Anatomical Determinants of Papillary Muscle Arrhythmias in Apparently Normal Hearts

Abstract

Background:

Myocardial connections of left ventricular (LV) papillary muscles (PM) are determinants of QRS variability in the case of PM arrhythmias. We investigated the anatomical substrate of monomorphic versus polymorphic LV PM arrhythmias in patients with apparently normal hearts, as well as ablation outcomes.

Methods:

Thirty-two patients were eligible for analysis. Thirteen patients underwent ablation. With advanced cardiac imaging (cardiac MRI or multidetector CT), we determined the number of PM–PM and PM–surrounding myocardium connections, PM architecture according to the number of strands and the level of LV trabeculation.

Results:

Combinations of unifocal (monomorphic premature ventricular complexes [PVCs]), multiform PVCs and/or runs of polymorphic PM arrhythmias (≥3 beats) were recorded in 24 patients. The remaining eight patients had only unifocal monomorphic PVCs. The mean [± SD] number of PM connections was higher than that of PM–PM or PM–myocardial connections in patients with multiform PVCs (30 ± 1.5 versus 4 ± 1, respectively) or polymorphic arrhythmias (136 ± 4 versus 26 ± 3, respectively; p=0.004). Compared with the unifocal group, the frequency of multistranded PMs was higher (1 versus 22, respectively; p<0.001) and LV trabeculation was more pronounced in the group with multiform arrhythmia (multiform PVCs and/or polymorphic arrhythmias). All patients ablated for unifocal PVCs remained free of recurrence, compared with only half of those ablated for multiform PVCs.

Conclusion:

Patients with multiform PM arrhythmias have more PM connections, PM strands and trabeculation than patients without QRS variability. The long-term effectiveness of catheter ablation in this patient group is limited.

The papillary muscles (PMs) of the left ventricle (LV) play an important role in normal cardiac function.1 When the heart is viewed in an attitudinally appropriate manner, rather than in cardiocentric terms, the superolateral PM (SLPM) originates from the apical to middle thirds of LV free wall and the inferoseptal PM (ILPM) originates from the diaphragmatic myocardium.1–4

Ventricular arrhythmias originating from the papillary muscles (PMs) in the LV often exhibit QRS variability (i.e. presenting as multiform premature ventricular complexes [PVCs] or polymorphic runs).5–11 Multiple anatomical variants of PMs have been identified in human hearts.12–14 Myocardial connections of an LV PM to the opposite PM or to the surrounding myocardium are frequent sites of the origin and determinants of such multiform PM arrhythmias (PVCs and/or polymorphic runs).15 Results of catheter ablation in patients with multiform PM arrhythmias appear to be suboptimal, which may be explained by the presence of multiple PM connections.15,16 Currently, several aspects regarding PM connections and arrhythmogenesis remain unclear. For example, it remains to be elucidated whether the presence and number of PM connections differs between patients with unifocal PVCs and those with multiform PM arrhythmias, or whether PM connections are part of a complex PM architecture and/or the endocardial trabecular network.

The present study provides novel insights into the anatomical determinants of LV PM arrhythmias with and without QRS variability in patients with apparently normal hearts. We focus on the relationship between PM connections and structural arrangement of the LV. The long-term outcome of catheter ablation guided by intracardiac echocardiography was also investigated.

Methods

Study Protocol

The study protocol was approved by the Internal Board of Review of The Anchorena Clinic and Accord Health Medical Center, Buenos Aires, Argentina. From a total of 763 patients treated for ventricular arrhythmias between 2018 and 2023, 58 presented with arrhythmia originating from LV PMs, as determined by non-invasive characteristics in all patients and invasive electrophysiological confirmation in those undergoing catheter ablation (n=13). Patient selection for ablation treatment was based on the PVC burden (>10%), related symptoms and ineffective antiarrhythmic drug therapy or drug intolerance, in compliance with European Society of Cardiology (ESC) guidelines for the management of patients with ventricular arrhythmias.17,18

After excluding patients with right ventricular PM arrhythmias, impaired LV ejection fraction (LVEF), coronary artery disease, mitral valve disease, structural heart disease or specific cardiomyopathies, 32 patients were eligible for analysis. Eight of these patients had unifocal PM arrhythmias and 24 had multiform PM arrhythmias.19 Patients with specific PM anatomical variations, such as single-base and accessory PMs (n=3 patients), were excluded from the study because these variations may influence the electrocardiographic characterisation of PM arrhythmias.20

Definitions

PM arrhythmias from the LV were defined as those with: a right bundle branch block pattern with a typical rSR or monophasic R, Rr, RR or qR pattern in lead V1; a QRS duration ≥130 ms; early precordial transition between V3 and V5; and QRS notching during PVCs.5–11,18,21

Unifocal PVCs were defined as arrhythmias with a single QRS morphology arising from the PM (Figure 1). Multiform PM arrhythmias were defined as those with QRS variability, determined by width, axis, notching and/or amplitude variation of the QRS complex between beats, and presenting as multiform PVCs or polymorphic runs, in accordance with the ESC guidelines.17

Figure 1: ECG Characteristics of Unifocal and Multiform Papillary Muscle Arrhythmias.

A: Unifocal PM arrhythmia as a PVC determined on the basis of a single QRS configuration without variability. B: Multiform PM arrhythmia, with QRS variability observed in the lead I amplitude (1), V1 pattern (2) and precordial transition zone (3). C: Non-sustained polymorphic VT originating from the LV PMs showing variability in the QRS amplitude, duration and coupling interval. PM = papillary muscle; PVC = premature ventricular complex; VT = ventricular tachycardia.

To determine the possible anatomical substrate in patients with PM arrhythmias, we focused on:

• The presence and total number of PM connections.

• The arrangement of the PM base, categorised according to the number of strands building the PM structure as either a solid PM (single strand), a bifid (two strands) PM or a multistranded (three or more strands) PM. Only strands of equal diameter were considered to differentiate them from the LV trabecular network and/or PM connections.

• The degree of LV trabeculation, assessed using established criteria.22–24

These variables were analysed in patients with unifocal and multifocal PM arrhythmias. Individuals without PM arrhythmias and no cardiomyopathies were assigned as the reference (control) group with normal anatomy. The presence of solid or bifid PMs, the absence of excess trabeculation and high counts of myocardial PM connections were considered normal.1-4

Catheter-based Electrophysiological Evaluation Using Intracardiac Echocardiography and Ablation

Catheter ablation was performed in patients under conscious sedation. Arrhythmia induction was attempted by programmed electrical stimulation, with the addition of isoproterenol infusion if necessary. IV heparin was administered to maintain an activated clotting time of ≥300 seconds.

Activation and pace mapping were performed in 13 patients (10 with multiform and 3 with unifocal PM arrhythmias) using intracardiac ultrasound (ICE) integration (CARTO 3 CARTOSOUND; Biosense Webster).9 Catheter position, contact and stability were assessed through ICE integration into the mapping system. Catheter stability was defined as the absence of back-and-forth movement of the catheter during energy delivery at the effective lesion site. High-density maps with small multielectrode catheters were not created to avoid mechanically induced ectopy, which confounds interpretation.

Radiofrequency energy was delivered at myocardial sites exhibiting the earliest bipolar activity or local unipolar QS pattern, or where Purkinje activity preceded the surface QRS onset by ≥25 ms during clinical arrhythmia at pace-mapping areas exhibiting a QRS match ≥95%. Ablation was performed with a 3.5 mm open-irrigated contact force-sensing radiofrequency ablation catheter (Thermocool SMART Touch SF; Biosense Webster).

Transaortic or transmitral access was chosen when the clinical arrhythmia originated from the inferoseptal (ISPM) or the superolateral (SLPM) PM, respectively. When a reduction in the incidence of ventricular tachycardia (VT) or PVCs was observed, radiofrequency energy was delivered for up to 90 s (45 W), with two posterior 45-s consolidation lesions in the same area; otherwise, energy delivery was terminated and the catheter was repositioned.

The endpoint of catheter ablation was the elimination and non-inducibility of ventricular arrhythmias during isoproterenol infusion (2–10 µg/min) and burst pacing from the right ventricle to a RR cycle length as short as 300 ms.

Advanced Cardiac Imaging

Cardiac MRI (CMR) was obtained in all patients unless contraindicated (n=7). In patients in whom CMR was contraindicated, multidetector cardiac CT was performed. Each patient was evaluated for possible anatomical connections of the PMs using an in-plane spatial resolution of 8 mm in cine sequences.15 To this end, the long and short axes of the LV were systematically examined. PMs were analysed in the long-axis view, where the muscles were deployed at their maximum length, from the site of chordae insertion to base insertion at the LV wall. Images were analysed by two cardiologists (MPR and RR) trained in cardiac imaging, who were blinded to the results. PM connections were evaluated in both the short and long axes. Some connections exhibited oblique trajectories, which created some difficulty in following the true trajectory in the short-axis view. In such cases, long-axis views helped determine the full trajectory, which usually exhibited an X or Y pattern. An interobserver agreement analysis was performed for the described anatomical variables, which were repeatedly measured by each observer.

LV trabeculation was assessed using the following two diagnostic criteria: a non-trabeculated (NT)/trabeculated (T) ratio ≥2.3; and a trabeculated LV mass ≥20%. The NT/T ratio was calculated in the most-trabeculated myocardial segments, excluding the apex to avoid overestimation.22–24

Control Group for Comparison

A control group, without PM-related arrhythmias, was established for comparison of LV trabeculations, PM strands and papillary muscle connections. The control group comprised patients without structural heart disease and right ventricular outflow tract arrhythmias (n=31) who had undergone routine CMR to assess the specified anatomical variables between 2023 and 2024.

Follow-up

All patients were monitored continuously for 24 h after the ablation procedure. Electrocardiography and echocardiography were performed before discharge and during follow-up. Patients underwent 24-h Holter monitoring and baseline electrocardiography before and 1, 3 and 6 months after the procedure. Successful long-term catheter ablation was defined as a significant reduction in (by ≥85%) or the absence of the clinical arrhythmia during follow-up.8

Statistical Analysis

Normally distributed continuous variables are presented as the mean ± SD, whereas those with a skewed distribution are presented as the median with interquartile range (IQR). Categorical variables are presented as numbers and frequencies. Group means for continuous variables with a normal and skewed distribution were compared using the Wilcoxon rank-sum test. Categorical variables were compared using the Chi-squared test or Fischer’s exact test, as appropriate. The log-rank test was used to evaluate the equality of survivor functions. Inter- and intrarater reliability was assessed by the kappa coefficient. If kappa was not appropriate due to a category imbalance, the simple agreement percentage was used as the concordance metric. All statistical analyses were performed using SSPS Statistics v24 (IBM.). The level of statistical significance was set at p<0.05 (two-tailed).

Results

For all 32 patients with PM arrhythmias, the mean LVEF was 60 ± 5%. No myocardial fibrosis was detected by CMR. Multistranded PMs were observed in 23 of 32 patients. Myocardial connections were observed in all patients, with a mean number of 9 ± 3 connections per individual.

Regarding trabeculation, the longest trabecula measured in the LV of each patient was 11 ± 3 mm. The mean ratio of trabecular length to regional LV wall thickness was 2.3 ± 0.4, and the mean trabecular mass was 20 ± 4%.

Patients in the control group had a mean LVEF of 61 ± 3% and no myocardial fibrosis. Multistranded PMs were observed in 3 of 31 (9.7%) patients in the control group, compared with 68% of patients with PM arrhythmias. The mean number of myocardial connections in the control group was 2 ± 1 per individual, almost one-fifth the number for patients with PM arrhythmias. The longest trabecula measured in the LV of each patient in the control group was 8 ± 3 mm, the mean trabecular length to regional LV wall thickness ratio was 1.7 ± 0.6 and the mean trabecular mass was 12 ± 4%.

Medical Treatment

Asymptomatic patients (n=18) with a PVC burden <10% and no (nonsustained) VT (n=5) were followed-up for LVEF impairment. Asymptomatic patients with a PVC burden <10% but documented non-sustained VT (n=1) or a PVC burden >10% (n=12) were treated with beta-blockers. Symptomatic patients (n=14) with a PVC burden <10% (n=1) were started on betablockers, whereas those with (non-sustained) VT or a PVC burden >10% (n=13) were initially treated with a combination of beta-blockers and flecainide.

Unifocal Versus Multiform Papillary Muscle Arrhythmias

The clinical and anatomical characteristics of patients with unifocal and multiform PM arrhythmia are presented in Table 1. Patients with unifocal PM arrhythmias had fewer (non-sustained) VTs. Of the patients with PM arrhythmia, 70% exhibited QRS variability (i.e. multiform PVCs or polymorphic tachycardia).

Table 1: Clinical and Anatomical Characteristics of Patients With Multiform and Unifocal Papillary Muscle Arrhythmias.

	Multiform PM Arrhythmias (n=24)	Unifocal PM Arrhythmias (n=8)	p-value
Age (years)	46 ± 15	46 ± 19	0.8
Female	11 (46)	4 (50)	1
LVEF (%)	61 ± 5	61 ± 5	1
PVCs as clinical presentation	22 (92)	7 (88)	1
Non-sustained VT	20 (83)	1 (14)	<0.001
VT	3 (13)	1 (14)	1
Arrhythmia origin
Inferoseptal PM	22 (92)	7 (88)	1
Superolateral PM	12 (50)	1 (14)	0.1
Both*	11 (46)	0	0.03
No. PM strands†	3 [3–3.75]	2 [2–2]	<0.001
Multistranded PM base‡	22 (92)	1 (14)	<0.001
Mean no. PM connections per patient§	11 ± 3	5 ± 3	<0.001
Total no. PM connections (all patients)	136 ± 4	26 ± 3	<0.001
Total no. SLPM connections (all patients)	58 ± 1	9 ± 4	<0.001
Total no. ISPM connections (all patients)	78 ± 1	11 ± 1	<0.001
Mean no. PM–PM connections per patient	3.3 ± 1.8	0.9 ± 1.5	0.002
Trabecular LV mass (%)	20 ± 4	14 ± 5	<0.001
Index LV non-trabecular mass (%)	46 ± 8	45 ± 5	0.7
Mean trabeculation ratio¶	2.5 ± 0.4	1.3 ± 0.4	<0.001
Mean trabeculation length (mm)	11 ± 2	7 ± 2	<0.001

Unless indicated otherwise, data are given as mean ± SD, median [interquartile range] or n (%). *Arrhythmia foci present at both papillary muscles (PMs) of the LV. †Mean number of strands for each PM. ‡Three or more strands per PM. ^§Mean number of all types of PM–myocardium connections per patient. ^¶Mean ratio between the highest trabecular mass in the LV and the wall thickness at the same position. LVEF = LV ejection fraction; PVC = premature ventricular complex; VT = ventricular tachycardia.

Arrhythmias originating from both PMs were more prevalent in the multiform than unifocal PM arrhythmias group (61% versus 0%, respectively, for SLPM [p=0.02]; 48% versus 0%, respectively, for ISPM [p=0.03]).

The overall number of PM connections was significantly higher in the multiform than unifocal PM arrhythmias group (30 ± 1 versus 4 ± 1, respectively, for PM–PM connections [p<0.001]; 136 ± 4 versus 26 ± 3, respectively, for PM–myocardial connections [p=0.004]).

Multistranded PMs were observed in 22 patients with multiform arrhythmia, compared with one patient with unifocal PM arrhythmia. The median number of muscle strands for each PM was significantly higher in patients with multiform PM arrhythmias than in those with unifocal PM arrhythmias (3 [IQR 3–3.75] versus 2 [IQR 2–2], respectively; p<0.001).

LV trabeculation was also more pronounced in the group with multiform arrhythmias than in the group with unifocal arrhythmias, with the trabecular ratio being 2.5 ± 0.4 versus 1.3 ± 0.4, respectively (p<0.001), and the trabecular mass being 20 ± 4% versus 14 ± 5%, respectively (p<0.001).

Inter- and intra-observer agreement was high (Supplementary Table 1).

Bidirectional Papillary Muscle Arrhythmias

Thirteen patients in the multiform arrhythmia group had bidirectional PM arrhythmia (Figure 2). Bidirectional PM arrhythmia was defined as arrhythmia originating at the PMs, exhibiting at least two consecutive beats with an alternating QRS axis and lead I polarity during the same arrhythmia run, suggesting activation of both PMs. Only couplets or nonsustained VT were observed spontaneously. Bidirectional PM arrhythmias in these patients represent a novel finding, correlating with connections between the PMs (Figure 3).

Figure 2: Bidirectional Papillary Muscle Arrhythmia.

A 12-lead ECG showing bidirectional ventricular couplets in a patient without structural heart disease and preserved ejection fraction. The QRS morphology of the first beat ‘A’ is compatible with an inferoseptal papillary muscle origin. The second beat ‘B’ suggests a superolateral papillary muscle origin. The coupling intervals were 400 ms.

Figure 3: Bidirectional Papillary Muscle Arrhythmia in the Presence of a Papillary Muscle–Papillary Muscle Anatomical Connection.

A: CARTOSOUND reconstruction and activation map of both PMs of the LV showing the exit sites (asterisks) of premature ventricular complexes ‘A’ (PM base) and ‘B’ (PM head) from the same patient as in Figure 2. B: Cardiac magnetic resonance long-axis view in the same patient: there is a PM–PM connection, exhibiting a complex multiple connection network (black arrow), and both PMs have a multistranded base (yellow arrows). ISPM = inferoseptal papillary muscle; PM = papillary muscle; SLPM = superolateral papillary muscle; LAT = local activation time; T.Time = time component associated with local activation data.

Catheter Ablation

Twenty-six patients were initially treated with antiarrhythmic drugs. Half the patients either did not respond to the drugs or were intolerant. Thirteen patients underwent catheter ablation.

Bipolar signals had a mean ventricular electrogram to QRS activation time of 32 ± 7 ms and a 96 ± 2% pattern-matching average of the clinical arrhythmia. Unipolar signals exhibited a QS pattern at the earliest activation site in all patients. The mean contact was 13 ± 2 g and 4 ± 1 g for base and non-basal regions of the PMs, respectively. During electrophysiological studies, 8 (61%) patients exhibited pre-potentials at the earliest activation site, five of which were low-amplitude (far-field) signals and three presented sharp (Purkinje-like) electrograms.

The overall long-term success was 62%. Catheter ablation was performed in 10 patients with multiform arrhythmias and in three with unifocal PM arrhythmias. The overall success of catheter ablation in patients with multiform arrhythmias was 50%, compared with 100% in patients with unifocal PM arrhythmias, who also experienced no recurrence over a mean follow-up period of 52 ± 18 months.

Control Group

Patients in the control group had similar anatomical characteristics to patients with unifocal PM arrhythmias. Specifically, in the control group, the overall number of PM connections was 64 ± 1 (20 ± 0.6 for SLPM and 44 ± 1 for ISPM). The mean number of PM–PM connections (found in only six patients) was 0.2 ± 0.5 per patient. Multistranded PMs were seen in three individuals. The median number of muscle strands in each PM was 2 [IQR 2–2]. The results for the PMA and control groups are presented in Supplementary Table 2.

Discussion

Patients without PM arrhythmias (control group) exhibit mostly single or bifid PM configurations. Multistranded PMs were observed in only 10% of patients in the control group. PM–myocardial connections were seen in the control group, but at low numbers per individual, almost one-fifth the number of PM–myocardial connections in patients with PM arrhythmias. In addition, patients in the control group did not meet any criteria of excess trabeculation (Supplementary Figure 1).

Patients with PM arrhythmias from the LV had a higher number of anatomical connections of the PMs than individuals without PM-related arrhythmias (control group). Moreover, patients with multiform arrhythmias from the LV PMs had a higher number of PM–PM and PM–surrounding myocardium connections than those with only unifocal arrhythmias (Supplementary Figure 2).

The more complex PM anatomy of patients with multiform arrhythmias also correlated with the presence of a multistranded base structure of the PMs and more prominent LV trabeculation (Figure 4). The anatomical findings of patients with PMA were not seen in the control group (patients without PM-related arrhythmias). The occurrence of multiform PM arrhythmias based on a more complex PM architecture decreases the effectiveness of catheter ablation in this group given our suboptimal follow-up results, whereas the absence of multistranded PMs, excess trabeculation and a low number of PM connections account for less complex arrhythmias and better ablation outcomes.

Figure 4: Anatomical Substrate of Unifocal Versus Multiform Arrhythmias.

The left-most ECG trace shows a multiform PM arrhythmia appearing as couplet of bidirectional PVCs. The first beat presents a QRS morphology compatible with an inferoseptal PM origin, whereas the morphology of the second beat suggests a superolateral PM origin. The CMR on the left show typical anatomical findings: PM–PM connections (top; white arrow); multistranded PMs (middle; black arrows); and prominent trabeculation of the LV (bottom; black arrow). The right-most ECG trace shows uniform PM arrhythmia. The trace shows a single-morphology PVC, which suggests an origin at the inferoseptal PM of the LV. The CMR on the right show the absence of PM–PM connections (top), a bifid posteromedial PM and solid superolateral PM (middle) and unremarkable ventricular trabeculation (bottom). CMR = cardiac MRI; PM = papillary muscle; PVC = premature ventricular complex.

Recent studies have proposed that complex multiform QRS morphologies of PM arrhythmias are often due to PVCs originating from different sites or PVCs from a single origin that are conducted via PM branches, with multiple arrhythmogenic mechanisms.15,16,25 However, in those studies, the authors did not correlate the anatomical substrates of PM structure, connections and relationship with the trabecular network to the specific pleomorphism of PM arrhythmia.15,16,25

Moreover, the study populations included patients with and without heart disease, whereas the specific aim of the present study was to highlight the importance of cardiac anatomy for elucidating the electroanatomical mechanisms of multiform PM arrhythmias in patients with apparently normal hearts (i.e. those in whom arrhythmogenesis cannot be explained by ischaemic or non-ischaemic cardiomyopathy).

Considerations

Excessive trabeculation is a ventricular phenotype identified most frequently by echocardiography and CMR. CMR has the advantage of greater contrast resolution and blood–muscle differentiation, enabling better visualisation of ventricular trabeculation criteria.26

Trabeculations are strut-like myocardial structures that connect to each other to form a complex network in the luminal part of the cardiac chambers. Trabeculations give rise to the PMs, the trabeculae carneae and the Purkinje network.27 The base of the PMs may join to the trabeculae carneae lining the ventricular cavity rather than directly to the solid portion of the ventricular wall (Supplementary Figure 3).3 This mesh-like attachment to the wall may reduce stress concentration in the wall near the base of the PMs.3 Multiple points of attachment for the PMs may provide some protection against mechanical failure.3

Although excessive trabeculation can be observed in different types of cardiomyopathies and congenital heart diseases, it also occurs in a substantial proportion of the healthy population.27 Trabeculations can recede when the underlying disease is treated or the trabecular layer can widen when the underlying disease progresses.27 The presence of excessive trabeculation itself does not indicate the presence of cardiomyopathy.26

Although excessive trabeculation has been linked to an increased risk of ventricular arrhythmias, most arrhythmias occur remote from the excessively trabeculated area, from the right ventricular outflow tract or the basal perivalvular LV.26 Late gadolinium enhancement revealing myocardial fibrosis, a well-known proarrhythmic factor, is typically low in the setting of excessive trabeculation.27 In asymptomatic individuals with excessive trabeculation, there is no relationship between the degree of trabeculation observed and diffuse fibrosis.26 Nevertheless, if trabeculations become fibrotic cords in failing hearts, with endocardial and subendocardial fibrosis, filling could be restricted by the increased stiffness of the trabecular layer, decreasing its compliance and consequently decreasing compliance of the total LV cavity. In these settings, the trabecular layer may be the substrate of subendocardial fibrosis and may have a negative impact on compliance and ventricular filling.27

Conceptually, the trabecular network offers a more tortuous and longer route between two distant positions in the LV than does the compact wall, and this may prolong activation time. Conduction delay could result from the small additional conduction time required for the activation wave front to reach the PMs through the somewhat more circuitous path via the trabeculae rather than directly from the wall.3

Although the complex myocardial architecture of excessive trabeculation may intuitively be linked to a propensity for re-entrant tachycardias, there is no convincing evidence to substantiate this.26 The present study provides further information on these aspects.

Possible mechanisms involved in the genesis of multiform PM arrhythmias include: impulse conduction from one PM to another via anatomical connections, resulting in multiple exit sites; abnormal impulse formation from the papillary muscle connection; and mechanical stretching by one PM of another.28 The resultant mechanical deformation could elicit various arrhythmia morphologies by mechanoelectrical coupling in the presence of an anatomical connection. The latter has been demonstrated in animal models, where mechanically induced PM traction can alter myocardial electrophysiological characteristics, inducing local premature ventricular activation by mechanical distortion and prolonging refractoriness in the traction zone.29 Presumably, mechanically induced premature ventricular activation may result from stretch distortion of either the PM itself or the Purkinje fibres. Transmembrane potential depolarisation caused by myocardial stretch has been associated with both transient afterpotentials and propagated action potentials.29 Myocardial stretching of the LV PMs may give rise to ventricular arrhythmias either through re-entry or triggered activity in patients with mitral annular disjunction (MAD) and mitral valve prolapse (MVP). Even patients without mitral regurgitation exhibit pleomorphic PVCs more commonly than the general population.30 Although none of the patients in this study presented with MAD, the significantly higher number of PM connections and excess trabeculation in patients with multiform PM arrhythmias, even without MVP and/or MAD, may also provoke mechanical stretch of the myocardium and arrhythmogenesis.

The concept of multiple impulse propagation pathways, Purkinje fibre involvement and/or mechanoelectrical coupling over labyrinth-like structures in the LV sets a new standard for mapping and catheter ablation. Although some PM ventricular arrhythmias can be mapped at one PM, the arrhythmogenic focus can also be located at the contralateral PM, so simultaneous mapping of both PMs is useful in specific cases, notably patients with documented arrhythmias from both PMs, and an indication for ablation. In this setting, precise anatomical evaluation is recommended.

Limitations

Most anatomical variables described in this study are not part of a standard cardiac imaging study and specific training may be required to depict them.

Invasive electrophysiological characterisation of the PM connections was not attempted in this study. Avoiding undesired recordings or capture of the surrounding myocardium is extremely challenging because, to date, most catheters are not suitable for recording and/or pacing such smallsized structures. Unstable catheter position, large electrode size and the lower tissue resolution of ICE may disable such fine discrimination because these can lead to large-field tissue capture, mechanically induced stretching and/or proarrhythmic effects through bumping. Although ICE has a better temporal and spatial resolution, CMR represents the gold standard in tissue resolution. Balanced steady state free precession (cine) sequences have a higher contrast resolution than ICE, discriminating endocardium and trabeculae from the blood pool.

We were able to depict PM connections by ICE imaging and integrate these into electroanatomical mapping systems. Nevertheless, we could not achieve a proper PM connection count using this method alone due to low resolution and a lack of full 3D visualisation, in agreement with a previous report.12 PM connections exhibit different anatomical patterns.

Usually, these are small structures exhibiting web-type configurations. Intracardiac ultrasound at 5 MHz, as used in the present study, does not have enough resolution to depict such structures appropriately.

This study was based on a retrospective analysis. Cardiac CT, which has better spatial resolution, was not available for all patients. CMR was chosen over CT to rule out the presence of cardiomyopathy (which was an exclusion criterion in this study) and to evaluate the presence of PM fibrosis (also an exclusion criterion) using late gadolinium enhancement. In addition, we used CMR to reliably determine the myocardial origin of PM connections given the higher tissue resolution of CMR over CT, evaluate myocardial strands and analyse the trabecular network, with the CMR performed in Accord Health Medical Center & Anchorena Clinic in accordance with previously described cardiac imaging methods.22–24 PM connections were assessed as described previously.15

Future studies are needed to further our understanding of arrhythmogenesis in the LV PMs. The present study provides the impetus to investigate stretch-induced, in addition to electrophysiological-only, mechanisms.

Conclusion

Patients with multiform PM arrhythmias have significantly more anatomical PM connections than those without such arrhythmias. Often, these patients also had a multistranded base structure of PMs and more prominent LV trabeculation in an otherwise apparently normal heart. Catheter ablation outcomes are poor in these settings.

Clinical Perspective

• PM–myocardial connections may be present in patients with LV PM arrhythmia and account for QRS variability during arrhythmia.

• A higher number of PM connections is associated with a multistranded PM base structure and excess LV trabeculation.

• In more complex PM arrhythmias (i.e. multiform PVCs or polymorphic runs), the number of PM connections is increased.

• Ablation results are categorical, suggesting that patients with PM arrhythmias are at risk of therapeutic failure if they have more PM connections.

Data availability:

The data that support the findings of this study are available in the article and/or supplementary material for this article.