Article ID: CJ-24-0438
Background: Asymptomatic premature ventricular complex (PVC) in childhood often disappears over time. However, predictive factors for persistent PVC are unknown. We examined predictive factors for persistent PVCs on initial Holter electrocardiogram (ECG) in pediatric patients with asymptomatic PVC.
Methods and Results: The initial Holter ECG findings of untreated PVC patients (n=216) between 2010 and 2021 were examined. Multivariable analysis was performed to clarify predictive factors for subsequent persistent PVC burden for each index (age, sex, PVC burden, PVC origin, minimum and maximum mean RR intervals [RRmin and RRmax, respectively]) of the 3 heartbeats of baseline sinus rhythm immediately before the PVC. The median age at initial Holter ECG was 11.6 years (range 5.8–18.8 years), the PVC burden was 5.22% (range 0.01–44.21%), RRmin was 660 ms, RRmax was 936 ms, RRrange (=RRmax−RRmin) was 273 ms, and 15 (7%) PVC runs were identified. The median follow-up period was 5.1 years (range 0.8–9.4 years), and the final Holter PVC burden was 3.99% (range 0–36.38%). In multivariate analysis, RRrange was the only independent risk factor for predicting a final Holter PVC burden >10%, with an area under the curve of 0.920 using an RRrange of 600 ms as the cut-off value.
Conclusions: A wide RRrange at the initial Holter ECG may be a predictive indicator for persistent PVC in childhood.
Premature ventricular complex (PVC) is relatively common in children, occurring in 1–2% of those with normal hearts and in up to 50–60% of adolescents.1,2 In Japan, many patients with asymptomatic PVC are identified by school-based cardiovascular screening programs. Due to the short duration of electrocardiography (ECG) screening, the detection frequency among students is 0.3–0.4%.3 Most cases of asymptomatic PVC disappear after several years of follow-up.4,5 However, when it persists, PVC can lead to impaired ventricular contraction, with impaired ventricular contraction potentially including a high burden of PVC (>24%), an epicardial origin, and a wide QRS (>150 ms).6–8 Nevertheless, predictive factors for the high burden of persistent PVC have not been verified. Although several reports have evaluated PVC prevalence and frequency from various parameters of heart rate (HR) variability (HRV), none has evaluated the subsequent PVC burden from patient-specific PVC characteristics.9,10 In patients without structural heart disease, the most likely mechanism of PVC is triggered activity.11,12 The burden of PVC increases under certain conditions, like decreased HR and sympathetic nerve activity. Therefore, we hypothesized that PVC observed across a wide range of HRs during baseline sinus rhythm (BSR) at initial diagnosis would persist. The aim of this study was to determine the factors evident on initial Holter ECG that are correlated with a high burden of subsequent PVC in childhood.
This single-center retrospective study was conducted at the Hokkaido University Graduate School of Medicine, Japan, after approval by the Institutional Review Board (IRB) of Hokkaido University Hospital for Clinical Research (IRB Approval no. 021-0192). The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki. The requirement for informed consent was waived due to the study’s retrospective nature. However, information regarding the present study was disclosed on the Hokkaido University website to allow patients to opt-out of the study.
Study PatientsIn Hokkaido, all students undergo the school-based cardiovascular screening program during their first year of elementary school and junior high school using questionnaires, a 12-lead ECG for 10 s, and phonocardiography. A second examination, if required, is performed in a hospital.
We enrolled 216 patients who were followed up at Hokkaido University Hospital for asymptomatic PVC diagnosed by a school-based cardiovascular screening program between January 2010 and December 2021. All patients were confirmed to have no structural heart disease by cardiac echocardiography. We analyzed correlations between the PVC burden in follow-up periods and data from patients’ medical records, which included patient profiles (age and sex), PVC origin identified by 12-lead ECG,13 and initial Holter ECG findings (PVC burden, the presence of PVC runs, the minimum RR time [RRmin], maximum RR time [RRmax], and RRrange [=RRmax−RRmin], where RR is defined as the baseline mean RR interval for 3 beats of sinus rhythm before a PVC, and the minimum, maximum, and range of PVC coupling time [CTmin, CTmax, and CTrange, respectively] from the previous beat of sinus QRS). Symptomatic patients, patients who were taking antiarrhythmic drugs, and those with idioventricular rhythm were excluded.
DefinitionsThe diagnosis of cardiac arrhythmia and the presence or absence of structural heart disease in all patients were confirmed by 4 pediatric cardiologists (G.I., H.Y., A.T., and S.S.). Figure 1 shows the algorithm for identifying the PVC origin by 12-lead ECG. The definitions of RRmin, RRmax, RRrange, and coupling time (CT) are shown in Figure 2. RRmin and RRmax were measured by selecting about 10 candidate sites each from the compressed 24-h Holter recordings, and RRrange was calculated from RRmin and RRmax. However, even among patients with a wide RRrange, there may have been cases in which PVCs occurred in various RRs or in only 2 fixed RRs. Therefore, we measured RRs at a minimum of 5 points in patients with an RRrange >400 ms to ensure that no patient had PVC occurrence in the 2 fixed RRs. CT was measured in the same way.
Algorithm for identifying the origin of a premature ventricular complex (PVC) by 12-lead electrocardiography (ECG). LBBB, left ventricular branch block type; LVIN, left ventricular inflow; LVOT, left ventricular outflow; RBBB, right ventricular branch block type; RVIN, right ventricular inflow; RVOT, right ventricular outflow; T-zone, transition zone.
Definitions of minimum, maximum, and range of the RR interval (RRmin, RRmax, and RRrange [=RRmax−RRmin], respectively), and coupling time (CT). RR is defined as the baseline mean RR interval for 3 beats of sinus rhythm before a premature ventricular complex (PVC). CT is defined the coupling time of the PVC from a previous sinus QRS beat. RRmin and RRmax were measured by selecting about 10 candidate sites each from the compressed 24-h Holter recordings, and RRrange was calculated from those 2 points. CT was measured in the same way.
Statistical Analysis
Data were analyzed using SPSS (version 22.0; IBM Japan, Ltd., Tokyo, Japan). Numerical variables are expressed as the median with range, and categorical data are presented as percentages. Factors significantly associated with persistent PVCs were subsequently analyzed in multivariable analyses using a logistic regression model. Pearson’s correlation coefficient was determined between the burden of persistent PVC and the RRrange. Moreover, receiver operating characteristic (ROC) curve analysis was use to determine the cut-off value for predicting a persistent PVC burden >10%, in accordance with recent reports. Statistical significance was set at two-tailed P values <0.05.
Patient characteristics are presented in Table 1. The median age at initial Holter ECG across all patients was 11.6 years (range 5.8–18.8 years) and the median PVC burden was 5.22%, with one patient having a high PVC burden of 44.21%. The most common PVC origin was the right ventricular outflow tract. RRmin represents the fastest BSR in which PVCs appear. When converted to HR, the median BSR-HR was 91 beats/min, with a maximum BSR-HR of 214 beats/min where PVCs appeared. RRmax represents the slowest BSR in which PVCs appear. The median BSR-HR was 64 beats/min, with a minimum BSR-HR of 32 beats/min where PVCs appeared. An RRrange of 0 indicates that PVCs appear at only 1 fixed rate, whereas an RRrange of 1,320 ms indicates that PVCs appear over a wide range of underlying rates, from a BSR-HR of 30 beats/min to over 200 beats/min. PVC runs were present in 7% of patients. Figure 3 shows the clinical course of the PVC burden in all patients. Most patients were initially detected with a PVC burden <10%, which gradually decreased over time. Some patients had a high initial burden that persisted, whereas others had a low initial burden that increased over time. No patient showed significant impairment of ventricular contraction during the median follow-up period of 5.1 years.
Patient Characteristics (n=216)
| Male sex | 112 (52) |
| Follow-up period (years) | 5.1 [0.8–9.4] |
| PVC origin identified by 12-lead ECG | |
| RVIN | 43 (20) |
| RVOT | 80 (37) |
| LVIN | 43 (20) |
| LVOT | 50 (23) |
| Septal/free wall | 108 (50)/108 (50) |
| Initial Holter ECG findings | |
| Age at initial Holter ECG (years) | 11.6 [5.8–18.8] |
| PVC burden (%) | 5.22 [0.01–44.21] |
| Presence of PVC run | 15 (7) |
| Presence of multifocal PVCs | 22 (10) |
| RRmin (ms) | 660 [280–1,120] |
| RRmax (ms) | 936 [480–1,880] |
| RRrange (ms) | 273 [0–1,320] |
| CTmin (ms) | 457 [240–1,120] |
| CTmax (ms) | 633 [360–1,600] |
| CTrange (ms) | 173 [0–1,320] |
| Final Holter ECG findings | |
| PVC burden (%) | 3.99 [0.00–36.38] |
Data are presented as the range or n (%). CT, coupling time; ECG, electrocardiogram; LVIN, left ventricular inflow; LVOT, left ventricular outflow; max, maximum; min, minimum; PVC, premature ventricular complex; RR, RR interval; RVIN, right ventricular inflow; RVOT, right ventricular outflow.
The clinical course of premature ventricular complex (PVC) burden in all patients. Most patients were detected with a PVC burden <10%, which gradually disappeared. Some had a high initial burden that persisted, whereas others had a low initial burden that increased afterwards. Initial, initial Holter ECG; 1yr after, 12 months after the initial Holter ECG; last, final Holter ECG performed a median of 5.1 years after the initial Holter ECG.
Analysis of Factors Predictive of Persistent PVC
Predictive factors and their odds ratios (ORs) are summarized in Table 2. Multivariate analysis revealed that the RRrange at initial Holter ECG (OR 2.56; 95% confidence interval [CI] 1.47–3.89) was an independent predictive factor for a persistent PVC burden >10%. Figure 4 shows that the RRrange was positively correlated with the burden of persistent PVC at the final Holter ECG (R2=0.461; P=0.02). ROC curve analysis identified an RRrange cut-off value of 600 ms (sensitivity, 88.0%; specificity, 87.0%; area under the curve, 0.920) for predicting the burden of persistent PVC in childhood (Figure 5).
Risk Factors for Persistent >10% Burden of PVC
| Univariable analysis | Multivariable analysis | |||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | |
| Male sex | NS | |||||
| PVC origin identified by 12-lead ECG | ||||||
| RVIN | NS | |||||
| RVOT | NS | |||||
| LVIN | NS | |||||
| LVOT | NS | |||||
| Septal/free wall | NS/NS | |||||
| Initial Holter ECG findings | ||||||
| Age at initial Holter ECG | NS | |||||
| PVC burden | 1.67 | 1.34–3.13 | <0.02 | NS | ||
| Presence of PVC run | NS | |||||
| Presence of multifocal PVCs | NS | |||||
| RRmin | NS | |||||
| RRmax | NS | |||||
| RRrange | 2.72 | 1.54–4.28 | <0.01 | 2.56 | 1.47–3.89 | <0.01 |
| CTmin | NS | |||||
| CTmax | NS | |||||
| CTrange | NS | |||||
CI, confidence interval; OR, odds ratio. Other abbreviations as in Table 1.
Scatterplots depicting RRrange at the initial Holter electrocardiogram (ECG) vs. the premature ventricular complex (PVC) burden at the final Holter ECG. RRrange was positively correlated with persistent PVC burden at the final Holter ECG.
Receiver operating characteristic curve for predicting a persistent premature ventricular complex (PVC) burden >10% in pediatric patients diagnosed with asymptomatic PVC, using RRrange on the initial Holter electrocardiogram. AUC, area under the curve.
Although most pediatric patients with PVC are asymptomatic, a high PVC burden can impair ventricular contraction.6,14 Therefore, it is important to predict the risk of persistent PVC in these patients, consider the follow-up period, and initiate early and effective therapy. This retrospective study investigated factors predictive of persistent PVC in pediatric patients diagnosed by a school-based cardiovascular screening program and included parameters obtained from the initial Holter ECG. A wide RRrange was found to be an independent predictive factor for the subsequent high burden of PVC in childhood.
Previous reports have shown that low values of time domain parameters of HRV predict PVC prevalence,9,10 whereas higher low frequency/high frequency ratios predict a high burden of PVC. These studies evaluated PVC prevalence and frequency trends based on the characteristics of each patient’ 24-h HRV.9,10 In contrast, the present study is the first to report the prediction of subsequent PVC frequency based on the characteristics of patient-specific PVC occurrence.
The initiation of PVCs is dependent on the underlying cardiac substrate and can be explained by re-entry,11 automaticity, and triggered activity. The mechanism of scar-related PVCs after myocardial infarction is re-entry. Re-entrant PVCs tend to occur under specific HR conditions. Conversely, the most likely mechanism of PVC in patients without structural heart disease is triggered activity. The conditions for the appearance of triggered activity PVCs are unique to each patient in terms of physical activity and sympathetic nerve regulation. This study demonstrated that PVCs in non-structural heart disease, which are more likely to occur under a wider range of baseline HR conditions, have a tendency to persist over time.
Ito et al evaluated the PVC frequency as a function of underlying HR.15 Observed patterns of correlation between PVC frequency and HR included: (1) a relatively linear increase in PVCs with HR (positive correlation); (2) a linear decrease (negative correlation); and (3) an increase at low HR and a decrease at high HR (bidirectional correlation).15 Ito et al concluded that tachycardia-accelerated PVCs may be related to triggered activity, because diltiazem or atenolol effectively suppressed PVCs in patients with a positive correlation.15 A wide RRrange corresponds to a “bidirectional correlation”, indicating that it can be a feature of being a PVC likely to persist.
The significance of predicting residual PVCs using the RRrange at the initial Holter has 3 aspects. First, even if the initial PVC burden is low, a wide RRrange indicates the need for careful follow-up. Second, even with a high PVC burden at the initial Holter, it may be possible to set an RRrange cut-off level to determine when no further follow-up is necessary. Third, it is crucial to refer pediatric patients with PVCs that do not seem to resolve spontaneously to an appropriate cardiologist as they grow older.
For the adequate management of patients with PVCs and to alleviate parents’ anxiety about their children, it is essential to consider predictive factors for persistent PVCs. This should be done in addition to evaluating other factors for impaired ventricular contraction, such as epicardial origin and wide QRS (>150 ms), when examining asymptomatic pediatric patients with PVCs.
Study LimitationsThere were 3 notable limitations of this study: (1) its single-center and retrospective case-control design; (2) the small number of patients; and (3) the nature of school screening, which means that not all patients with PVCs were detected.
The present study demonstrated that RRrange at the initial Holter ECG is a simple and convenient predictive factor for the subsequent persistence of PVC. Therefore, RRrange analysis should be performed in outpatient clinics to determine the follow-up period, considering the possibility of impaired ventricular contraction in these patients.
The authors thank Editage (www.editage.com) for English language editing a draft of this paper.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
The authors have no conflicts of interest to disclose.
G.I. and S.S. conceived the study and interpreted the data. G.I. drafted the manuscript. A.T., H.Y., N.K. and S.S. revised the manuscript for critically important intellectual content. All authors approved the final submitted manuscript.
This study was conducted at the Hokkaido University Graduate School of Medicine, Japan, after approval by the IRB of Hokkaido University Hospital for Clinical Research (IRB Approval no. 021-0192).
The deidentified participant data will be shared on a request basis. Please contact the corresponding author directly to request data sharing. The entire dataset used will be available, including the study protocol. Data will be shared as soon as it is approved by the IRB at Hokkaido University Graduate School of Medicine, and will be available until end of March 2034. Data will be shared with anyone wishing to access it. Any analyses on the data will be approved and data will be shared as an Excel file via email.
