Arrhythmia-Induced Cardiomyopathy: JACC State-of-the-Art Review

Abstract

Arrhythmias coexist in patients with heart failure (HF) and LV dysfunction. Tachycardias, AF and premature ventricular contractions (PVCs) are known to trigger a reversible dilated cardiomyopathy (CM) referred as arrhythmia-induced CM (AiCM). It remains unclear why some patients are more prone to develop AiCM despite similar arrhythmia burdens. The challenge is to determine whether arrhythmias are fully, partially or at all responsible for an observed LV dysfunction. AiCM should be suspected in patients with mean HR above 100 bpm, AF and/or PVCs burden equal or greater than 10%. Reversal of cardiomyopathy by elimination of the arrhythmia confirms AiCM. Therapeutic choice depends on the culprit arrhythmia, patient comorbidities and preferences. Following recovery of LV function, patients require continued follow up if an abnormal myocardial substrate is present. Appropriate diagnosis and treatment of AiCM is likely to improve quality of life, clinical outcomes and reduce hospital admission and health care spending.

Condensed Abstract

Persistent tachycardia, atrial fibrillation and frequent ventricular ectopy can trigger a reversible form of dilated cardiomyopathy, referred as arrhythmia-induced cardiomyopathy (AiCM). Clinical presentation may vary from asymptomatic to severe heart failure. Clinicians should have a high suspicion especially in patients without another obvious etiology. Ambulatory ECG monitors are key to screen and properly diagnose AiCM. Arrhythmia suppression will not only reverse LV dysfunction with its associated morbidity and mortality, but most importantly improve patient’s health, quality of life and long-term prognosis.

Introduction

Arrhythmias have been long considered part of the clinical presentation of heart failure (HF) and cardiomyopathy (CM). Yet, supraventricular or ventricular tachyarrhythmias alone can result or trigger a reversible non-ischemic CM. Most recently, atrial fibrillation (AF) despite adequate rate control, as well as, premature ventricular contractions (PVCs) have been recognized as a unique etiology of non-ischemic dilated cardiomyopathy (1,2,3). Since the recognition of PVC-Cardiomyopathy. Thus, a more inclusive term of Arrhythmia-induced CM (AiCM) has emerged to include both, tachycardia-induced CM (T-CM), AF-induced CM (AF-CM) and PVC-Cardiomyopathy (PVC-CM) (Central Illustration). However, this term does not include recently recognized CM due to conduction abnormalities / dyssynchrony, such as chronic RV pacing, left bundle branch block and pre-excitation (1,4).

Central Illustration. Arrhythmia-induced Cardiomyopathies: Possible Triggers, Mediators, Effect, and Recovery.

While T-CM was first described in a patient with atrial fibrillation at the beginning of the 20^th century, PVC-CM was only recognized nine decades later in 1998 (6,7). Significant skepticism remained on the cause-relationship between arrhythmias and cardiomyopathies until experimental animal models developed in 1962 and 2011 (8,9), respectively, proved that these sustained arrhythmias could result in LV dysfunction in structurally normal hearts. This review article presents an update of the current understanding of AiCM.

Tachycardia-induced Cardiomyopathy (T-CM)

Definition and prevalence

T-CM refers to the presence of a reversible LV dysfunction solely due to increase in ventricular rates, regardless of tachycardia origin. The risk of developing T-CM depends not only on the type but also duration and rate of tachycardia. The overall prevalence and incidence of T-CM is unclear and likely underestimated.

A study reported T-CM in 2.7% patients referred for radiofrequency ablation (RFA), however it also included patients referred for PVC ablation (10). T-CM has been reported in 10% of patients with atrial tachycardia (AT) (11), and as high as 37% in patients with incessant AT. Moreover, permanent junctional reciprocating tachycardia (PJRT) appears to have the highest association with T-CM (20–50%) as it frequently presents as an incessant supraventricular tachycardia (SVT) (12). Despite atrial fibrillation (AF) being the most prevalent arrhythmia, there is no clear data on the prevalence of T-CM in this population. Only a single study reported T-CM in 4% of patients referred for pulmonary vein isolation (13), however, this data is confounded by selection and referral biases. Children are also prone to T-CM resulting most frequently from AT (59%), PJRT (23%) and ventricular tachycardia (7%) (12).

Causes

T-CM can manifest in the setting of either an incessant or paroxysmal tachycardia and it should be suspected if no other cause of LV dysfunction is identified. A frequent challenge is to identify a superimposed T-CM when tachycardia worsens a known CM. T-CM has been reported to present weeks or months to years after the onset of tachycardia (14).

Supraventricular arrhythmias are the most common etiology, namely AF and atrial flutter with rapid ventricular response. Although rare, other arrhythmias such as incessant or very frequent paroxysmal AT, persistent AV reciprocating tachycardia (AVRT) and AV nodal reentrant tachycardia (AVNRT), sustained sinus tachycardia, frequent ventricular tachycardias (idiopathic, bundle branch and fascicular) and pacemaker mediated tachycardia can also be responsible (11,14). In general, it is suspected that faster tachycardia rates and ventricular arrhythmias cause a more severe T-CM although no studies exist to support this hypothesis (1).

Pathophysiology & Mechanism

Animal models have been key to understand the pathophysiology and mechanism of T-CM. Similar to humans, animals exposed to persistent tachycardia using a continuous rapid atrial or ventricular pacing develop heart failure symptoms, LV systolic dysfunction and dilatation, decrease in LV dP/dtmax and myocardial blood flow and increase in LV wall stress and end-diastolic pressure and volume (9,15,16). Dilatation tends to be biventricular with mild thinning or no associated hypertrophy or change in heart mass (9,15). The progression of these physiological changes include a decrease in systemic blood pressure and increase in LV and pulmonary artery pressure, which plateaus at 1 week, while cardiac output, ejection fraction and volumes continue to deteriorate the following 4 weeks with development of symptomatic HF within 2–3 weeks (9).

T-CM is characterized by structural and functional myocardial changes (Table 1, Central Illustration). Similar to human studies, T-CM models have also demonstrated electrical remodeling and abnormal Ca homeostasis thought to be responsible for impaired excitation-contraction coupling and diastolic dysfunction (9,16–18). Only total Ca cycling, Ca-channel inhibition and basal ATPase activity have demonstrated a statistical correlation with decrease in LV ejection fraction (LVEF) (17).

Table 1.

Summary of abnormalities found in tachycardia and PVC-Cardiomyopathies based in human and animal models.

	Feature	T-CM	PVC- CM
Tissue remodeling (8,9,16,18,19,59)	Myocardial blood flow	Decreased	??
	Fibrosis	Mild	Absent or mild / ↑ collagen type I
	Extracellular matrix and myocyte basal membrane	Disarray	??
	Myocyte misalignment	Present	??
	Inflammation	Macrophage -dominated ↑↑ MHC- class II; ↑ MMP-9) ↑ D68 macrophages	Absent
Myocyte remodeling (8,16,18,64)	Myocyte loss	Present	??
	Apoptosis	Increased	Unchanged
	T-tubules	Depletion	Depletion
	Sarcomere	Loss and change in sarcomere	??
	Mitochondria	Abnormal size, architecture and function ↑ number at intercalated disc ↑ MPC1	Unchanged Normal oxidative phosphorylation
	ATP and Na-K ATPase	Reduced	??
	β -adrenergic receptors	↓ number of β-receptors ↓ G stimulator protein density & adenylate cyclase activity ↑ G inhibitory protein density	??
	Oxidative stress	Increased	??
	Glucose metabolism	Impaired	??
Electrical remodeling (8,16,17,24,59,63,64)	Action potential duration	Prolonged	Prolonged (heterogeneity)
	VERP	Increased	Increased
	Ion Currents	↓ ICa	↓ ICa and Cav1.2 ↓ Ito and Kv4.3 ↓ IKr
	Ca2+ Transient / SR release	Decreased	Decreased
	SR Ca2+-uptake (SERCA2a)	Decreased	Decreased
	CaMKII-alpha	Increased	Increased
	SR Ca2+ store	Decreased	Unchanged
	SR Ca2+ leak	Decreased	Unchanged
	Na+/Ca2+ exchanger (NCX) - Cai extrusion	Increased	Unclear (contradictory data)
	Dyad remodeling	??	↓ JPH-2 & BIN-1

Note: MHC-class II, major histocompatibility complex Class II; MMP-9, anti-matrix metallopeptidase; MPC1, mitochondrial pyruvate carrier; VERP, Ventricular effective refractory period; Cav1.2, L-Type Ca²⁺ channel pore-forming subunit; ICa, L-type Ca²⁺ current; Ito, Potassium transient outward current; IKr, rapid delay potassium current; SR, sarcoplasmic reticulum; SERCA2a, Sarcoplasmic/endoplasmic reticulum Ca²⁺ ATPase-2a; JPH-2, junctophilin-2 (a dyad scaffolding protein); BIN-1, a key protein involved in Cav1.2 targeting to T-tubules; CaMKII-alpha, Ca²⁺/calmodulin-dependent protein kinase II; ??, unknown.

Cessation of tachy-pacing results in normalization of right atrial and arterial pressure with significant recovery of LVEF and cardiac output by 48 hours, and full normalization after 1 to 2 weeks (9). However, a week after resolution of tachycardia, LV mass increases by 26% and LV remains dilated and myocytes continue to demonstrate contractile dysfunction (15). Moreover, only Ca cycling (sum of Ca uptake and Ca release), Ca-uptake and CK activity significantly normalized (17) 4 weeks after cessation of tachy-pacing. Importantly, some changes such as fibrosis appear to persist despite elimination of the tachycardia and normalization of LV function (16,19).

Clinical presentation, diagnosis and imaging features

Clinical studies have found a variable time from onset of arrhythmia symptoms to development of T-CM, ranging from 3 to 120 days with an overall LVEF of 32% (14). Regardless of tachyarrhythmia, heart failure symptoms will manifest earlier at higher tachycardia rates (1,9,16), such as patients with persistent atrial flutter or tachycardia with 2:1 AV conduction with rates greater than 150 bpm. A recent clinical study found a more severe LV dysfunction (LVEF 29.3 ± 6.6%) in T-CM when compared to dilated and inflammatory CM (32.1 ± 10.2% and 41.9 ± 12.9%, respectively; p <0.001) (18).

Major reported symptoms include palpitations (29%), HF class III-IV (47%) and syncope/presyncope (12%), while the remaining may have no symptoms (10). Sudden cardiac death is uncommon but has been reported in up to 8–12% despite treatment and resolution of cardiomyopathy (14,20).

T-CM should be suspected in patients with LV dysfunction and a prior or persistent or frequent paroxysmal tachycardia without obvious etiology (Table 2). A superimposed T-CM should be considered despite underlying secondary CM (ischemic, infiltrative or toxic/drug-related) if the tachycardia is present. Thus, an ambulatory ECG monitor for at least a 2-week period is key to confirm or exclude T-CM.

Table 2.

Reversible and irreversible causes of cardiomyopathy

Reversible	Irreversible
Transient ischemia / post-cardiac arrest	Extensive / multiple myocardial infarctions
Subacute valvular heart disease	Hypertrophic cardiomyopathy
Uncontrolled hypertension	Cardiac sarcoidosis
LBBB - cardiomyopathy	End-stage valvular heart disease
Pacing-induced cardiomyopathy	Infectious (e.g. Chagas disease)
Drug or alcohol abuse
Endocrine (severe hypothyroidism)
PVC-induced Cardiomyopathy
Stress-induced Cardiomyopathy
Peripartum Cardiomyopathy
Inflammatory / Infectious (e.g. myocarditis, sepsis)

Echocardiogram or cardiac magnetic resonance can assist in excluding other etiologies. T-CM is characterized by a dilated CM (increased LV end-diastolic dimension and area) with moderate to severe biventricular systolic dysfunction and normal LV septal and posterior wall thickness (lack of hypertrophy). Mitral insufficiency may be present due to LV and mitral annular dilatation with lack of leaflet coaptation (16).

Neurohormonal markers such as brain natriuretic peptide (BNP) and pro-BNP are commonly elevated depending on the degree of heart failure and CM (14,21). Moreover, a sudden drop of pro-BNP within a week of elimination of tachycardia is supportive of T-CM (21). However, the final diagnosis of T-CM can be only confirmed after recovery or improvement of LV systolic function within 1 −6 months after elimination of the tachyarrhythmia.

Treatment

A major feature of T-CM is its reversibility once tachycardia is eliminated. Thus, the mainstay treatment consists on suppression of tachycardia based on the culprit arrhythmia (Table 3) with antiarrhythmics (AADs) and/or RFA. Nevertheless, initial treatment of T-CM, should include initiation and optimization of medical therapy for heart failure and LV systolic dysfunction (beta blockers, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, diuretics and aldosterone blockers) to optimize reverse remodeling.

Table 3.

Treatment options of T-CM based on tachy-arrhythmia.

Arrhythmia	Treatment
Sinus tachycardia / Thyrotoxicosis	BB + treatment of underlying disease
Atrial fibrillation with RVR	Rhythm (PVI +/− AADs) vs. Rate control vs. AVJ
Atrial flutter with RVR	Radiofrequency ablation
Atrial Tachycardia	Radiofrequency ablation vs. AADs
AVRT / AVNRT	Radiofrequency ablation
Rapid Atrial/Ventricular pacing (pacemaker mediated tachycardia)	Reprogram pacemaker
Sustained Ventricular Tachycardia	Radiofrequency ablation +/− AADs

Note: BB, beta-blockers; PVI, pulmonary vein isolation; AADs, antiarrhythmic drugs; AVJ, atrioventricular nodal ablation

Elimination of tachyarrhythmia not only resolves LV function within 4–12 weeks, but also improves heart failure symptoms by at least one NYHA class in most patients (9,14,16). Multivariate analysis demonstrated that age, tachycardia rate and baseline LVEF and LVEDD were predictors of recovery in a pediatric population (12). Unfortunately, the recovery of T-CM is not always complete. Histopathologic abnormalities, diastolic dysfunction and ventricular dilatation with a hypertrophic response may persist despite normalization of LVEF (15,16,18,19).

In the presence of superimposed T-CM, full reversibility of LV function is unlikely after treatment of tachyarrhythmia. Nevertheless, treatment should not be discouraged as it may have small yet significant benefits.

Tachycardia recurrence in prior history of T-CM

Studies have documented that recurrence of T-CM and heart failure symptoms upon arrhythmia recurrence sooner and with at least the same severity compared to initial presentation (14,20). It is speculated that persistence of underlying histopathological abnormalities from initial presentation are likely responsible for a more rapid and severe presentation if arrhythmia recurs (16,19). Yet, LV dysfunction recovers to prior or normal levels with elimination of new or recurrent tachycardia. Thus, a permanent treatment such as ablation therapy should be especially considered in arrhythmias with a high success or cure rate such as atrial flutter, AVNRT, AVRT and AT.

AF-induced Cardiomyopathy (AF-CM)

AF is frequently considered as a primary cause for T-CM. Limited evidence suggests that ventricular rate during AF does not predict reversibility of CM (22). This has raised the question of whether AF duration and/or irregularity predicts LV dysfunction rather than ventricular rate. More recently, AF ablation trials in HF have questioned current recommendations that rate control alone is appropriate in patients with paroxysmal / persistent AF and associated CM / HF (3,23). This is in light of the significant improvement of LV function and HF symptoms after AF ablation when compared to medical therapy, supporting the premise that AF alone can lead to CM despite appropriate rate control.

AF-CM is defined as LV systolic dysfunction in patients with paroxysmal or persistent atrial fibrillation despite appropriate rate control. Thus, an ambulatory Holter monitor is key to rule out poor rate control and T-CM. Despite AF being the most prevalent arrhythmia, the prevalence and factors that predispose or prevent AF-CM are unknown. A frequent clinical challenge is to recognize whether the AF is due to HF and cardiomyopathy or vice versa.

The mechanism of AF-CM is unknown (Central Figure). It is believed that AF-CM is triggered in part due to 1) HR irregularity with calcium mishandling (24) and 2) loss of atrial contraction/ emptying associated with sympathetic activation contributing to limited ventricular filling and increased filling pressures, functional mitral regurgitation and diastolic dysfunction (2,25). Unfortunately, no AF-CM animal models exist to better understand causality, risk factors and/or mechanism involved in its pathogenesis.

AF-CM is a diagnosis of exclusion and should be primarily suspected in patients with non-ischemic CM and persistent AF that do not improve after appropriate medical therapy and rate control. Due to the overlap with T-CM and lack of animal models, it is unclear how the time course, clinical, laboratory or imaging features differ between AF-CM and T-CM. A final AF-CM diagnosis can only be corroborated if LV systolic function improves or normalizes after elimination of AF.

Restoration of sinus rhythm should be considered if AF-CM is suspected. AF ablation has been reported to achieve sinus rhythm from 50–88% in both paroxysmal and persistent AF patients with HF and CM (2,3,23,26). While complication rate of AF ablation is low (2–3%), a second ablation is frequently required. In contrast, AADs have an overall 30–50% success rate to maintain sinus rhythm with frequent discontinuation due to side effects (2,26).Further evidence suggests that AF ablation maybe superior to AADs in AF-CM (2). Landmark AF trials with AADs have failed to demonstrate outcome benefits including HF admissions in patients with and without HF or CM (2,27,28). This is in contrast to randomized clinical studies comparing AF ablation as a rhythm control versus rate control strategy, reporting an 8–18% absolute increase in LVEF in 60–70% in patients with AF and CM randomized to ablation (2,13,23,29). The only trial comparing rhythm control strategies (ablation vs. amiodarone) in patients with AF and CM demonstrated ablation to be superior by improving freedom of AF (70 vs. 34%), quality of life, HF admissions (31 vs. 57%) and mortality (8 vs. 18%) after 2-year follow-up (26). Lastly, the CAMERA-MRI study supports the use of cardiac MRI with late-gadolinium enhancement as a screening tool to predict CM reversibility before ablation. The absence of ventricular scar or scar burden <10% predicted reversibility of AF-CM (Figure 1) (23).

Figure 1.

Representative examples of patients with AF with late gadolinium enhancement (LGE) in a patient without (A) and with AF-CM (B). Top panels demonstrate the presence (A) and absence (B) of scar in cardiac MR with late-gadolinium enhancement (LGE) in a patient without and with AF-CM, respectively. Panel C demonstrate change in LVEF from baseline stratified by the presence or absence of scar, while Panel D demonstrates correlation between % of scar (LGE) and change in LVEF from baseline after AF ablation (permission obtained Prabhu S, et al. The CAMERA-MRI Study. J Am Coll Cardiol 2017;70:1949–1961) (23).

PVC-Cardiomyopathy

Due to its unique features and lack of tachycardia, PVC-CM is now recognized as a distinct clinical entity by the most recent 2016 AHA Scientific Statement of dilated cardiomyopathies (1). PVC-CM is defined as the development of LV dysfunction caused solely by frequent PVCs. Moreover, superimposed PVC-CM can be defined as worsening of LVEF by at least 10% due to frequent PVCs in a previously known CM. Even though, frequent PVCs are commonly referred as PVC burden greater than 5%, a PVC burden ≥ 10% is often considered high and significant to trigger PVC-CM.

Epidemiology of PVCs and PVC-cardiomyopathy

The incidence of PVCs in a 10-second 12-lead ECG is estimated between 1% to 4% of patients without heart disease (30,31). However, the prevalence of PVCs is significantly higher during an ambulatory ECG recordings, (40% and 75% of participants on 24- to 48-hour ambulatory Holter monitoring) (31). This can be explained by a significant variability of PVC frequency with time (32).

Prevalence of PVCs is also age-dependent with <1% in children younger than 11 years and near 70% in subjects 75 years and older (30,31). PVCs are more frequently associated with post-myocardial infarction, coronary heart disease and dilated CM and HF (30,34). Furthermore, almost half of patients with HF class II and III had frequent PVCs (>1,000 PVCs/day) (34).

Clinical studies have found high PVC burden associated with LV dysfunction, increased risk of systolic heart failure (HR 1.48–1.8) and mortality (HR 1.31) (35–41) even after adjusting for age and other ECG abnormalities (30). Surprisingly, a six-fold increased risk (HR 6.5) of systolic HF has been reported in subjects <65 years with PVCs without other cardiovascular risk factors (42) and a higher mortality risk for those with HR >100 bpm (30). Lastly, a significant increased risk of incident stroke (HR 1.71, CI 1.14–2.59) has been reported on a secondary analysis of the Atherosclerosis Risk in Communities (ARIC) study, which the investigators attribute to a possible atrioventricular remodeling (43).

Prevalence of PVC-CM has been reported at 7% amongst patients with frequent PVC burden greater than 10% (44). Nevertheless, PVC-CM is likely underestimated (31). Clinical studies have reported a diagnosis of PVC-CM in 9–30% of patients referred for RFA of PVC (35,38,45–47). Similarly, a secondary analysis of the CHF-STAT study (48) (>10 PVCs/hour and LVEF <40%) demonstrated an estimated rate of PVC-CM of 40% in all patients with CM and up to 66% in non-ischemic CM (49). This data supports frequent PVCs as a significant and modifiable risk factor for systolic HF and increased mortality.

Acute effects of PVCs and potential triggers of PVC-CM

PVCs have acute intrinsic effects that are innate to their ectopic and premature origin, including heart rate irregularity and post-extrasystolic potentiation, LV dyssynchrony, atrioventricular (AV) dyssynchrony and increased heart rate (Table 4) (41,50–54). Furthermore, it is unclear if and how these triggers by altering hemodynamics and autonomic nervous system (ANS) contribute to the development of PVC-CM (55,56).

Table 4.

Acute effects and potential triggers of PVC-CM.

Intrinsic PVC effects / Triggers
Post-extrasystolic potentiation
Increase in contractility that follows either an atrial or ventricular extrasystole, associated with Ca2+ overload (54).
Has an inverse relationship to PVC coupling interval (prematurity): shorter PVC coupling intervals (early PVCs) have a greater intracellular Ca2+ and post-extrasystolic potentiation (54).
Increase in O2 consumption with little change in cardiac output despite reduction LVEDP (mean 13mmHg) and increase in coronary flow (54).
LV dyssynchrony
Uncoordinated contraction of the LV segments.
Cause disruption and progression of dyssynergic LV wall motion resulting in LV dysfunction (41,85,100).
Directly proportional to the PVC coupling interval: longer PVC coupling intervals (late PVCs) demonstrate a greater LV dyssynchrony (52).
Tachycardia
Mean heart rate in ambulatory ECG Holters in PVC-CM is frequently normal, likely due to compensatory pauses.
Clinical and animal studies have not found any difference in mean heart rate between PVC-CM and other etiologies(41,50,65).
AV dyssynchrony
Occurs when atrial contraction takes place against a closed AV valve during PVC, demonstrated as a prominent “a” wave in the pulmonary capillary wedge pressure (PCWP) tracing.
PVC-PCWP augmentation (defined as prominent a wave > 15mmHg) was associated with a lower LVEF (52±9 vs. 62±10%, p<0.01) and shorter coupling interval (432±41 vs. 522±54ms, p<0.0001) compared to those without PVC-PCWP augmentation despite having similar PVC burden (53).
Autonomic Nervous System (ANS)
PVCs alone have shown to acutely increase sympathetic nerve activity not only at a cardiac level but also peripherally (56)
PVCs elicit a greater neuronal response than other stimuli such as aortic or inferior vena cava occlusion (55)
PVCs with variable prematurity had the greatest neuronal response (convergent, sympathetic and parasympathetic input) when compared to early and late PVCs (55)

Post-extrasystolic potentiation together with HR irregularity, LV dyssynchrony, AV dissociation and alteration in ANS are possible triggers of PVC-CM. However, differences in cellular and electrical remodeling between PVC- and T-CM (Table 1) reinforce the argument that tachycardia is unlikely to be the sole trigger for PVC-CM.

The role of post-extrasystolic potentiation and HR irregularity could be addressed by evaluating the chronic effects of frequent PACs, which lack LV dyssynchrony. While, Pacchia et.al. (57) demonstrated that CM could be induced by atrial bigeminy, this has not been observed clinically and in other animal models (58,59)(60). These findings suggest that heart rate irregularity and post-extrasystolic potentiation play a limited role, if any, on the pathophysiology of PVC-CM.

Finally, Gerstenfeld et.al. demonstrated in a bigeminy PVC animal model that LV dysfunction is more pronounced in PVCs from LV epicardium as they demonstrate a higher degree of dyssynchrony when compared to endocardial RV free wall (59). This is also consistent with clinical studies where epicardial PVCs and QRS >150ms were predictors of PVC-CM (61,62).

Potential mechanism(s) of PVC-CM

In contrast to T-CM, the cellular mechanism of PVC-CM has not been extensively studied. Yet, it is clear that there are distinct differences in histopathological and cellular features compared to other HF models including T-CM (Central Figure, Table 1) (8,9,15,16,18,19,59,63,64).

The primary cause of contractile dysfunction in PVC-CM appears to be disorders of the calcium-induced calcium release mechanism itself, with alterations of dyad (L-type Ca channel and Ryanodine receptor) function proposed as a potential mechanism. Similar to other cardiomyopathies, this PVC-CM model has revealed electrophysiological remodeling (Table 1). Histopathological abnormalities are distinct without evidence of increased inflammation or apoptosis and minimal or no fibrosis. Mitochondrial studies have demonstrated no changes in oxidative phosphorylation (8). These findings are supported clinically by the lack of scar on cardiac MRI of patients with PVC-CM (38). These findings further confirm a primary functional abnormality as a primary mechanism of this reversible CM (8,38,59,63,64). Whether all the cellular and molecular changes are in response to the CM rather than the cause of the CM remains unclear.

Predictors of PVC-CM

Clinical evaluation of underlying effects of frequent PVCs have limitations due to a variability of PVC features (origin, prematurity, frequency, QRS width), presence of confounding patient comorbidities and small sample population.

PVC burden has been shown to be a major predictor of PVC-CM (OR 1.25 per each percent increase in PVC burden; CI 1.1–1.42) (35,45,62,65–67). Two main studies have shown that PVC burden greater than 16% and 24% best identifies patients with a diagnosis of PVC-CM (sensitivity and specificity of 79–100 and 78–87%, respectively) (35,44). While these and other studies suggest that a PVCs burden of at least 10% is required to induce PVC-CM (35,44,68–70), other studies question this minimal PVC thresholds since they have shown improvement in LV function with PVC burden as low as 6–8% (48,51,66,71–73). The length of ambulatory ECG monitoring has important implications, since increasing the duration from a 24-hr to a 7-day ambulatory Holter monitor can doubled the number of patients who reach the 10% threshold (33).

Nevertheless, some patients do not develop a CM even with high PVC burden. Thus, it is likely that other patient’s characteristics and/or PVC features play a role in the pathophysiology of PVC-CM. Some other PVC features have been found to be independent predictors for PVC-CM such as male gender (74), lack of symptoms (adjusted OR 13.1; CI 4.1–37) or duration of palpitations >30 months, variability of PVC coupling interval (OR 1.04; CI 1.03–1.07(39)), QRS duration of PVC >150ms, and epicardial origin (37,44,45,61,62,65,67,74,75) (Figure 2). Other less frequently reported independent predictors are BMI >30kg/m² (OR 3.03; CI 1.2–7.7) (39), less variability in circadian PVC distribution (OR 16.3; CI 1.7–155) (76) and presence of retrograde P wave (OR 2.79; CI 1.08–7.19) (67). AV dyssynchrony during PVCs could potentially be a predictor but yet remains to be studied (53,67). Except for PVC burden, most predictors have been variably reported reflecting the heterogeneity of different populations. Although further validation is required, a PVC-CM index, including PVC burden, PVC-QRS width and epicardial origin, has been developed in an attempt to identify patients with high probability of PVC-CM (62).

Figure 2.

Representative cases of patients similar high PVC burden from left coronary cusp without (A) and with (B) associated CM (LVEF 40%). PVC-CM has a wider PVC QRS duration (172ms) when compared to preserved LV function (150ms) (permission obtained Carballeira Pol L, et al. Heart Rhythm 2014; 11:299–306) (75). (C) Representative case of PVC-CM with dispersion of CI of 144ms (cutoff of CI dispersion >99ms best identified patients with and without PVC-CM) (permission obtained Kawamura M, et al. J Cardiovasc Electrophysiol 2014;25:756–62) (39).

Even though, short PVC coupling intervals have been associated with idiopathic ventricular fibrillation (77), most studies have not found a clear relationship between PVC coupling interval and CM (37). While some studies have reported that interpolated PVCs or coupling interval <450ms may be predictors to develop PVC-CM (68,78), others reported that PVC coupling interval variability (dispersion) is not only associated with a higher risk of PVC-CM but also cardiovascular mortality (39,79,80). A potential explanation (demonstrated in animal data) is the large cardiac neuronal disturbance demonstrated in PVCs with variable coupling interval, beyond what is seen in short or long coupling intervals (55). Animal studies are currently underway to investigate the effects of PVC coupling interval on LVEF.

PVC locations, other than epicardial origin has not shown to be predictor of PVC-CM (45,68). One of the largest series have found that PVC-CM have a PVC origin from coronary sinus (epicardial) in 24%, RV outflow in 21%, LV outflow tract in 28%, mitral annulus (endocardial) in 7%, RV/LV septum in 5% and RV/LV apex in 4% (65).

Lastly, genetic predisposition could explain why some patients are prone to develop PVC-CM while others do not despite similar PVC burden. For instance, R222Q missense variant of the Nav1.5 subunit of sodium channel, resulting in a greater and earlier sodium current, has been attributed to a rate-dependent ectopy of Purkinje and associated with a CM reversible upon treatment with amiodarone or flecainide (81).

Clinical presentation, diagnosis and imaging features

The time course for the development of PVC-CM is unclear, but it is estimated to occur within months up to several years (62,67,75). While animal studies with persistent high PVC burden (33–50%) develop CM within 4 weeks (8,50), human studies are not consistent in part due to the unclear onset and variability of PVCs.

PVC-CM may have a wide range of presentations, from asymptomatic or vague symptomatology to heart failure and even syncope. It is unclear why some patients have symptoms related to PVCs while others do not, but a PVC coupling interval ratio <0.5 (PVC CI ratio: PVC coupling interval/Sinus coupling interval) has been proposed as an important marker of symptoms (82). A careful history and pertinent testing should be obtained to rule out other causes of cardiomyopathy (Table 2), while physical examination is frequently normal with the exception of irregular heart sounds and mild signs of heart failure.

PVC-CM is a diagnosis of exclusion, to be suspected in patients with frequent PVCs greater than 10%, especially in non-ischemic CM. A challenge is to identify when PVCs are the etiology of a CM or just “innocent bystanders” in patients with CM. Even if PVCs are the result of CM, these PVCs if frequent may contribute to and further worsen CM and HF symptoms, referred to as “superimposed’ PVC-CM (73,83). In selected cases, echocardiographic and PVC features can help identify these patients (Table 5) (1).

Table 5.

Clinical and PVC features to identify PVC-Cardiomyopathy.

	CM resulting in PVCs	PVCs causing CM
Patient characteristics	Older with known heart disease	Healthy otherwise
Comorbidities	CAD, myocarditis, RV dysplasia	No prior cardiac hx
Echocardiogram	Segmental hypokinesis, LVEF < 25%	Global hypokinesis, LVEF 37 ± 10%
Cardiac MRI (late-gadolinium enhancement)	Significant scar	Absence or minimal scar burden (≤ 9gm)
PVC Frequency	< 5,000 / 24 hrs (< 5%)	≥ 10,000 / 24 hours (>10%)
PVC pattern	Multifocal	Monomorphic
QRS morphology	Non-specific	RVOT / LVOT / Epicardial
Response to PVC suppression	No change in LV function	Improvement of LV function

Even though the ECG is important, a prolonged ambulatory ECG monitor is essential to improve the diagnostic yield of high PVC burden. As noted by Loring et.al. (33), a minimum of 6 days is required to detect an individual’s maximum PVC burden. In contrast, a 24-hour Holter identified only 53% of patients with a PVC burden greater than 10%, which would likely miss almost half (47%) of patients with potential PVC-CM diagnosis (33).

PVC-CM is characterized by mild to moderate LV systolic dysfunction, LV dilatation, mild MR and LA enlargement, which resolved within 2–12 weeks after elimination of PVCs (8,41,50). Cardiac imaging is key to identify LV dysfunction and suspect the diagnosis of PVC-CM in patients with high PVC burden (≥10%) (Table 5). Cardiac MR with late-gadolinium enhancement has the advantage of identifying scar and quantify scar burden, which in turn potentially predicts response to PVC suppression (84). Few clinical and animal studies have also demonstrated diastolic dysfunction after 12-weeks of chronic ventricular bigeminy (50,85). Interestingly, a subclinical form of PVC-CM (LVEF ≥50%) has been reported using speckle tracking by a decrease in radial, circumferential and longitudinal strain that can reverse after RFA (86,87). This is further supported by translational studies demonstrating a mild and linear decrease of LV systolic function with 7, 14 and 25% PVC burden (Figure 3) (50).

Figure 3.

Progression of LV ejection fraction in a large animal PVC model after 4 and 8weeks of a progressive incremental PVC burden starting from 0% (baseline) to 7%, 14%, 24%, 33%, and 50%. (p <0.0001, 1-way ANOVA). Even though PVC-CM is not seen until PVC burden of 33%, a decline in LVEF can be noted at lower PVC burdens (Permission obtained Tan AY, et al. Heart Rhythm 2016;13:755–61) (50).

Most recently, myocarditis has been implicated as a potential trigger for frequent PVCs and CM (88), while elevated hs-CRP has been reported to be an independent predictor of PVCs in a Chinese population study (89). Thus, it may be clinically challenging to determine a causal versus bystander role of PVCs in an inflammatory process.

Treatment

Currently, a PVC suppression strategy with RFA or AADs is a widely accepted intervention to treat a CM that might be caused or exacerbated by frequent PVCs (1). However, the treatment of frequent PVCs (≥ 10% burden) without LV dysfunction (LVEF ≥50%), symptoms or idiopathic ventricular fibrillation is less clear. Due to the lack of data and potential risk of developing PVC-CM, these patients require close monitoring every 6–12 months or more closely if heart failure symptoms develop. Echocardiogram should be repeated to confirm a normal LV function, while a prolonged ambulatory ECG monitor should be considered to reassess PVC burden. Even though spontaneous resolution of frequent PVCs has not been evaluated, data from the CHF-STAT trial demonstrated that 12% of patients on placebo had spontaneous and significant decrease in PVC burden at 6 months (49).

PVC suppression is considered successful if burden is decreased by >80% of baseline PVCs as it likely represents a true effect of treatment rather than spontaneous PVC variability (90). However, this criterion was based on 24-hour Holter data and it is unclear if this is different nowadays with extended 2- or even 4-week ambulatory monitors. Current therapies, RFA and AADs, have similar long-term PVC suppression success rate between 70 to 80% (45,48,62,68,70,91). Successful RFA may be limited in patients with PVCs originating from papillary muscle, epicardium or nearby critical structures such as coronary arteries and conduction system (40,45,68). Thus, antiarrhythmic therapy may be necessary in about 5–15% of patients after RFA (68). PVC suppression strategies (RFA or AADs) carries an overall low risk. While the complication rates of RFA have been reported between 5–8%, antiarrhythmics have a discontinuation rate of near 10% due to short- and long-term side effects (45,68,84,91,92) in addition to potentially decreased efficacy overtime (93).

Randomized clinical studies of AADs have only been performed prior to the recognition of PVC-CM as a unique entity. While the CAST trial demonstrated an increase mortality with class IC agents in patients with frequent ectopy after MI, the GESICA, CAMIAT and CHF-STAT trials demonstrated at least a trend towards a decrease in mortality after MI and in non-ischemic CM with the use of amiodarone (48,94). No randomized-prospective studies has compared the efficacy and outcomes between RFA and AAD therapy. A contemporary retrospective study has shown that PVC reduction was greater with RFA than antiarrhythmics (mean reduction: RFA −15.5 ± 1.3% vs. AADs −4.8 ± o.8, p<0.001) (68). While RFA and antiarrhythmic drugs can successfully suppress high PVC burdens, a single retrospective study suggests that RFA may be more effective in patients with lower PVC burdens (68).

PVC suppression in PVC-CM has been shown to improve LV function, LV dilatation, mitral regurgitation and BNP levels (1,70). The mean improvement of LVEF after RFA in most studies is between 10 to 15% (Table 6) (45,65,68,70,91,92) even in superimposed PVC-CM. A recent multicenter retrospective study of 245 patients with non-ischemic CM and frequent PVCs (mean PVC burden 20 ± 13%) demonstrated improvement of LV function in 67% of patients after RFA (45). Similarly, a prospective study demonstrated a significant decrease in BNP levels while primary prophylaxis ICD implantation was avoided in 80% of all patients with PVC burden >13% due to significant improvement of LVEF after successful RFA (95). Interestingly, another study (40) found that 81% of patients (n=36) with abnormal baseline eGFR (<60mL/min) had significant improvement in renal function (eGFR 51 to 57mL/min) after RFA of PVCs.

Table 6.

Summary of studies of PVC-Cardiomyopathy.

							LV Ejection Fraction
Author	# Pts	Baseline PVC Burden	PVC-CM rate	Suppression strategy	CM	Successful Suppression	Pre-PVC suppression	Post-PVC suppression	Follow-up (months)	Notes/ Outcome/Predictors (P)
Singh (1995) (48,49)	336	>10 PVC/hr	40%	Amiodarone vs. Placebo (RC)*	I+NI	72% (Amio)12% (placebo)	A: 24.9 ± 8% P: 25.7 ± 8%	A: 33.7 ± 11% P: 29.2 ± 11%	45	* 15% pts in placebo had LVEF improvement (delta EF ≥10%) despite PVC suppression <80%
Duffee (1998) (6)	5	>20,000/day	--	AADs	NI	35%	27 ± 10 %	49 ± 17 %	6 ± 3
Yargadala (2005) (71)	8	17,541/day	--	RFA	NI*	87%	39 ± 6 %	62 +6 %	3	* included PVC from RVOT only
Bogun (2007) (36)	18	37 ± 13%	82%	RFA	NI	80%	34 ± 13 %	59 ± 7 %	46
Taieb (2007) (101)	6	17,717/day	--	RFA	NI	?	42 ± 2.5 %	57 ± 3.7 %	6
Sarrazin (2009) (73)	12	22 ± 12%	66%	RFA	I*	100%	38 ± 11 %	51 ± 0.1	14 ± 13	*included pts with remote hx of MI referred for IC & PVC ≥5%; NYHA improved after RFA
Baman (2010) (35)	57	33 ± 13%	81%	RFA	NI*	84%	35 ± 9%	54 ± 10%	48	* excluded CAD only; Burden (P)
Hasdemir (2011) (44)	17	29 ± 9%	--	RFA	NI*	75%	38 ± 7%	53 ± 7%	4	* excluded prior cardiac or ischemic disease
Mountantonakis (2011) (91)	69	29 ± 13%	--	RFA	NI	78%	35 ± 9%	48 ± 10%*	11 ± 6	LVEF improvement correlated with ablation outcome and decline in PVC burden.
Lu (2012) (46)	24	15 ± 6%	58%	RFA	I+NI	?	32 ± 15%	43 ± 14%	8
Ban (2013) (67)	28	26 ± 10%	75%	RFA	NI	91%	44 ± 5%	55 ± 6%	19 ± 17	Burden (P), retrograde P waves (P)
Yokokawa (2013) (47)	87	26 ± 11%	86%	RFA	NI	86%	39 ± 10%	59 ± 4%	6 – 45	32% with delayed recovery of LVEF (5–45 mo). Epicardial origin predicted delayed recovery.
Zhong (2014) (68)	121	25% (RFA) 22% (AAD)	32%	RFA vs. AAD (NR)	I+NI	86% (RFA) 49% (AAD)	42%	55%*	6. 3 ± 2.4	* EF restored 47% (RFA) vs. 21% (AAD); PVC CI <450ms (P)
El Kadri (2015) (83)	30	23 ± 8.8%	50%	RFA	NI*	60%	34 ± 15%	45 ± 17%	30 ± 28	*Included pts with scar or prior CM; NYHA improved after RFA
Panela (2015) (70)	66	21 ± 12%	26%	RFA	I+NI	76%	28 ± 4%	42 ± 12%	12	Included only LVEF <35% referred for ICD
Latchamsetty (2015) (45)	245	27 ± 13%	67%	RFA	NI	71%	38 %	50 %	20 ± 22	Burden (P), Male gender (P), epicardial origin (P)
Sadron Blaye- Felice (2016) (65)	96	26 ± 12%	--	RFA	NI*	80%	38 ± 10%	50 ± 13%	24 ± 21	*39% had HD (ischemic, HTN< valvular) with high suspicious of PVC-CM; Burden (P), Epicardial origin (P) & lack of palpitations (P)
Hamon (2016) (62)	58	23 ± 12%	54%	RFA	I+NI+SHD	91%	38 ± 9%	55 ± 9%	22 ± 15%	Burden (P), Epicardial origin (P), PVC-QRS duration (P), SHD (P)
Lee (2018) (66)	54	28% (19–44)	61%	RFA	I+NI	73%	40% (30–46)	52% (45–56)	7 (median)	Burden (P), Male gender (P)

Note: All studies were observational, except CHF-STAT study (48) which was a randomized control trial; PVC-CM rate refers to the percentage of patients with improvement of LV function (normalized or increased by ≥10–15%) after PVC suppression. I, Ischemic CM; NI, Non-ischemic CM; SHD, structural heart disease including valvular heart disease, ischemic, hypertensive and dilated CM

^*see column “notes”; P, predictors of PVC-CM only by multivariate analysis.

Successful ablation (OR 15.7; 1.4–180), myocardial scar mass <9gm (OR 0.9; 0.81 – 99) and reduction of mean PVC burden (OR 1.09; 1.01–1.16) have been shown to be independent predictors of response to RFA (84,91). This supports assessment of scar burden using cardiac MR with late-gadolinium enhancement to predict responders versus non-responders to PVC suppression (Figure 4). However, if significant PVC burden reduction (>20%) is achieved, the presence of myocardial scar seems to be less relevant to predict response (84). In contrast, PVC location does not appear to predict improvement of LVEF (92). Recently, post-extrasystolic potentiation assessed by invasive BP monitoring has been described as a predictor of LV function recovery after radiofrequency ablation (96).

Figure 4. Representative PVC-cardiomyopathy.

A 53-year-old man with 21% PVC burden and LVEF of 40% underwent PVC ablation. Successful ablation was achieved at the mid-septal RVOT just above pulmonary valve with a PVC burden of 1.5% after RFA. (A) 12-lead ECG of representative PVC. (B) Baseline CMR demonstrate absence of scar with late-gadolinium enhancement. LVEF normalized to 55% after 3 months, diagnostic of PVC-CM.

Clinical studies have consistently demonstrated a significant increase in LV function after PVC suppression with antiarrhythmics, ranging from 10 – 13% (48,68). Because the CAST trial included patients with ischemic CM, the guidelines discourage the use of Class IC antiarrhythmics not only in ischemic but also in non-ischemic CM patients. Most recently, a small retrospective study (93) demonstrated that flecainde and propafenone (Class IC) can also improve LV function (LVEF from 37.4 ± 2 to 49 ± 1.9%), even without an 80% PVC suppression (PVC burden from 36.2 ± 3.5% to 10 ± 2.4%) without an increase in ventricular arrhythmias and/or death (93).

While current literature supports improvement of LV function and symptoms by elimination of PVCs, there is limited data that PVC suppression will subsequently modify the risk of cardiovascular events including heart failure and death (97,98). Over the past few years it has become clear that comparative effectiveness trials are needed to understand how to best treat patients with frequent PVCs and CM (95,97). Currently, a pilot multicenter study “Prospective Assessment of PVC Suppression in Cardiomyopathy: A pilot study” (PAPS: Pilot) is ongoing to better understand the prevalence of frequent PVCs and CM, prove feasibility of a large-scale randomized clinical trial (ClinicalTrials.gov Identifier: NCT03228823). Focused studies to understand the prevalence of PVC-CM are key to provide a better perspective of the magnitude of this clinical entity and potential impact in the HF population.

Clinical significance of arrhythmia-induced CM

Tachycardia, AF and PVCs are highly prevalent in patients with CM and HF, and they must be considered as potential cause of a HF and CM (Figure 5). While, it is unclear why some patients with high arrhythmia burden (frequent HR > 100bpm and/or PVC burden ≥ 10%) do not develop CM, these patients are at risk and should undergo close monitoring every 6–12 months or sooner if symptoms develop. The risk factors for developing PVC-CM include male gender, lack of symptoms, and PVCs QRS duration >150ms, epicardial origin and/or variable coupling interval, As any other CM, AiCM can lead to heart failure admissions and implantation of defibrillators and resynchronization devices (51,73,95,99). Thus, AiCM carries a significant financial burden if untreated, which makes diagnosis and treatment paramount to improve morbidity and potentially decrease health care costs. A better understanding of the mechanism of tachycardia and PVC induced CM could lead to novel therapies to prevent and improve outcomes especially when antiarrhythmics or RFA are not feasible or unsuccessful.

Figure 5. Proposed management of potential Arrhythmia-induced CM.

Footnotes: (~) Consider following the algorithm even if CAD is documented or worsening of prior CM is noted (superimposed AiCM). (^) Two-week ambulatory Holter is preferred as increases the diagnosis yield of high PVC burden (≥10%). (*) Consider cardiac MR to assess scar burden and predict response to PVC suppression. Short-term observation is reasonable for PVC-CM as 15% of cases may improve without PVC suppression strategy (49). (©) Continue close surveillance and HF med Rx in those with abnormal LV dimensions and presence of scar (cardiac MRI)

Summary / conclusions

Arrhythmia-induced CM, a reversible CM, has a significant variety of presentations from asymptomatic to severe heart failure symptoms. Clinicians should have a high index of suspicion of superimposed AiCM even in patients with an obvious etiology (Figure 5). T-CM should be strongly considered in patients with paroxysmal or persistent SVTs, primarily atrial fibrillation / flutter, atrial tachycardia and persistent junctional reciprocating tachycardia with heart rates above 100 bpm. AF-CM should be suspected in patients with non-ischemic CM and paroxysmal, persistent or permanent AF even with appropriate rate control, while, PVC-CM should be considered in patients with non-ischemic CM and PVC burden equal or greater than 10%. Appropriate diagnosis and treatment of AiCM will not only reverse LV dysfunction with its associated morbidity, mortality and healthcare spending, but most importantly improve quality of life and long-term prognosis. Future clinical studies are needed to compare standard treatment strategies and identify best long-term PVC suppression and prevention of recurrence of PVC-CM.

Highlights.

• Tachycardias, atrial fibrillation, and premature ventricular contractions are known to trigger a reversible dilated cardiomyopathy.

• Arrhythmia-induced cardiomyopathy should be highly suspected in patients without an obvious etiology.

• Ambulatory ECG monitors are key to screen and properly diagnose arrhythmia-induced cardiomyopathy.

• Reversal of cardiomyopathy by elimination of arrhythmia not only confirms the diagnosis but may significantly improve outcomes.

Abbreviations

AADs

Antiarrhythmic drugs

AiCM

Arrhythmia-induced cardiomyopathy

AFiCM

Atrial fibrillation-induced cardiomyopathy

ANS

Autonomic nervous system

AT

Atrial tachycardia

CM

Cardiomyopathy

PJRT

Permanent Junctional reciprocating tachycardia

PVC-CM

Premature ventricular contraction-cardiomyopathy

RFA

Radiofrequency ablation

T-CM

Tachycardia-induced cardiomyopathy