Risk factors for the development of premature ventricular complex-induced cardiomyopathy: a systematic review and meta-analysis

Background
 Premature ventricular complexes (PVCs) are a potentially reversible cause of heart failure. However, the characteristics of patients most likely to develop impaired left ventricular function are unclear. Hence, the objective of this study is to systematically assess risk factors for the development of PVC-induced cardiomyopathy. Methods We performed a structured database search of the scientific literature for studies investigating risk factors for the development of PVC-induced cardiomyopathy (PVC-CM). We investigated the reporting of PVC-CM risk factors (RF) and assessed the comparative association of the different RF using random-effect meta-analysis. Results A total of 26 studies (9 prospective and 17 retrospective studies) involving 16,764,641 patients were analyzed (mean age 55 years, 58% women, mean PVC burden 17%). Eleven RF were suitable for quantitative analysis (≥ 3 occurrences in multivariable model assessing a binary change in left ventricular (LV) function). Among these, age (OR 1.02 per increase in the year of age, 95% CI [1.01, 1.02]), the presence of symptoms (OR 0.18, 95% CI [0.05, 0.64]), non-sustained ventricular tachycardias (VT) (OR 3.01, 95% CI [1.39, 6.50]), LV origin (OR 2.20, 95% CI [1.14, 4.23]), epicardial origin (OR 4.72, 95% CI [1.81, 12.34]), the presence of interpolation (OR 4.93, 95% CI [1.66, 14.69]), PVC duration (OR 1.05 per ms increase in QRS-PVC duration [1.004; 1.096]), and PVC burden (OR 1.06, 95% CI [1.04, 1.08]) were all significantly associated with PVC-CM. Conclusions In this meta-analysis, the most consistent risk factors for PVC-CM were age, non-sustained VT, LV, epicardial origin, interpolation, and PVC burden, whereas the presence of symptoms significantly reduced the risk. These findings help tailor stringent follow-up of patients presenting with frequent PVCs and normal LV function. Supplementary Information The online version contains supplementary material available at 10.1007/s10840-022-01421-8.


Introduction
Premature ventricular complex-induced cardiomyopathy (PVC-CM) is defined as the development of left ventricular dysfunction (left ventricular ejection fraction (LVEF) of < 50%) caused solely by frequent PVCs [1]. Superimposed PVC-CM can be defined as worsening of LVEF by at least 10% due to frequent PVCs in a previously known CM [1]. Currently, diagnosis of PVC-induced CM can only be made during follow-up, by showing documentation of complete LVEF recovery in absence of PVCs after successful treatment [2].
Clinical studies have found that a high PVC burden is associated with an increased risk of systolic heart failure 1 3 (HF) (hazard ratio [HR]: 1.48 to 1.8) [3,4]. Two main studies have shown that PVC burden > 16% and 24% best identifies patients with a diagnosis of PVC-CM [5,6]. Nevertheless, some patients do not develop CM even with a high PVC burden, whereas other patients develop CM with a burden as low as 6% [7]. Thus, it is likely that other patients' characteristics and/or PVC features besides PVC burden play a role in the pathophysiology of PVC-CM. Multiple predictors of PVC-CM were described including male sex, lack of symptoms or duration of palpitations [8], variability of PVC coupling interval (dispersion) [9], interpolation of PVCs [10], QRS duration of PVC > 150 ms [11], or epicardial origin [12].
Prior studies investigating risk factors for PVC-CM were retrospective and were not designed with the main objective of assessing these RFs [3-8, 11, 12]. In addition, the assessed study populations were very heterogeneous and often the main endpoint was not defined with enough precision. Thus, most predictors have been variably reported and further validation is required.
We therefore conducted a systematic review and metaanalysis of studies addressing clinical, ECG, Holter, or echocardiographic risk factors able to differentiate patients having a PVC-induced CM from other forms of CM.

Methods
This systematic review and meta-analysis received approval from the ethics committee and was registered on PROS-PERO (CRD42021243622). The reporting of our results was done according to the PRISMA statement about systematic reviews and meta-analyses [13] (Supplemental Table 1) and followed the latest guidelines about reporting systematic reviews and meta-analyses of prognostic factors studies [14].

Data sources and search
A comprehensive systematic search was conducted in Pub-Med, MEDLINE, and Embase by combining keywords synonyms of PVC, heart failure, and risk factors as detailed in the Supplemental appendix. The study registry Clinicaltrial. gov was manually searched using the same terms. The search was conducted once on February 27, 2021, accounting for all articles published between January 1, 2000, and February 27, 2021.

Study selection
Studies that met the following pre-specified criteria were included: (1) RCTs, prospective, or retrospective observational studies and registers; (2) with at least 50 patients total (with and without PVC-CM); (3) assessing adult patients with at least part of the cohort diagnosed with PVC-CM and at least part of the cohort presenting with PVCs; (4) investigating risk factors for the development of PVC-CM (which were not defined beforehand); (5) reporting summary statistics such as regression coefficients, odds ratios (OR) or HR; (6) assessing the incidence, prevalence, or recovery of heart failure thought to be related to PVCs or the change in ejection fraction (EF) due to the presence, increase, or reduction of PVCs; (7) providing either time-to-event data or cross-sectional data; and (8) providing at least one adjusted (multivariable) risk-factor model.

Endpoints
The primary endpoint of this meta-analysis was the quantitative meta-analysis of risk factors for the development of PVC-CM. We pre-defined that risk factors should be reported in at least 3 different studies with a compatible definition in order to allow for a meaningful quantitative summary.
Secondary endpoints were either the qualitative analysis of risk factors reported in ≥ 3 different studies or important study characteristics, such as (1) the prevalence of comprehensive work-up to ensure patients diagnosed with PVC-CM did not present with another cause for heart failure; (2) the differences in the reported definitions of PVC-CM; and (3) the assessment of study quality using the validated QUIPS (Quality in Prognosis Studies) tool [15].

Primary outcome
The primary outcome of this meta-analysis was the presence of PVC-CM, which we pre-defined either as the development, presence, or recovery from heart failure with reduced ejection fraction (HFrEF) in patients with CMP in whom no other cause of heart failure was evident. Further details are available in the supplemental.

Analysis of risk factors
A meta-analysis was conducted on risk factors presenting ≥ 3 times throughout the studies. When continuous risk factors were presented using cutoffs, the exposure per group (above and below the respective cutoff) was derived as recommended in previous dose-exposure meta-analyses and corresponding guidelines [16][17][18][19].
Further details regarding the analysis of risk factors are given in the supplemental.

Assessment of study quality
Study quality was assessed according to the QUIPS tool [15] and summarized graphically.

Statistical analysis
The analysis was performed according to the recommendations of the Cochrane Collaboration [20] and the reporting was in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [13] and according to recent guidelines on the conduction of review and meta-analyses of prognostic factor studies [14].
We recorded quantitative measures of baseline characteristics as mean with standard deviation (SD) or median with interquartile range (IQR). To allow for quantitative summaries, we transformed the median with IQR into mean with SDs using a mathematical transformation as proposed in previous research [21].
For the main analysis, in order to increase the number of studies available for the quantitative summary of each risk factor, we summarized odds ratios and hazard ratios as a common measure of risk ratio, as it has been conducted in previous meta-analyses [16,22].
To allow for the expected heterogeneity in effect measures across studies, summary relative risk estimates and their 95% CIs were estimated from a random effect model [23] that used the inverse variance method as proposed by the metagen package [24], which considers both within-and between-study variation. To estimate the between-study variance, the Tau estimator was calculated according to the DerSimonian-Laird estimator [23,25]. Statistical heterogeneity among studies was evaluated using the I 2 statistic [26].
Details of the dose-response analysis are available in the supplemental.
Significant heterogeneity was defined as an I 2 statistic of > 50%.
Evidence for publication bias was assessed for PVC burden graphically using contour-enhanced funnel plots [27] and the Egger test.
The risk of bias within each study was assessed using the QUIPS tool.
All statistical analyses were performed using the Statistical Software "R" (R Foundation for Statistical Computing, Vienna, Austria). P values < 0.05 were considered as significant.
Baseline study characteristics are presented in Table 1. The included studies reported data on patients treated between 1989 and 2019. They consisted of 9 prospective and 17 retrospective studies. One of the retrospective studies was a re-analysis of a register (the California Health Care Cost and Utilization Project (CHCCUP)) evaluating 16,757,903 patients that was qualitatively analyzed but was eventually excluded from the meta-analysis because of the bias caused by its extreme weight. The 25 other studies provided a total of 6738 patients.
Further details regarding inclusion and exclusion criteria for each study and definitions of both PVC-CM and PVCs are presented in Supplemental Table 3. Fifteen of 26 (57.7%) studies provided a definition of PVC-CM: the CMP was mostly defined as an LVEF < 50% and 9/26 (34.6%) studies took a time component into account (e.g., normalization or increase in the EF over time). The requirement for LVEF improvement in the PVC-CM definition varied from 10 to 15% in these studies.

Baseline patient characteristics
Often, several groups were analyzed in each study, which did not always report data for the overall cohort. The analyzed groups are presented in Table 2. In summary, the overall patient population was rather young (weighted mean age of 50.2 years old, 55.0 years old when excluding data from the predominant CHCCUP study) and with a weighted mean PVC burden of 16.5% (not reported in the CHCCUP study). The weighted mean percentage of women in the overall analyzed dataset was 57.6%, which decreased to 44.2% when excluding data from the CHCCUP. In a significant proportion of the studies and reported groups, there was no described attempt to assess for the presence of underlying structural heart disease or this detail was not reported (8/26 studies, Supplemental Tables 4 and 5).

Assessment of outcomes
Most of the studies assessed the presence of PVC-CM (17/26), the recovery of LVEF after PVC-CM 4/27 (defined as a binary variable), or the worsening of LVEF suspected to be due to PVC-CM 2/27 (also defined as a binary variable). We conducted a pooled analysis for these three outcomes, as these are solely different ways to define a PVC-CM. Studies reporting continuous LVEF change over time (3/26) were rare (Table 3).  Table 4 presents the occurrence of all risk factors throughout the selected studies and the occurrence of reporting which were suitable for quantitative analysis (≥ 3 occurrences in multivariable model assessing a binary change in LV function). Supplemental table 6 presents the risk factors analyzed by each study. The exact definitions of each risk factor, as provided by the individual studies, are presented in the supplemental.

Assessed risk factors
PVC burden was the most commonly analyzed risk factor (24/26 studies, 20/26 studies for quantitative summary), followed by sex (13/26), PVC origin (11/26), PVC and morphology (10/26), and PVC and QRS duration (each in 8/26 studies). Only few other risk factors (age, coupling interval, non-sustained VTs, interpolation, and the presence of symptoms) were investigated in ≥ 3 studies and suitable for quantitative summary. Further investigated risk factors were baseline LVEF, coupling interval, polymorphic PVCs, and outflow tract origins. These risk factors did not appear often enough (< 3 appearances) or were differently defined, hence not suitable for quantitative summary.  Fig. 1).

Dose-response analysis of PVC burden
In the dose-response analysis encompassing 7 studies reporting PVC burden at different cutoffs, there was a highly significant association between increase in PVC burden and an exponential increase in risk for PVC-CM (at 10% PVC burden,

Modification of the risk associated with PVC burden through meta-regression
When assessing the risk modification associated with the publication year or with study quality, older studies and studies with higher quality were associated with a nonsignificant trend in increased risk for the development of PVC-CM with a growing PVC burden.

Publication bias
On funnel plot analysis of PVC burden, study distribution was mildly asymmetric (Fig. 11) but the Egger test did not suggest any publication bias (P = 0.07).

Quality assessment
As presented in Fig. 12, all of the studies presented with at least a moderate risk of bias. The uncontrolled risk of confounding appeared as the most problematic throughout all recorded studies.

Discussion
This systematic review and meta-analysis investigated 26 studies to investigate risk factors associated with the development of PVC-CM. We report four major findings. First, despite screening abstracts published over 30 years of scientific research, only few studies presented a multivariable assessment of risk factors potentially associated with PVC-CM and the quality of the research currently does not allow for definitive conclusion. Second, although many candidate risk factors were proposed by the analyzed studies, only 13 risk factors   Fig. 11 Assessment of publication bias using a contour-enhanced funnel plot. The contour-enhanced funnel plot represents the different studies reporting estimated for the association between PVC burden (continuous increase in %) and assess the risk for publication bias. The 7 studies reporting a cutoff of PVC burden were summarized beforehand as the "dose-response analysis." The dotted line represents the overall estimate using all available studies and the dashed line represents a classical funnel plot with the expected distribution of the studies if no publication bias is present. The contour-enhanced funnel plot is centered at 0 (i.e., the value under the null hypothesis of no relationship) and various levels of statistical significance are indicated by the shaded region. The white region corresponds to nonsignificant P values. Highly significant P values appear in the light gray region (age, PVC burden, PVC origin from epicardial, outflow tract or LV, interpolation, non-sustained VTs, presence of symptoms, coupling interval, PVC morphology and duration, QRS duration, and sex) were reported often enough with appropriate statistics to allow for a quantitative summary. Many other predictors remain possible candidates for the risk stratification of PVC-CM development. Third, age, non-sustained VTs, LV and epicardial origin, interpolation, PVC duration, and PVC burden were all associated with an increased risk for PVC-CM, whereas the presence of symptoms significantly reduced the risk. Fourth, there was a clear association between increasing PVC burden and increasing PVC-CM risk. In the dose-response analysis encompassing 7 studies reporting PVC burden at different cutoffs, there was a highly significant association between increase in PVC burden and an increasing risk for PVC-CM. Specifically, per % increase in PVC burden, there was an exponential increase in the absolute risk of PVC-CM. This association was not significantly impacted by the study publication year, suggesting that despite improvements in heart failure treatments and prevention over years, the burden remains an important predictor of PVC-CM development.
To the best of our knowledge, this is the first systematic review and meta-analysis comprehensively assessing the risk factors for the development of PVC-induced cardiomyopathy. The optimal approach to frequent PVCs (> 10% burden) without LV dysfunction, symptoms, or idiopathic ventricular fibrillation is unclear, but patients should probably be monitored every 6-12 months with echocardiography and PVC burden assessment [47]. Therefore, until PVC-induced cardiomyopathy can be predicted, these results help to focus on patients at the highest risk of developing PVC-CM. The role of early rhythm control with catheter ablation or AAD of frequent PVCs without LV dysfunction and symptoms, but risk factors, needs to be defined.
Several studies have confirmed a correlation between a higher PVC burden and development of cardiomyopathy, although no precise burden of PVCs consistently predicts the development of a cardiomyopathy. In this meta-analysis, we found a highly significant association between an increase in PVC burden and increasing risk for PVC-CM.

Limitations
This systematic review and meta-analysis has several limitations. First, the quantitative summary of risk factors we are presenting summarizes different measures of risks (odds and hazard ratios) together. While this has been conducted in previous research and is acknowledged by recent guidelines as a possible necessary simplification [14], this might have biased absolute risk estimated. Second, most of the articles had different definitions for the risk factors. As such, only 15 of the 26 analyzed studies (57%) provided a definition for PVC-CM and only 9 of the 26 (34.6%) assessed the evolution of EF into the model. The latest literature on PVC-CM [2,48,49] recommends assessing the temporal course of worsening or recovery of EF over time. Thus, about three fourth of the studies we investigated did not define their main endpoint with enough precision. At the same time, none of the three included studies provided a standardized definition for non-sustained tachycardia, limiting the credibility of the result. Fig. 12 Assessment of study quality. Evaluation of study quality according to the QUIPS tool. Five domains of bias (participation, attrition, prognostic factor measurement, outcome measurement, confounding and statistical analysis and reporting) are represented with the associated risk of bias (high in red, moderate in yellow, and low in green). The overall column represents the mean risk of bias from the 6 domains Third, as several studies did not thoroughly assess other underlying heart failure etiologies in their patients collectively, our estimates may have been occasionally confounded by other causes of heart failure. Fourth, as most of the studies providing a PVC burden cutoff only provided two categories, we had to assume a linear trend between PVC exposure and the associated increase in risk (thereby leading to an exponentially growing risk after back-transformation of the log-odds). With more detailed data, quadratic estimations could lead to more accurate dose-response relationship modelling.

Conclusion
In this meta-analysis, the most consistent risk factors for PVC-CM were age, non-sustained VTs, LV and epicardial origin, interpolation, PVC duration, and PVC burden, while the presence of symptoms significantly reduced the risk. These findings help tailor stringent follow-up to patients presenting with frequent PVCs and normal LV function.

Supplementary Information
The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s10840-022-01421-8.Funding Open access funding provided by University of Basel

Declarations
Ethics approval The study received approval of the ethics' committee.

Consent to participate Not applicable.
Conflict of interest Sven Knecht has received funding of the "Stiftung für Herzschrittmacher und Elektrophysiologie." Michael Kühne reports personal fees from Bayer, personal fees from Böhringer Ingelheim, personal fees from Pfizer BMS, personal fees from Daiichi Sankyo, personal fees from Medtronic, personal fees from Biotronik, personal fees from Boston Scientific, personal fees from Johnson & Johnson, personal fees from Roche, grants from Bayer, grants from Pfizer, grants from Boston Scientific, grants from BMS, grants from Biotronik, and grants from Daiichi Sankyo, all outside the submitted work. Christian Sticherling is a Member of Medtronic Advisory Board Europe and Boston Scientific Advisory Board Europe, received educational grants from Biosense Webster and Biotronik and a research grant from the European Union's FP7 program and Biosense Webster, and lecture and consulting fees from Abbott, Medtronic, Biosense-Webster, Boston Scientific, Microport, and Biotronik, all outside the submitted work. Patrick Badertscher has received research funding from the "University of Basel," the "Stiftung für Herzschrittmacher und Elektrophysiologie," the "Freiwillige Akademische Gesellschaft Basel," and Johnson & Johnson, all outside the submitted work and reports personal fees from Abbott. Jeanne du Fay de Lavallaz has received research funding from the "University of Basel" and from the "Swiss Heart Foundation." Henry Huang has received research funding from Medtronic, educational grants from Medtronic, Biotronik, Abbott, and Boston Scientific, and consulting fees from Biosense-Webster and Cardiofocus.
Michael Gold is a consultant to Boston Scientific and Medtronic, as well as on steering committees with Boston Scientific, Abbott, and Medtronic. Others have nothing to declare.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.