection for the study. ECG ¼ electrocardiogram; LVEF ¼ left ventricle ejection fraction; VF ¼ ventricular fibrillation; VPD ¼ ventricular premature depolarization; VT ¼ ventricular tachycardia.
were performed by using SPSS version 16 (SPSS Inc, Chicago, IL).
Results During the study period, a total of 605 VPD ablations were performed on 540 patients. One hundred twenty (22%) patients were excluded owing to the presence of structural or inherited heart disease (Figure 1). From the remainder, only 110 (20%) patients had a documented VPD burden of Z10% VPDs on preprocedural 24-hour Holter monitoring along with a baseline echocardiogram and ECG at least 6 months before ablation. Forty-five (8%) of these patients had a normal baseline LVEF and are the focus of this investigation. Seventeen subsequently developed VPD-induced CMP before ablation (group 1) and 28 continued to have preserved LVEF (group 2).
Baseline patient and VPD characteristics The baseline characteristics of the normal LVEF group and the group who developed CMP are summarized in Table 1. We found no significant differences in cardiovascular risk factors, use of medication, echocardiographic data, daily VPD burden, or prevalence of nonsustained ventricular tachycardia between groups. Patients that remained with normal LVEF were more likely to present with palpitations (71% vs 41% in the VPD-induced CMP group; P ¼ .04), and those who developed VPD-induced CMP had a higher prevalence of atypical symptoms. Eight patients in the normal LVEF group and 7 patients in the group who developed CMP had a magnetic resonance imaging (MRI) performed; minor abnormalities in the form of regional wall motion defects or delayed gadolinium enhancement were found in 2 patients in each group.
Figure 2 shows the distribution of the sites of origin across both groups. Non–outflow tract sites of origin were more common among patients who developed CMP (53% vs 11%; P ¼ .005). Most common non–outflow tract sites of origin were the anterolateral basal LV and great cardiac vein. There was a nonsignificant trend toward more LV sites of origin in patients who developed VPD-induced CMP (82% vs 58%; P ¼ .09). Compared with the patients that remained with normal LVEF, those who developed VPD-induced CMP had a significantly longer baseline conducted sinus QRS duration (97.0 ms vs 88.6 ms; P ¼ .04) and VPD QRS duration (159.1 ms vs 142.3 ms; P o .001). Shortest mean VPD QRS duration was among VPDs originating from the RV outflow tract septal sites/pulmonary artery/right coronary cusp (142 ms), followed by the LV outflow tract sites (148 ms), RV outflow tract free wall sites (154 ms), and the other LV sites (160 ms). We did not find any difference in the VPD coupling interval between groups. We found a weak linear correlation between left ventricular end-diastolic diameter (LVEDD) and native QRS duration (r ¼ .37; P ¼ .03), but not between LVEDD and VPD QRS duration (r ¼ .43; P ¼ .57). A P-wave notch was identified after the VPD QRS in 19 patients in the normal ejection fraction (EF) group and in 14 patients in the VPD-induced CMP group.16 We did not find any significant differences in the VPD P-wave notch interval between groups (200 ms [IQR 178–207 ms] vs 200 ms [IQR 161–212 ms]; P ¼ .9).
Changes in echocardiographic and electrocardiographic parameters during follow-up before ablation The median time from the initial presentation to ablation was 14 months (IQR 8–32 months), with no differences between
302 Table 1
Heart Rhythm, Vol 11, No 2, February 2014 Baseline and electrocardiographic characteristics of patients stratified by development of VPD-induced CMP P
Characteristic
Normal LVEF
Developed CMP
n Baseline characteristics Sex: female (%) Age (y) Hypertension (%) Diabetes (%) Smoking (%) Atrial fibrillation (%) LVEF (%) LVEDD (mm) Symptoms Asymptomatic Palpitations/skipped beats Atypical symptoms* Symptom duration (mo) Medication use β-Blockers (%) Calcium channel blockers (%) ACEI or ARB (%) Antiarrhythmic drugs (%) Holter monitoring % VPD 24 h VPDs per 24 h NSVT (45 beats) (%) Cardiac MRI Performed (%) Abnormal (%) VPD classification† (%) Left ventricular site of origin only Midline site of origin only Non–outflow tract site of origin only Electrocardiographic characteristics VPD QRS duration (ms) Conducted QRS duration (ms) Heart rate (beats/min) VPD coupling interval (ms) Retrograde P-wave visible
28
17
67.0 46.0 ⫾ 14.9 35.0 14.0 25.0 4.0 59.1 ⫾ 5.1 49.0 ⫾ 7.0
50.0 53.0 ⫾ 16.5 29.0 5.0 47.0 17.0 56.2 ⫾ 4.9 53.0 ⫾ 5.0
.3 .2 .8 .6 .1 .1 .1 .08
18.0 71.0 11.0 48.0 (30–99)
24.0 41.0 35.0 24 (13–63)
.6 .04 .04 .2
29.0 14.0 29.0 14.0
41.0 12.0 29.0 12.0
26.5 (19.3–39.5) 29,015 (20,450–40,089) 46.0
26.0 (16.4–41.0) 27,477(14,910–41,000) 53.0
28.5 25.0
41.0 28.5
58.0 82.0 11.0
82.0 82.0 53.0
142.3 ⫾ 12.6 88.6 ⫾ 10.5 70.0 ⫾ 3.0 479.7 ⫾ 63.0 67.0
159.1 ⫾ 13.9 97.0 ⫾ 16.0 75.0 ⫾ 3.0 459.7 ⫾ 57.0 82.0
.4 .8 1.0 1.0 .4 .4 .7 .4 1.0 .09 1 .005 o.001 .04 .4 .3 .3
Values are presented as mean ⫾ SD and as median (interquartile range). P values are for comparison between VPD-induced CMP and normal LVEF group. ACEI ¼ angiotensin-converting enzyme inhibitor; ARB ¼ angiotensin receptor blocker; CMP ¼ cardiomyopathy; LVEDD ¼ left ventricular end-diastolic diameter; LVEF ¼ left ventricular ejection fraction; MRI ¼ magnetic resonance imaging; VPD ¼ ventricular premature depolarization. * Includes fatigue, chest pain, and shortness of breath or dizziness. †Excluded patients with multiple sites of origin.
Figure 2 Anatomical distribution of VPD sites of origin across both groups, as determined at ablation sites. AIV ¼ anterior interventricular vein; AMC ¼ aortomitral continuity; CC ¼ coronary cusp; GCV ¼ great cardiac vein; LV ¼ left ventricle; LVEF ¼ left ventricle ejection fraction; PA ¼ pulmonary artery; RVOT ¼ right ventricular outflow tract; VPD ¼ ventricular premature depolarization.
Carballeira Pol et al Table 2
VPD QRS Duration as a Marker of Risk for Cardiomyopathy
303
Univariate predictors of the development of VPD-induced cardiomyopathy
Predictor Baseline characteristics Sex: female Age (per 1-y increase) Hypertension Diabetes Smoking Symptoms Symptom duration (per mo) Atrial fibrillation (%) Medication use Baseline β-blockers Baseline ACEI/ARB Antiarrhythmic drugs (%) Holter monitoring % VPD 24 h (per 1% increase) NSVT (45 beats) Electrocardiographic characteristic VPD QRS duration (per 10-ms increase) Conducted QRS duration (per 10-ms increase) Heart rate (per 1-beat/min increase) VPD coupling interval Retrograde P wave VPD site of origin Left ventricular (vs RV) Non–midline (vs midline) Non–outflow tract (vs outflow tract)
P
Odds ratio
95% CI
0.53 1.03 0.75 0.38 2.67 0.71 0.99 5.78
0.15–1.84 0.98–1.07 0.20–2.75 0.03–3.67 0.74–9.59 0.16–3.10 0.98–1.00 0.55–60.87
.32 .19 .66 .39 .13 .64 .44 .14
1.70 1.04 0.80
0.49–6.02 0.27–3.93 0.13–4.91
.39 .95 .81
1.01 1.30
0.94–1.04 0.39–4.34
.76 .67
2.67 1.67 1.03 0.99 2.20
1.4–4.03 0.95–2.75 0.98–1.06 0.98–1 0.5–9.6
.002 .07 .22 .29 .29
3.40 0.99 9.00
0.78–14.88 0.21–4.92 1.96–50
.10 1.0 .012
ACEI ¼ angiotensin-converting enzyme inhibitor; ARB ¼ angiotensin receptor blocker; CI ¼ confidence interval; NSVT ¼ nonsustained ventricular tachycardia; RV ¼ right ventricle; VPD ¼ ventricular premature depolarization.
the 2 groups. In patients who developed VPD-induced CMP, the LVEF decreased from 56% ⫾ 5% at baseline to 38% ⫾ 6% (P o .0001), with no significant difference in the degree of LV dilatation (LVEDD: 53 ⫾ 5 mm vs 54 ⫾ 4 mm; P ¼ .4). In the group that remained with a normal EF, there were no changes in LVEF or LV diameter (LVEF: 59% ⫾ 5% to 59% ⫾ 7%; P ¼ .5; LVEDD: 49 ⫾ 7 mm vs 48 ⫾ 6 mm; P ¼ .4). The mean VPD QRS duration did not change significantly from baseline to ablation in either of the groups (from 142 ⫾ 13 to 141 ⫾ 13 ms for the group with preserved LVEF, P ¼ .7 and from 159 ⫾ 13 to 160 ⫾ 12 ms, P ¼ .64 for the group who developed VPD-induced CMP). In addition, the mean sinus rhythm QRS duration remained constant over time (from 89 ⫾ 10 to 88 ⫾ 14 ms, P ¼ .9 for the normal EF group and from 97 ⫾ 16 to 91 ⫾ 16 ms, P ¼ .9 for the VPD-induced CMP group). There was no significant difference in the VPD coupling interval at baseline or at the time of ablation between groups (from 459 ⫾ 57 to 470 ⫾ 36 ms, P ¼ .3 for patients with normal EF and from 476 ⫾ 65 to 465 ⫾ 53 ms, P ¼ .5 for patients with VPD-induced CMP).
VPD site of origin were independently associated with the development of LV dysfunction. The odds ratio for the development of VPD-induced CMP was 2.94 (95% CI 1.36– 6.55) for every 10-ms increase in the baseline VPD QRS duration. The odds ratio for non–outflow tract location was 14 (95% CI 1.55–126.84). The area under the receiver operating characteristic curve for the final model was 0.82.
Characteristics associated with the development of LV dysfunction In univariate analysis (Table 2), longer VPD QRS duration, longer conducted QRS duration, LV site of origin, and non– outflow tract site of origin were associated with an increased risk of development of VPD-induced CMP. In multivariate analysis, only VPD QRS duration and non–outflow tract
Figure 3 Scatter plot illustrating VPD durations on baseline electrocardiogram for the group that continued to have a normal ejection fraction during the observation period and the group who developed VPD-induced cardiomyopathy. The 153-ms cutoff favored the development of cardiomyopathy (see text). VPD ¼ ventricular premature depolarization.
304
Heart Rhythm, Vol 11, No 2, February 2014
Figure 4 Electrocardiogram showing baseline VPD QRS duration from 2 patients with a high VPD burden originating from the left coronary cusp: LVEF remained normal for patient A. Patient B developed VPD-induced cardiomyopathy (LVEF 40%) that subsequently reversed after long-term VPD suppression. VPD is wider in the patient who developed cardiomyopathy. LVEF ¼ left ventricle ejection fraction; VPD ¼ ventricular premature depolarization.
Baseline VPD QRS duration for the 2 groups is displayed in Figure 3. Despite the heterogeneous distribution of sites of origin among the 2 populations, we found longer VPD durations for similar anatomical sites of origin among the group who developed VPD-induced CMP (Figure 4). The best cutoff for the development of VPD-induced CMP was a VPD QRS duration of 153 ms (82% sensitivity and 75% specificity; Figure 5). Most (93%) of the patients who
maintained a normal EF had a baseline VPD QRS duration of o158 ms, and most of those who developed VPDinduced CMP (94%) had a baseline QRS duration of 4138 ms. For non–outflow tract locations, a VPD QRS duration cutoff of 147 ms demonstrated a sensitivity of 89% and a specificity of 100% (area under the curve ¼ 0.96; P ¼ .02) for the development of CMP, whereas for outflow tract locations, the best cutoff was 153 ms (88% sensitivity and 71% specificity; area under the curve ¼ 0.78; P ¼ .02).
Effect of VPD burden on the development of LV dysfunction The statistically significant association between VPD QRS width and CMP development was not lessened by adjustment for VPD burden (P ¼ .002 before and after adjustment for VPD burden). Furthermore, the R2 value of the statistical model did not improve with the addition of VPD burden (R2 ¼ .38 with and without VPD burden). Thus, in our population, the effects of VPD QRS width and VPD burden were not additive.
Ablation outcomes and follow-up
Figure 5 Receiver operating characteristic curve of the VPD QRS duration. The best diagnostic accuracy cutoff for the development of VPD-induced cardiomyopathy was a VPD QRS duration of 153 ms (82% sensitivity and 75% specificity). AUC ¼ area under the curve; VPD ¼ ventricular premature depolarization.
Twelve (41%) patients in the normal LVEF group and 6 (35%) patients in the VPD-induced CMP group had undergone at least 1 prior failed ablation procedure (P ¼ NS). The overall acute success rate of the index ablation was 80%, with an additional 4% of the patients achieving partial procedural success. There was no significant difference between the 2 groups in terms of total procedure time (normal LVEF group: 240 ⫾ 97 minutes; VPD-induced CMP group: 261 ⫾ 94 minutes; P ¼ .493) and total fluoroscopy time (normal LVEF
Carballeira Pol et al
VPD QRS Duration as a Marker of Risk for Cardiomyopathy
group: 39 ⫾ 27 minutes; VPD-induced CMP group: 46 ⫾ 31 minutes; P ¼ .448). There was only 1 periprocedural complication in the form of left femoral artery pseudoaneurysm managed percutaneously. In 26 patients (58%), a followup 24-hour Holter monitoring was available, with 96% achieving long-term success (Z80% reduction in the initial VPD burden) and 61% having r1% residual VPD burden. In the patients who developed VPD-induced CMP, echocardiographic and Holter follow-up was available after ablation in 14 of 17 patients. During a mean follow-up period of 7.3 ⫾ 7 months from ablation, all achieved Z80% reduction of the initial VPD burden, 3 of them with adjunctive antiarrhythmic drug treatment. The CMP was reversible in 93%, with only 1 patient not experiencing a recovery in his LVEF by 410% after 7.4 months of follow-up.
Intra- and interobserver variability in the measurements The intra- and interobserver correlation coefficients for VPD QRS duration were .95 and .85, respectively. The interobserver correlation coefficient for the reported echocardiographic measurements was .86 and .88 for LVEDD and LVEF, respectively.
Discussion This longitudinal study is the first to evaluate systematically the risk factors implicated in the development of VPDinduced CMP rather than to associate risk factors in a crosssectional fashion. In our population, with a preset inclusion threshold of Z10% daily VPD, the arrhythmia burden was not associated with the development of LV dysfunction. This contrasts with the observations of previous cross-sectional studies.8,11 For example, Baman et al9 found that 424% daily VPD burden demonstrated a sensitivity of 79% and a specificity of 78% for the presence of impaired LVEF. Importantly, their study had no minimum threshold VPD burden for inclusion. The lack of association of VPD burden with CMP development in our population is explained both by the low sensitivity and specificity of this parameter in the range of VPD burdens included in our study population (median 26% VPD burden for both groups) and by the wide variability in VPD burdens on repeat Holter monitoring. Interestingly, in this relatively young population, other traditional cardiovascular risk factors such as hypertension, diabetes, and older age were also not associated with the development of CMP. The fact that traditional risk factors were not associated with the development of CMP emphasizes that the mechanism of CMP was likely VPD induced rather than via the development of coronary artery disease, valvular heart disease, or hypertensive heart disease. Although there was a significant difference in sinus QRS and VPD QRS durations between the 2 groups, the difference in VPD QRS duration was more significant and remained independently associated with the development of VPDinduced in multivariate modeling, even after adjusting for VPD site of origin. We believe that VPD QRS width, more
305
than the conducted QRS complex during sinus rhythm, better reflects the subtle underlying myocardial substrate abnormality that might predispose to LV dysfunction with arrhythmia stress. By assessing myocardial cell-to-cell conduction rather than using the specialized conduction system, VPD QRS duration may be a more accurate measurement of the degree of cell-to-cell electrical uncoupling that precedes the development of mechanical CMP. This finding is consistent with previous observations showing that the VPD QRS duration at the time of ablation was associated with the reversibility of VPD-induced CMP after successful ablation.13 It is possible that remodeling caused by a repetitive abnormal electrical and mechanical activation pattern of the ventricles in the context of a subtly altered substrate, suggested by the wider VPD QRS duration, could potentiate clinical CMP. A complete characterization of the microscopic substrate that predisposes to an overt CMP is still lacking, as available imaging techniques lack sufficient resolution in the early stages of CMP. It has been shown that even with a normal LVEF, quite often patients with a high VPD burden of significant duration have some degree of mechanical impairment, which is detectable only by more accurate imaging methods, such as 2D speckle tracking strain imaging.7 Nevertheless, new tools for the characterization of diffuse and subtle myocardial fibrosis are evolving, such as MRI with T1 mapping, that might help detect earlier preclinical changes.17,18 Whether the substrate abnormality at early stages is caused exclusively by VPDs or another inherited or acquired predisposing factor with VPD unmasking the background abnormality still needs to be elucidated. In contrast to arrhythmia stress acting on a subtle underlying substrate abnormality, it is also possible that the arrhythmia burden coupled with specific VPD sites of origin may predispose to dysfunction even in the absence of structural abnormalities. Recently, Huizar et al19 described a canine model of VPD-induced CMP where they found no relevant histopathologic abnormalities. VPDs can be hemodynamically deleterious. A short diastolic ventricular filling time and an asynchronous systolic beat with lower mechanical efficiency can lead to wedge pulmonary pressure augmentation.16 This same dyssynchrony could also explain why the risk for the development of CMP appears greater with non–outflow tract sites of origin. Lateral and free wall LV locations will cause a greater amount of dyssynchrony in biventricular contraction compared with VPDs in the outflow tract that arise closer to the ventricular septum. This has been demonstrated in MRI-tagged studies of ventricular activation during LV pacing in dogs.20,21 Ventricular dyssynchrony results in compromised global cardiac mechanical efficiency, asymmetrically increased wall thickness, and work overload in the late activated regions, altered myocardial blood flow, and local changes in myocardial protein expression; all these mechanisms could contribute to the pathogenesis of VPD-induced CMP.22 The current findings related to the importance of the VPD site of origin are consistent with our observations in a previous cross-sectional study, in which VPDs originating from epicardial sites were associated with VPD-induced CMP.12
306 The results of our study appear to have important clinical implications. Given the high efficacy and low risk of VPD ablation, the risk-benefit ratio may favor ablation in asymptomatic patients with normal LVEF who are at increased risk for the development of CMP on the basis of sufficient VPD burden, VPD QRS duration, and site of origin. Of course, this approach would need to be validated by additional observational and prospective randomized trials before being routinely recommended.
Study limitations Limiting the analysis to patients known to have a normal LVEF for at least 6 months before ablation resulted in a relatively small sample size. In addition, there was a higher proportion of left-sided VPDs owing to referral bias to our tertiary care center, which might limit extrapolation of our results to the overall population of patients with high VPD burden. Unfortunately, we will not know whether any of the patients who continued to have a normal EF would have developed LV dysfunction with a longer time of exposure to the arrhythmia burden, as previous observations showed that there might be a decline in LV function only after a prolonged period (4–8 years).2 It is also worth noting that only patients with high VPD burden were included in our study (lowest burden 11,000 per 24 hours). It is quite possible that if patients with lower VPD burdens had been included, VPD burden would have become a significant predictor of the development of CMP, in addition to VPD QRS width. Finally, only a limited number of patients underwent cardiac MRI in our study, limiting the power to detect significant MRI predictors of the development of CMP. Our effort must be considered a hypothesis-generating study based on a highly selected group of patients. Further studies with systematic prospective follow-up of patients with frequent VPDs and normal LV function should be performed to confirm our observations. Additional observational and prospective randomized trials are needed before routinely recommending prophylactic VPD ablation on the basis of VPD QRS width or site of frequent VPD origin.
Conclusions This study sheds light on the temporal association between VPD widening and development of CMP. We found that VPD QRS duration longer than 153 ms and non–outflow tract sites of VPD origin were associated with greater risk of developing VPD-induced CMP, above and beyond VPD burden. This is the first prospective study of risk factors for the development of VPD-induced CMP. By improving our ability to define risk of development of CMP, more informed decisions regarding VPD ablation can be made.
Heart Rhythm, Vol 11, No 2, February 2014
References 1. Lee V, Hemingway H, Harb R, Crake T, Lambiase P. The prognostic significance of premature ventricular complexes in adults without clinically apparent heart disease: a meta-analysis and systematic review. Heart 2012;98:1290–1298. 2. Niwano S, Wakisaka Y, Niwano H, et al. Prognostic significance of frequent premature ventricular contractions originating from the ventricular outflow tract in patients with normal left ventricular function. Heart 2009;95:1230–1237. 3. Hasdemir C, Ulucan C, Yavuzgil O, et al. Tachycardia-induced cardiomyopathy in patients with idiopathic ventricular arrhythmias: the incidence, clinical and electrophysiologic characteristics, and the predictors. J Cardiovasc Electrophysiol 2011;22:663–668. 4. Sekiguchi Y, Aonuma K, Yamauchi Y, et al. Chronic hemodynamic effects after radiofrequency catheter ablation of frequent monomorphic ventricular premature beats. J Cardiovasc Electrophysiol 2005;16:1057–1063. 5. Yarlagadda RK, Iwai S, Stein KM, et al. Reversal of cardiomyopathy in patients with repetitive monomorphic ventricular ectopy originating from the right ventricular outflow tract. Circulation 2005;112:1092–1097. 6. Takemoto M, Yoshimura H, Ohba Y, et al. Radiofrequency catheter ablation of premature ventricular complexes from right ventricular outflow tract improves left ventricular dilation and clinical status in patients without structural heart disease. J Am Coll Cardiol 2005;45:1259–1265. 7. Wijnmaalen AP, Delgado V, Schalij MJ, et al. Beneficial effects of catheter ablation on left ventricular and right ventricular function in patients with frequent premature ventricular contractions and preserved ejection fraction. Heart 2010;96:1275–1280. 8. Bogun F, Crawford T, Reich S, et al. Radiofrequency ablation of frequent, idiopathic premature ventricular complexes: comparison with a control group without intervention. Heart Rhythm 2007;4:863–867. 9. Baman TS, Lange DC, Ilg KJ, et al. Relationship between burden of premature ventricular complexes and left ventricular function. Heart Rhythm 2010;7: 865–869. 10. Kanei Y, Friedman M, Ogawa N, Hanon S, Lam P, Schweitzer P. Frequent premature ventricular complexes originating from the right ventricular outflow tract are associated with left ventricular dysfunction. Ann Noninvasive Electrocardiol 2008;13:81–85. 11. Del Carpio Munoz F, Syed FF, Noheria A, et al. Characteristics of premature ventricular complexes as correlates of reduced left ventricular systolic function: study of the burden, duration, coupling interval, morphology and site of origin of PVCs. J Cardiovasc Electrophysiol 2011;22:791–798. 12. Yokokawa M, Kim HM, Good E, et al. Impact of QRS duration of frequent premature ventricular complexes on the development of cardiomyopathy. Heart Rhythm 2012;9:1460–1464. 13. Deyell MW, Park K-M, Han Y, et al. Predictors of recovery of left ventricular dysfunction after ablation of frequent ventricular premature depolarizations. Heart Rhythm 2012;9:1465–1472. 14. Mountantonakis SE, Frankel DS, Gerstenfeld EP, et al. Reversal of outflow tract ventricular premature depolarization–induced cardiomyopathy with ablation: effect of residual arrhythmia burden and preexisting cardiomyopathy on outcome. Heart Rhythm 2011;8:1608–1614. 15. Malm S, Frigstad S, Sagberg E, Larsson H, Skjaerpe T. Accurate and reproducible measurement of left ventricular volume and ejection fraction by contrast echocardiography: a comparison with magnetic resonance imaging. J Am Coll Cardiol 2004;44:1030–1035. 16. Kuroki K, Tada H, Seo Y, et al. Prediction and mechanism of frequent ventricular premature contractions related to haemodynamic deterioration. Eur J Heart Fail 2012;14:1112–1120. 17. Han Y, Peters DC, Dokhan B, Manning WJ. Shorter difference between myocardium and blood optimal inversion time suggests diffuse fibrosis in dilated cardiomyopathy. J Magn Reson Imaging 2009;30:967–972. 18. Sibley CT, Noureldin RA, Mudd JO, et al. T1 mapping in cardiomyopathy at cardiac MR: comparison with endomyocardial biopsy. Radiology 2012;265:724–732. 19. Huizar JF, Kaszala K, Potfay J, et al. Left ventricular systolic dysfunction induced by ventricular ectopy: a novel model for premature ventricular contractioninduced cardiomyopathy. Circ Arrhythm Electrophysiol 2011;4:543–549. 20. Wyman BT, Hunter WC, Prinzen FW, et al. Mapping propagation of mechanical activation in the paced heart with MRI tagging. Am J Physiol 1999;276:H881–H891. 21. Prinzen FW, Peschar M. Relation between the pacing induced sequence of activation and left ventricular pump. Pacing Clin Electrophysiol 2002;25:484–498. 22. Spragg DD, Kass DA. Pathobiology of left ventricular dyssynchrony and resynchronization. Prog Cardiovasc Dis 2006;49:26–41.
Introduction Ventricular premature depolarizations (VPDs) are usually regarded as benign in the absence of structural heart disease. However, a subset of patients with frequent VPDs develop a cardiomyopathy (CMP) that usually reverses after successful This work was supported in part by the Mark S. Marchlinski Research Fund at the University of Pennsylvania. Dr Carballeira Pol has received a research grant from the Sociedad Española de Cardiología. Dr Squara is supported by a research grant from the Fédération Française de Cardiologie. Address reprint requests and correspondence: Dr Francis E. Marchlinski, Electrophysiology Section, Cardiovascular Division, Department of Medicine, Hospital of the University of Pennsylvania, 9 Founders Pavilion – Cardiology, 3400 Spruce St, Philadelphia, PA 19104. E-mail address: francis. [email protected].
1547-5271/$-see front matter B 2014 Heart Rhythm Society. All rights reserved.
longer sinus QRS duration (97 ms vs 89 ms; P ¼ .04). A VPD QRS duration of Z153 ms best predicted development of VPD CMP (82% sensitivity and 75% specificity). Longer VPD QRS duration and a non– outflow tract site of VPD origin were independent risk factors for left ventricular dysfunction after multivariate analysis. CONCLUSION VPD QRS duration longer than 153 ms and a non– outflow tract site of origin might be useful predictors of the subsequent development of VPD-induced CMP. KEYWORDS Ventricular premature depolarizations; Cardiomyopathy; Ventricular outflow tract; Ventricular arrhythmia; Electrocardiogram ABBREVIATIONS CMP ¼ cardiomyopathy; ECG ¼ electrocardiogram; EF ¼ ejection fraction; IQR ¼ interquartile range; LV ¼ left ventricle/ventricular; LVEDD ¼ left ventricular end-diastolic diameter; LVEF ¼ left ventricular ejection fraction; MRI ¼ magnetic resonance imaging; VPD ¼ ventricular premature depolarization (Heart Rhythm 2014;11:299–306) I 2014 Heart Rhythm Society. All rights reserved.
ablation.1–6 The spectrum of dysfunction that can occur with frequent VPDs is wide, ranging from severe CMP to more frequent and subtle forms of biventricular impairment.7 Currently, the mechanisms underlying the development of VPD-induced CMP remain unclear. Although it seems that a critical VPD burden is required, VPD burden remains a weak predictor of the development of CMP.2,3,8–11 A longer VPD QRS duration has been associated with the presence of CMP in previous cross-sectional studies.11,12 We have previously shown that shorter VPD QRS duration is associated with the recovery of left ventricular (LV) function after successful ablation, but the temporal relationship of VPD QRS duration to CMP is unknown.13 We therefore sought to determine whether VPD QRS duration measured before the onset of http://dx.doi.org/10.1016/j.hrthm.2013.10.055
300 CMP would be a useful predictor of the subsequent development of CMP.
Methods Study population We identified all patients who underwent VPD ablation at the Hospital of the University of Pennsylvania between January 1, 2006, and April 2, 2013. As previous data suggested that a significant VPD burden is required for the development of VPD-induced CMP, we restricted our study to patients with Z10% VPDs on preprocedural 24-hour Holter monitoring.9 We further identified for inclusion those patients with an electrocardiogram (ECG) of the targeted VPD obtained at least 6 months before ablation. In addition, the echocardiographic documentation of normal left ventricular ejection fraction (LVEF) at the time of initial VPD measurement as well as subsequent echocardiographic monitoring of LV function before or immediately after ablation were required. We excluded the following patients: (1) those with structural or genetic heart disease and (2) those with a history of sustained ventricular tachycardia or sudden cardiac death.
Data collection Baseline demographic and clinical characteristics were collected through a detailed review of the electronic medical records from the beginning of follow-up in the outpatient arrhythmia center. For patients not followed at the University of Pennsylvania, the referring cardiologist was contacted and medical records reviewed. The presence of structural or inherited heart disease was excluded by clinical history, physical examination, echocardiographic and electrocardiographic assessment, and genetic testing when appropriate. In selected patients with coronary risk factors or symptoms, the presence of significant ischemic heart disease was excluded by stress test or coronary angiography. All electrocardiographic measurements were performed initially by using digital calipers by a blinded investigator. An averaged VPD coupling interval was calculated by measuring all the coupling intervals on the initial ECG. VPD QRS durations were reassessed by the same and an additional blinded independent observer in 20 randomly selected patients. In addition, the electrocardiographic measurements were repeated at the ablation procedure recordings by using digital calipers at 100 mm/s on Cardiolab (version 6.5.4.1858, GE medical Systems, Waukesha, WI). We averaged the QRS duration on 5 separate VPDs before the introduction of the catheters. In the case of multiple VPD morphologies, the dominant VPD was measured or, if 41 dominant VPD was present, the measurements were averaged between morphologies. Transthoracic echocardiograms were formally interpreted by echocardiographers. When a range of LVEF was reported, the lower value was analyzed. Furthermore, the validity of echocardiographic data was reassessed in 20 randomly selected studies by 2 independent investigators.
Heart Rhythm, Vol 11, No 2, February 2014 Informed consent was obtained from all patients undergoing ablation in accordance with the institutional guidelines of the University of Pennsylvania Health System. Acute procedural success was defined as the absence of spontaneous or inducible VPDs after isoproterenol infusion and burst pacing from the right ventricular apex for at least 30 minutes after ablation. Partial acute procedural success was defined as 80% reduction in VPD frequency or persistence of infrequent nondominant VPDs from a nontargeted morphology. Long-term success was defined as Z80% reduction in VPD burden in follow-up.14
Patient classification VPD-induced CMP was defined as decrease in LVEF to o50% with Z10% decrease from baseline on echocardiogram recorded before or immediately after ablation. A Z10% LVEF reduction was chosen to account for interobserver variability of noncontrast echocardiography (reported to range from 9% to 14%).15 The site of VPD origin was defined during ablation as the location of earliest activation and/or best pace map. VPD sites of origin were categorized as follows: (1) septal RV outflow tract/pulmonary artery/right coronary cusp, (2) RV outflow tract free wall, (3) para-Hisian, (4) left coronary cusp/left-right coronary cusp junction/aortomitral continuity/anterior interventricular vein, (5) other LV sites (great cardiac vein and anterolateral basal LV sites), (6) papillary muscle, and (7) multiple sites of origin. For logistic regression analysis, sites of origin were further dichotomized as follows: (1) LV vs RV, (2) midline (septal locations, coronary cusps, LV apex, posteromedial papillary muscle, and basal anterior LV) vs non–midline, and (3) outflow tract vs non–outflow tract. On the basis of the response of the CMP to the VPD suppression (either with ablation or antiarrhythmic drugs if ablation was unsuccessful), the CMP was classified as reversible (LVEF improvement of Z10% for a final LVEF Z50%) or irreversible (LVEF improvement of o10%).
Statistical Analyses Continuous variables were expressed as mean ⫾ SD and for nonparametric variables as median with interquartile range (IQR). Categorical variables were expressed as percentages. The Student t test and the Mann-Whitney test were used to compare normally and nonnormally distributed continuous variables, respectively. Categorical variables were compared by using the Fisher exact test or the χ2 test. Pearson and Spearman product-moment correlation coefficients were calculated to evaluate intra- and interobserver variability and linear relationships. To determine independent predictors of the development of VPD-induced CMP, multivariate logistic regression analysis was performed. Covariates were assessed initially in a univariate fashion. Those with a P value of r.1 were entered into the multivariate model. All tests were 2-tailed, and a P value of r.05 was considered as statistically significant. Analyses
Carballeira Pol et al
VPD QRS Duration as a Marker of Risk for Cardiomyopathy
301
Figure 1 Flowchart of patient selection for the study. ECG ¼ electrocardiogram; LVEF ¼ left ventricle ejection fraction; VF ¼ ventricular fibrillation; VPD ¼ ventricular premature depolarization; VT ¼ ventricular tachycardia.
were performed by using SPSS version 16 (SPSS Inc, Chicago, IL).
Results During the study period, a total of 605 VPD ablations were performed on 540 patients. One hundred twenty (22%) patients were excluded owing to the presence of structural or inherited heart disease (Figure 1). From the remainder, only 110 (20%) patients had a documented VPD burden of Z10% VPDs on preprocedural 24-hour Holter monitoring along with a baseline echocardiogram and ECG at least 6 months before ablation. Forty-five (8%) of these patients had a normal baseline LVEF and are the focus of this investigation. Seventeen subsequently developed VPD-induced CMP before ablation (group 1) and 28 continued to have preserved LVEF (group 2).
Baseline patient and VPD characteristics The baseline characteristics of the normal LVEF group and the group who developed CMP are summarized in Table 1. We found no significant differences in cardiovascular risk factors, use of medication, echocardiographic data, daily VPD burden, or prevalence of nonsustained ventricular tachycardia between groups. Patients that remained with normal LVEF were more likely to present with palpitations (71% vs 41% in the VPD-induced CMP group; P ¼ .04), and those who developed VPD-induced CMP had a higher prevalence of atypical symptoms. Eight patients in the normal LVEF group and 7 patients in the group who developed CMP had a magnetic resonance imaging (MRI) performed; minor abnormalities in the form of regional wall motion defects or delayed gadolinium enhancement were found in 2 patients in each group.
Figure 2 shows the distribution of the sites of origin across both groups. Non–outflow tract sites of origin were more common among patients who developed CMP (53% vs 11%; P ¼ .005). Most common non–outflow tract sites of origin were the anterolateral basal LV and great cardiac vein. There was a nonsignificant trend toward more LV sites of origin in patients who developed VPD-induced CMP (82% vs 58%; P ¼ .09). Compared with the patients that remained with normal LVEF, those who developed VPD-induced CMP had a significantly longer baseline conducted sinus QRS duration (97.0 ms vs 88.6 ms; P ¼ .04) and VPD QRS duration (159.1 ms vs 142.3 ms; P o .001). Shortest mean VPD QRS duration was among VPDs originating from the RV outflow tract septal sites/pulmonary artery/right coronary cusp (142 ms), followed by the LV outflow tract sites (148 ms), RV outflow tract free wall sites (154 ms), and the other LV sites (160 ms). We did not find any difference in the VPD coupling interval between groups. We found a weak linear correlation between left ventricular end-diastolic diameter (LVEDD) and native QRS duration (r ¼ .37; P ¼ .03), but not between LVEDD and VPD QRS duration (r ¼ .43; P ¼ .57). A P-wave notch was identified after the VPD QRS in 19 patients in the normal ejection fraction (EF) group and in 14 patients in the VPD-induced CMP group.16 We did not find any significant differences in the VPD P-wave notch interval between groups (200 ms [IQR 178–207 ms] vs 200 ms [IQR 161–212 ms]; P ¼ .9).
Changes in echocardiographic and electrocardiographic parameters during follow-up before ablation The median time from the initial presentation to ablation was 14 months (IQR 8–32 months), with no differences between
302 Table 1
Heart Rhythm, Vol 11, No 2, February 2014 Baseline and electrocardiographic characteristics of patients stratified by development of VPD-induced CMP P
Characteristic
Normal LVEF
Developed CMP
n Baseline characteristics Sex: female (%) Age (y) Hypertension (%) Diabetes (%) Smoking (%) Atrial fibrillation (%) LVEF (%) LVEDD (mm) Symptoms Asymptomatic Palpitations/skipped beats Atypical symptoms* Symptom duration (mo) Medication use β-Blockers (%) Calcium channel blockers (%) ACEI or ARB (%) Antiarrhythmic drugs (%) Holter monitoring % VPD 24 h VPDs per 24 h NSVT (45 beats) (%) Cardiac MRI Performed (%) Abnormal (%) VPD classification† (%) Left ventricular site of origin only Midline site of origin only Non–outflow tract site of origin only Electrocardiographic characteristics VPD QRS duration (ms) Conducted QRS duration (ms) Heart rate (beats/min) VPD coupling interval (ms) Retrograde P-wave visible
28
17
67.0 46.0 ⫾ 14.9 35.0 14.0 25.0 4.0 59.1 ⫾ 5.1 49.0 ⫾ 7.0
50.0 53.0 ⫾ 16.5 29.0 5.0 47.0 17.0 56.2 ⫾ 4.9 53.0 ⫾ 5.0
.3 .2 .8 .6 .1 .1 .1 .08
18.0 71.0 11.0 48.0 (30–99)
24.0 41.0 35.0 24 (13–63)
.6 .04 .04 .2
29.0 14.0 29.0 14.0
41.0 12.0 29.0 12.0
26.5 (19.3–39.5) 29,015 (20,450–40,089) 46.0
26.0 (16.4–41.0) 27,477(14,910–41,000) 53.0
28.5 25.0
41.0 28.5
58.0 82.0 11.0
82.0 82.0 53.0
142.3 ⫾ 12.6 88.6 ⫾ 10.5 70.0 ⫾ 3.0 479.7 ⫾ 63.0 67.0
159.1 ⫾ 13.9 97.0 ⫾ 16.0 75.0 ⫾ 3.0 459.7 ⫾ 57.0 82.0
.4 .8 1.0 1.0 .4 .4 .7 .4 1.0 .09 1 .005 o.001 .04 .4 .3 .3
Values are presented as mean ⫾ SD and as median (interquartile range). P values are for comparison between VPD-induced CMP and normal LVEF group. ACEI ¼ angiotensin-converting enzyme inhibitor; ARB ¼ angiotensin receptor blocker; CMP ¼ cardiomyopathy; LVEDD ¼ left ventricular end-diastolic diameter; LVEF ¼ left ventricular ejection fraction; MRI ¼ magnetic resonance imaging; VPD ¼ ventricular premature depolarization. * Includes fatigue, chest pain, and shortness of breath or dizziness. †Excluded patients with multiple sites of origin.
Figure 2 Anatomical distribution of VPD sites of origin across both groups, as determined at ablation sites. AIV ¼ anterior interventricular vein; AMC ¼ aortomitral continuity; CC ¼ coronary cusp; GCV ¼ great cardiac vein; LV ¼ left ventricle; LVEF ¼ left ventricle ejection fraction; PA ¼ pulmonary artery; RVOT ¼ right ventricular outflow tract; VPD ¼ ventricular premature depolarization.
Carballeira Pol et al Table 2
VPD QRS Duration as a Marker of Risk for Cardiomyopathy
303
Univariate predictors of the development of VPD-induced cardiomyopathy
Predictor Baseline characteristics Sex: female Age (per 1-y increase) Hypertension Diabetes Smoking Symptoms Symptom duration (per mo) Atrial fibrillation (%) Medication use Baseline β-blockers Baseline ACEI/ARB Antiarrhythmic drugs (%) Holter monitoring % VPD 24 h (per 1% increase) NSVT (45 beats) Electrocardiographic characteristic VPD QRS duration (per 10-ms increase) Conducted QRS duration (per 10-ms increase) Heart rate (per 1-beat/min increase) VPD coupling interval Retrograde P wave VPD site of origin Left ventricular (vs RV) Non–midline (vs midline) Non–outflow tract (vs outflow tract)
P
Odds ratio
95% CI
0.53 1.03 0.75 0.38 2.67 0.71 0.99 5.78
0.15–1.84 0.98–1.07 0.20–2.75 0.03–3.67 0.74–9.59 0.16–3.10 0.98–1.00 0.55–60.87
.32 .19 .66 .39 .13 .64 .44 .14
1.70 1.04 0.80
0.49–6.02 0.27–3.93 0.13–4.91
.39 .95 .81
1.01 1.30
0.94–1.04 0.39–4.34
.76 .67
2.67 1.67 1.03 0.99 2.20
1.4–4.03 0.95–2.75 0.98–1.06 0.98–1 0.5–9.6
.002 .07 .22 .29 .29
3.40 0.99 9.00
0.78–14.88 0.21–4.92 1.96–50
.10 1.0 .012
ACEI ¼ angiotensin-converting enzyme inhibitor; ARB ¼ angiotensin receptor blocker; CI ¼ confidence interval; NSVT ¼ nonsustained ventricular tachycardia; RV ¼ right ventricle; VPD ¼ ventricular premature depolarization.
the 2 groups. In patients who developed VPD-induced CMP, the LVEF decreased from 56% ⫾ 5% at baseline to 38% ⫾ 6% (P o .0001), with no significant difference in the degree of LV dilatation (LVEDD: 53 ⫾ 5 mm vs 54 ⫾ 4 mm; P ¼ .4). In the group that remained with a normal EF, there were no changes in LVEF or LV diameter (LVEF: 59% ⫾ 5% to 59% ⫾ 7%; P ¼ .5; LVEDD: 49 ⫾ 7 mm vs 48 ⫾ 6 mm; P ¼ .4). The mean VPD QRS duration did not change significantly from baseline to ablation in either of the groups (from 142 ⫾ 13 to 141 ⫾ 13 ms for the group with preserved LVEF, P ¼ .7 and from 159 ⫾ 13 to 160 ⫾ 12 ms, P ¼ .64 for the group who developed VPD-induced CMP). In addition, the mean sinus rhythm QRS duration remained constant over time (from 89 ⫾ 10 to 88 ⫾ 14 ms, P ¼ .9 for the normal EF group and from 97 ⫾ 16 to 91 ⫾ 16 ms, P ¼ .9 for the VPD-induced CMP group). There was no significant difference in the VPD coupling interval at baseline or at the time of ablation between groups (from 459 ⫾ 57 to 470 ⫾ 36 ms, P ¼ .3 for patients with normal EF and from 476 ⫾ 65 to 465 ⫾ 53 ms, P ¼ .5 for patients with VPD-induced CMP).
VPD site of origin were independently associated with the development of LV dysfunction. The odds ratio for the development of VPD-induced CMP was 2.94 (95% CI 1.36– 6.55) for every 10-ms increase in the baseline VPD QRS duration. The odds ratio for non–outflow tract location was 14 (95% CI 1.55–126.84). The area under the receiver operating characteristic curve for the final model was 0.82.
Characteristics associated with the development of LV dysfunction In univariate analysis (Table 2), longer VPD QRS duration, longer conducted QRS duration, LV site of origin, and non– outflow tract site of origin were associated with an increased risk of development of VPD-induced CMP. In multivariate analysis, only VPD QRS duration and non–outflow tract
Figure 3 Scatter plot illustrating VPD durations on baseline electrocardiogram for the group that continued to have a normal ejection fraction during the observation period and the group who developed VPD-induced cardiomyopathy. The 153-ms cutoff favored the development of cardiomyopathy (see text). VPD ¼ ventricular premature depolarization.
304
Heart Rhythm, Vol 11, No 2, February 2014
Figure 4 Electrocardiogram showing baseline VPD QRS duration from 2 patients with a high VPD burden originating from the left coronary cusp: LVEF remained normal for patient A. Patient B developed VPD-induced cardiomyopathy (LVEF 40%) that subsequently reversed after long-term VPD suppression. VPD is wider in the patient who developed cardiomyopathy. LVEF ¼ left ventricle ejection fraction; VPD ¼ ventricular premature depolarization.
Baseline VPD QRS duration for the 2 groups is displayed in Figure 3. Despite the heterogeneous distribution of sites of origin among the 2 populations, we found longer VPD durations for similar anatomical sites of origin among the group who developed VPD-induced CMP (Figure 4). The best cutoff for the development of VPD-induced CMP was a VPD QRS duration of 153 ms (82% sensitivity and 75% specificity; Figure 5). Most (93%) of the patients who
maintained a normal EF had a baseline VPD QRS duration of o158 ms, and most of those who developed VPDinduced CMP (94%) had a baseline QRS duration of 4138 ms. For non–outflow tract locations, a VPD QRS duration cutoff of 147 ms demonstrated a sensitivity of 89% and a specificity of 100% (area under the curve ¼ 0.96; P ¼ .02) for the development of CMP, whereas for outflow tract locations, the best cutoff was 153 ms (88% sensitivity and 71% specificity; area under the curve ¼ 0.78; P ¼ .02).
Effect of VPD burden on the development of LV dysfunction The statistically significant association between VPD QRS width and CMP development was not lessened by adjustment for VPD burden (P ¼ .002 before and after adjustment for VPD burden). Furthermore, the R2 value of the statistical model did not improve with the addition of VPD burden (R2 ¼ .38 with and without VPD burden). Thus, in our population, the effects of VPD QRS width and VPD burden were not additive.
Ablation outcomes and follow-up
Figure 5 Receiver operating characteristic curve of the VPD QRS duration. The best diagnostic accuracy cutoff for the development of VPD-induced cardiomyopathy was a VPD QRS duration of 153 ms (82% sensitivity and 75% specificity). AUC ¼ area under the curve; VPD ¼ ventricular premature depolarization.
Twelve (41%) patients in the normal LVEF group and 6 (35%) patients in the VPD-induced CMP group had undergone at least 1 prior failed ablation procedure (P ¼ NS). The overall acute success rate of the index ablation was 80%, with an additional 4% of the patients achieving partial procedural success. There was no significant difference between the 2 groups in terms of total procedure time (normal LVEF group: 240 ⫾ 97 minutes; VPD-induced CMP group: 261 ⫾ 94 minutes; P ¼ .493) and total fluoroscopy time (normal LVEF
Carballeira Pol et al
VPD QRS Duration as a Marker of Risk for Cardiomyopathy
group: 39 ⫾ 27 minutes; VPD-induced CMP group: 46 ⫾ 31 minutes; P ¼ .448). There was only 1 periprocedural complication in the form of left femoral artery pseudoaneurysm managed percutaneously. In 26 patients (58%), a followup 24-hour Holter monitoring was available, with 96% achieving long-term success (Z80% reduction in the initial VPD burden) and 61% having r1% residual VPD burden. In the patients who developed VPD-induced CMP, echocardiographic and Holter follow-up was available after ablation in 14 of 17 patients. During a mean follow-up period of 7.3 ⫾ 7 months from ablation, all achieved Z80% reduction of the initial VPD burden, 3 of them with adjunctive antiarrhythmic drug treatment. The CMP was reversible in 93%, with only 1 patient not experiencing a recovery in his LVEF by 410% after 7.4 months of follow-up.
Intra- and interobserver variability in the measurements The intra- and interobserver correlation coefficients for VPD QRS duration were .95 and .85, respectively. The interobserver correlation coefficient for the reported echocardiographic measurements was .86 and .88 for LVEDD and LVEF, respectively.
Discussion This longitudinal study is the first to evaluate systematically the risk factors implicated in the development of VPDinduced CMP rather than to associate risk factors in a crosssectional fashion. In our population, with a preset inclusion threshold of Z10% daily VPD, the arrhythmia burden was not associated with the development of LV dysfunction. This contrasts with the observations of previous cross-sectional studies.8,11 For example, Baman et al9 found that 424% daily VPD burden demonstrated a sensitivity of 79% and a specificity of 78% for the presence of impaired LVEF. Importantly, their study had no minimum threshold VPD burden for inclusion. The lack of association of VPD burden with CMP development in our population is explained both by the low sensitivity and specificity of this parameter in the range of VPD burdens included in our study population (median 26% VPD burden for both groups) and by the wide variability in VPD burdens on repeat Holter monitoring. Interestingly, in this relatively young population, other traditional cardiovascular risk factors such as hypertension, diabetes, and older age were also not associated with the development of CMP. The fact that traditional risk factors were not associated with the development of CMP emphasizes that the mechanism of CMP was likely VPD induced rather than via the development of coronary artery disease, valvular heart disease, or hypertensive heart disease. Although there was a significant difference in sinus QRS and VPD QRS durations between the 2 groups, the difference in VPD QRS duration was more significant and remained independently associated with the development of VPDinduced in multivariate modeling, even after adjusting for VPD site of origin. We believe that VPD QRS width, more
305
than the conducted QRS complex during sinus rhythm, better reflects the subtle underlying myocardial substrate abnormality that might predispose to LV dysfunction with arrhythmia stress. By assessing myocardial cell-to-cell conduction rather than using the specialized conduction system, VPD QRS duration may be a more accurate measurement of the degree of cell-to-cell electrical uncoupling that precedes the development of mechanical CMP. This finding is consistent with previous observations showing that the VPD QRS duration at the time of ablation was associated with the reversibility of VPD-induced CMP after successful ablation.13 It is possible that remodeling caused by a repetitive abnormal electrical and mechanical activation pattern of the ventricles in the context of a subtly altered substrate, suggested by the wider VPD QRS duration, could potentiate clinical CMP. A complete characterization of the microscopic substrate that predisposes to an overt CMP is still lacking, as available imaging techniques lack sufficient resolution in the early stages of CMP. It has been shown that even with a normal LVEF, quite often patients with a high VPD burden of significant duration have some degree of mechanical impairment, which is detectable only by more accurate imaging methods, such as 2D speckle tracking strain imaging.7 Nevertheless, new tools for the characterization of diffuse and subtle myocardial fibrosis are evolving, such as MRI with T1 mapping, that might help detect earlier preclinical changes.17,18 Whether the substrate abnormality at early stages is caused exclusively by VPDs or another inherited or acquired predisposing factor with VPD unmasking the background abnormality still needs to be elucidated. In contrast to arrhythmia stress acting on a subtle underlying substrate abnormality, it is also possible that the arrhythmia burden coupled with specific VPD sites of origin may predispose to dysfunction even in the absence of structural abnormalities. Recently, Huizar et al19 described a canine model of VPD-induced CMP where they found no relevant histopathologic abnormalities. VPDs can be hemodynamically deleterious. A short diastolic ventricular filling time and an asynchronous systolic beat with lower mechanical efficiency can lead to wedge pulmonary pressure augmentation.16 This same dyssynchrony could also explain why the risk for the development of CMP appears greater with non–outflow tract sites of origin. Lateral and free wall LV locations will cause a greater amount of dyssynchrony in biventricular contraction compared with VPDs in the outflow tract that arise closer to the ventricular septum. This has been demonstrated in MRI-tagged studies of ventricular activation during LV pacing in dogs.20,21 Ventricular dyssynchrony results in compromised global cardiac mechanical efficiency, asymmetrically increased wall thickness, and work overload in the late activated regions, altered myocardial blood flow, and local changes in myocardial protein expression; all these mechanisms could contribute to the pathogenesis of VPD-induced CMP.22 The current findings related to the importance of the VPD site of origin are consistent with our observations in a previous cross-sectional study, in which VPDs originating from epicardial sites were associated with VPD-induced CMP.12
306 The results of our study appear to have important clinical implications. Given the high efficacy and low risk of VPD ablation, the risk-benefit ratio may favor ablation in asymptomatic patients with normal LVEF who are at increased risk for the development of CMP on the basis of sufficient VPD burden, VPD QRS duration, and site of origin. Of course, this approach would need to be validated by additional observational and prospective randomized trials before being routinely recommended.
Study limitations Limiting the analysis to patients known to have a normal LVEF for at least 6 months before ablation resulted in a relatively small sample size. In addition, there was a higher proportion of left-sided VPDs owing to referral bias to our tertiary care center, which might limit extrapolation of our results to the overall population of patients with high VPD burden. Unfortunately, we will not know whether any of the patients who continued to have a normal EF would have developed LV dysfunction with a longer time of exposure to the arrhythmia burden, as previous observations showed that there might be a decline in LV function only after a prolonged period (4–8 years).2 It is also worth noting that only patients with high VPD burden were included in our study (lowest burden 11,000 per 24 hours). It is quite possible that if patients with lower VPD burdens had been included, VPD burden would have become a significant predictor of the development of CMP, in addition to VPD QRS width. Finally, only a limited number of patients underwent cardiac MRI in our study, limiting the power to detect significant MRI predictors of the development of CMP. Our effort must be considered a hypothesis-generating study based on a highly selected group of patients. Further studies with systematic prospective follow-up of patients with frequent VPDs and normal LV function should be performed to confirm our observations. Additional observational and prospective randomized trials are needed before routinely recommending prophylactic VPD ablation on the basis of VPD QRS width or site of frequent VPD origin.
Conclusions This study sheds light on the temporal association between VPD widening and development of CMP. We found that VPD QRS duration longer than 153 ms and non–outflow tract sites of VPD origin were associated with greater risk of developing VPD-induced CMP, above and beyond VPD burden. This is the first prospective study of risk factors for the development of VPD-induced CMP. By improving our ability to define risk of development of CMP, more informed decisions regarding VPD ablation can be made.
Heart Rhythm, Vol 11, No 2, February 2014
References 1. Lee V, Hemingway H, Harb R, Crake T, Lambiase P. The prognostic significance of premature ventricular complexes in adults without clinically apparent heart disease: a meta-analysis and systematic review. Heart 2012;98:1290–1298. 2. Niwano S, Wakisaka Y, Niwano H, et al. Prognostic significance of frequent premature ventricular contractions originating from the ventricular outflow tract in patients with normal left ventricular function. Heart 2009;95:1230–1237. 3. Hasdemir C, Ulucan C, Yavuzgil O, et al. Tachycardia-induced cardiomyopathy in patients with idiopathic ventricular arrhythmias: the incidence, clinical and electrophysiologic characteristics, and the predictors. J Cardiovasc Electrophysiol 2011;22:663–668. 4. Sekiguchi Y, Aonuma K, Yamauchi Y, et al. Chronic hemodynamic effects after radiofrequency catheter ablation of frequent monomorphic ventricular premature beats. J Cardiovasc Electrophysiol 2005;16:1057–1063. 5. Yarlagadda RK, Iwai S, Stein KM, et al. Reversal of cardiomyopathy in patients with repetitive monomorphic ventricular ectopy originating from the right ventricular outflow tract. Circulation 2005;112:1092–1097. 6. Takemoto M, Yoshimura H, Ohba Y, et al. Radiofrequency catheter ablation of premature ventricular complexes from right ventricular outflow tract improves left ventricular dilation and clinical status in patients without structural heart disease. J Am Coll Cardiol 2005;45:1259–1265. 7. Wijnmaalen AP, Delgado V, Schalij MJ, et al. Beneficial effects of catheter ablation on left ventricular and right ventricular function in patients with frequent premature ventricular contractions and preserved ejection fraction. Heart 2010;96:1275–1280. 8. Bogun F, Crawford T, Reich S, et al. Radiofrequency ablation of frequent, idiopathic premature ventricular complexes: comparison with a control group without intervention. Heart Rhythm 2007;4:863–867. 9. Baman TS, Lange DC, Ilg KJ, et al. Relationship between burden of premature ventricular complexes and left ventricular function. Heart Rhythm 2010;7: 865–869. 10. Kanei Y, Friedman M, Ogawa N, Hanon S, Lam P, Schweitzer P. Frequent premature ventricular complexes originating from the right ventricular outflow tract are associated with left ventricular dysfunction. Ann Noninvasive Electrocardiol 2008;13:81–85. 11. Del Carpio Munoz F, Syed FF, Noheria A, et al. Characteristics of premature ventricular complexes as correlates of reduced left ventricular systolic function: study of the burden, duration, coupling interval, morphology and site of origin of PVCs. J Cardiovasc Electrophysiol 2011;22:791–798. 12. Yokokawa M, Kim HM, Good E, et al. Impact of QRS duration of frequent premature ventricular complexes on the development of cardiomyopathy. Heart Rhythm 2012;9:1460–1464. 13. Deyell MW, Park K-M, Han Y, et al. Predictors of recovery of left ventricular dysfunction after ablation of frequent ventricular premature depolarizations. Heart Rhythm 2012;9:1465–1472. 14. Mountantonakis SE, Frankel DS, Gerstenfeld EP, et al. Reversal of outflow tract ventricular premature depolarization–induced cardiomyopathy with ablation: effect of residual arrhythmia burden and preexisting cardiomyopathy on outcome. Heart Rhythm 2011;8:1608–1614. 15. Malm S, Frigstad S, Sagberg E, Larsson H, Skjaerpe T. Accurate and reproducible measurement of left ventricular volume and ejection fraction by contrast echocardiography: a comparison with magnetic resonance imaging. J Am Coll Cardiol 2004;44:1030–1035. 16. Kuroki K, Tada H, Seo Y, et al. Prediction and mechanism of frequent ventricular premature contractions related to haemodynamic deterioration. Eur J Heart Fail 2012;14:1112–1120. 17. Han Y, Peters DC, Dokhan B, Manning WJ. Shorter difference between myocardium and blood optimal inversion time suggests diffuse fibrosis in dilated cardiomyopathy. J Magn Reson Imaging 2009;30:967–972. 18. Sibley CT, Noureldin RA, Mudd JO, et al. T1 mapping in cardiomyopathy at cardiac MR: comparison with endomyocardial biopsy. Radiology 2012;265:724–732. 19. Huizar JF, Kaszala K, Potfay J, et al. Left ventricular systolic dysfunction induced by ventricular ectopy: a novel model for premature ventricular contractioninduced cardiomyopathy. Circ Arrhythm Electrophysiol 2011;4:543–549. 20. Wyman BT, Hunter WC, Prinzen FW, et al. Mapping propagation of mechanical activation in the paced heart with MRI tagging. Am J Physiol 1999;276:H881–H891. 21. Prinzen FW, Peschar M. Relation between the pacing induced sequence of activation and left ventricular pump. Pacing Clin Electrophysiol 2002;25:484–498. 22. Spragg DD, Kass DA. Pathobiology of left ventricular dyssynchrony and resynchronization. Prog Cardiovasc Dis 2006;49:26–41.