All enrolled patients were diagnosed with frequent VPCs at Konkuk University Medical Center from January 2012 to February 2014. The study protocol was approved by the hospital's Institutional Review Board. Inclusion criteria were predominant VPCs exhibiting ECG characteristics suggestive of outflow tract (OT) origin (right or left bundle branch morphology and an inferior axis), age≥19 years, frequent VPCs (>10% VPC burden per 24 hours) according to 24-hr Holter ECG monitoring at enrollment without undergo medical or procedural VPC suppression, full description of symptoms in the medical records including questionnaire and Holter monitoring, and baseline echocardiography and Holter monitoring at enrollment. The exclusion criteria were history of atrial fibrillation, atrial flutter, atrial tachycardia, nonsustained ventricular tachycardia, sustained ventricular tachycardia, or evidence for any of these arrhythmias by 12-lead ECG or Holter ECG monitoring; history of myocardial infarction, structural heart disease, or heart valve replacement/repair; and any evidence of ischemic/structural heart disease based on information obtained from the echocardiogram, radionuclide evaluation, and/or cardiac catheterization. In the patients with VPC-induced CMP, transthoracic echocardiography was performed after the procedure to assess any structural changes and to confirm the status of the left ventricular ejection fraction (LVEF) in the absence of VPCs after a successful radiofrequency (RF) ablation or medical treatment. If the patients had a depressed LVEF (<50%), we enrolled only patients who recovered their left ventricular (LV) function back to normal (an improvement in the LVEF of ≥10% to a final LVEF of ≥50%) after successful VPC suppression. Successful treatment was defined as at least an 80% reduction in the 24-hr burden of VPCs, based on our previously published experience.27) Patients were divided into two subgroups (symptomatic or asymptomatic) according to the presence or absence of typical VPC-related symptoms. Symptom evaluation was determined by reviewing the cardiology records (medical records created by a cardiologist). If the patient did not feel any palpitations or "dropped beats" related to VPCs observed on an ECG, the patient was assigned to Group A. If a patient felt palpitations or dropped beats when VPCs appeared on an ECG or Holter monitoring, this was defined as a typical VPC-related symptom, and the patient was assigned to Group B. Especially, a Holter monitoring and a questionnaire at enrollment was evaluated in detail to match the correlation between the clinical VPC and the symptom. If a patient recorded any symptoms (palpitation or dropped beats) during an episode of VPCs on Holter monitoring, the patient was judged to have a typical VPC-related symptom. All asymptomatic patients were diagnosed with frequent VPCs during regular ECG check-ups or before non-cardiac procedures or operations. They were referred to our institution for further management of their VPCs.

The 12-lead ECG was recorded at a sweep speed of 100 ms and analyzed off-line with a Muse® Cardiology Information System (GE Healthcare, Piscataway, NJ, USA) using digital calipers. VPCs were evaluated with respect to the QRS duration, amplitude, and morphology. VPC notching in the precordial and limb leads was evaluated. Notching was defined as a >40 msec (ms) interval between the initial and second peaks of the QRS complex in the precordial or limb leads. Several other ECG parameters were measured on the surface ECG (Fig. 1):

• Baseline sinus cycle length (ms): from the R peak of one sinus beat to the R peak of the next sinus beat

• VPC QRS width (ms): from the onset of the VPC to the terminal S wave

• VPC CI (ms): from the onset of the R wave of the previous sinus beat to the onset of the VPC

• VPC CI ratio (%): VPC CI/sinus cycle length×100%

• Post-VPC CI: from the onset of the VPC to initiation of the next sinus beat

• Post-VPC CI ratio (%): post-VPC CI/sinus cycle length×100%

• VPC amplitude (mV): highest amplitude of the VPC in the precordial leads

The Muse® Cardiology Information System (GE Healthcare, Piscataway, NJ, USA) was used to measure the width and amplitude of the VPCs, as well as CI and cycle length of both the VPCs and sinus beats. To assess intra-observer variation, parameters were measured for five consecutive normal sinus rhythms and VPC beats.

For the patients who underwent successful RF ablation procedure, the baseline demographic data, history, and other clinical characteristics were collected prospectively. Additional clinical and ECG parameters were assessed retrospectively in all enrolled patients by a detailed medical record review. All ECG measurements were performed using digital calipers at 100 mm/s, on the Muse® Cardiology Information System (GE Healthcare, Piscataway, NJ, USA). For each measured parameter, the mean of five measurements was used for analysis to minimize the influence of the measurement error. For patients with multiple VPC morphologies, only the dominant type of VPCs was measured.

Pearson's product moment correlation coefficient was calculated to quantify inter-rater variability in the measurement of both VPC QRS duration and VPC coupling interval. Continuous data are expressed as mean±standard deviation unless otherwise specified. Between-groups comparisons were performed using Student's t-tests, whereas within-group comparisons were performed using paired t-tests. For comparisons of non-continuous variables, chi-square tests were used. SPSS Version 18.0 for Windows (SPSS, Chicago, IL, USA) software was employed for the analysis. For all tests, p<0.05 was considered statistically significant.

There were no significant differences between groups for age (Group A vs. Group B: 46.1±16.6 vs. 50.1±13.9 years, p=0.22), body surface area (p=0.34), daily VPC burden (Group A vs. Group B: 22.1±11.9 vs. 26.5±12.8%, p=0.15), and VPC site of origin (p=0.39) (Table 1). Except for antiarrhythmic drugs (p=0.01), there were no other significant differences between groups in the use of medications including beta-blockers (p=0.19), calcium channel blockers (p=0.60), angiotensin converting enzyme inhibitors (p=0.73), and angiotensin receptor blockers (p=0.48). VPC-related typical symptoms (palpitations and dropped beats) occurred only in Group B. The incidence of atypical VPC symptoms (shortness of breath, syncope, dizziness, and fatigue) was similar between Group A and Group B. LVEF was significantly lower in Group A (Group A vs. Group 44.3±16.5 vs. 57.2±14.1%, p=0.02), and the LV end-systolic dimension (Group A vs. Group B: 32.3±5.2 vs. 28.5±4.3 mm, p=0.02). The LV end-diastolic dimension (Group A vs. Group B: 54.4±6.3 vs. 49.7±5.6 mm, p=0.01) were significantly greater in Group A.

For all enrolled patients, there were no significant differences in sinus QRS width (p=0.17), sinus cycle length (p=0.86), VPC CI (p=0.82), VPC QRS width (p=0.12), post-VPC CI (p=0.68), and post-VPC CI ratio (p=0.16) between Group A and Group B (Table 2). The VPC CI ratio was significantly higher in Group B (Group A vs. Group B: 149±22 vs. 60±15%, p=0.01). The VPC amplitude in the precordial (p=0.86) and limb leads (p=0.53), and the incidence of a notch in the precordial leads (p=0.09) were not significantly different between Groups A and B. However, the incidence of a notch in the limb leads was significantly higher in Group A (Group A vs. Group B: 50% vs. 26%, p=0.03). There was no significant different in the bundle branch block pattern (p=0.31) between Groups A and B (Table 3). A right bundle branch block pattern was observed in nine patients (20.8%) in Group A and 16 (15.8%) in Group B. The distribution of precordial R wave transition did not differ between Groups A and B (p=0.88). In Group A, early transition (before V₃) occurred in seven patients (23.3%) and late transition (after V₃) occurred in 17 patients (57.4%). In Group B, early transition occurred in 15 patients (18.8%) and late transition in 47 patients (59.4%). The distribution of QRS axis also did not significantly differ between groups (p=1.00). An inferior axis was observed in all patients in both groups. In the normal EF group, there were no significant differences in sinus QRS width (p=0.37), sinus cycle length (p=0.05), VPC QRS width (p=0.62), VPC CI (p=0.82), post-VPC CI (p=0.45), and post-VPCs CI ratio (p=0.37) between Groups A and B (Table 4). The VPC CI ratio was significantly higher in Group B than in Group A (Group A vs. Group B: 48±20 vs. 59±11%, p<0.01). The incidence of a notch in both the precordial and limb leads was similar for Groups A and B. In the depressed EF group, there were no significant difference in sinus QRS width (p=0.21), sinus cycle length (p=0.08), VPC QRS width (p=0.61), VPC CI (p=0.88), post-VPC CI (p=0.63), and post-VPC CI ratio (p=0.67) between Group A and Group B (Table 4). The VPC CI ratio was significantly longer in Group B than Group A (Group A vs. Group B: 50±31 vs. 61±23%, p=0.01). The incidence of a notch in the limb leads was significantly higher in Group B (p=0.04).

The correlation coefficient for measurements of VPC QRS duration was 0.871 and for VPC coupling interval was 0.906.

We specifically selected VPCs that originated from the OT of both ventricles, so the results may not represent symptom-related hemodynamics of other VPC sites of origin. Determining whether patients did or did not have symptoms associated with VPCs was a critical aspect of the study, as it was used to divide the patients into the appropriate groups. However, this process was difficult using a simple medical record analysis. Symptoms may show day-to-day and time-to-time variation. Even though we measured the CI at five time points to check reliability, VPC related symptoms are also associated with the change of autonomic nervous system and emotional status in addition to the coupling interval. Also, we could not match the VPC-related symptoms to each VPC with just surface ECG and Holter monitoring. Another potential limitation of the study was that the sinus cycle length may continuously change throughout the day. Therefore, the CI ratio and other parameters used in this study would vary (and not be a single value) during each sampling period. To minimize the influence of measurement error because of this variation, the mean of five ECG measurements was used for our analysis. In addition, the main indication for VPC ablation in patients with normal ventricular function was the presence of severe symptoms. Consequently, the asymptomatic group was smaller than the symptomatic group. Another limitation was that the patients included in the study were those who were referred to our institution for further management (ablation and/or medications), and thus our findings may have been influenced by referral bias. We were unable to systematically track those patients who were not referred to our institution or those who refused to take medications or undergo ablation. This limitation is shared with prior reports of ablation in this population.