The association between mitral valve disease and sudden cardiac death has drawn increased attention. Typical characteristics of patients include female gender, Barlow’s disease, inferior T wave inversion, and polymorphic/right bundle branch block type ventricular arrhythmias.(1-4) Moderate-to-severe mitral regurgitation (MR) has been reported to be an independent risk factor for arrhythmias.(5, 6) Despite these findings, there are limited data on the prevalence of ventricular arrhythmia in this cohort due to the absence of longitudinal studies(7). The effect of surgery on PVC burden has also not been clearly delineated and may provide insight into the arrhythmogenic substrate.

Premature ventricular contractions (PVCs) are the most common ventricular arrhythmia in patients with mitral valve disease.(8) Studies based on 12-lead ECG and Holter recordings have reported that PVCs with increased QRS duration,(9) short coupling interval, highly variable coupling interval(10-12) and polymorphic origin(13) are associated with cardiomyopathy, heart failure and ventricular tachyarrhythmia in other cohorts. Invasive electrophysiological studies in mitral valve disease patients have identified the most common sites of origin (SOO) of PVCs to be the inferobasal LV(2), the left and right ventricular outflow tracts, and the papillary muscles.(1) However, such studies may alter the PVC pattern. The accuracy of 12-lead ECG is limited by variability in lead placement. A better understanding of in vivo PVC characteristics may allow the identification of MR patients at risk for malignant arrhythmia and facilitate catheter or surgical ablation. 

Noninvasive Electrocardiographic Imaging (ECGI) reconstructs electric potentials on the ventricular epicardial surface and has been successfully used to study human cardiac electrical activity under physiological conditions.(14, 15) ECGI can map dynamic changes in the origin and sequence of ventricular arrhythmia.(16, 17)  ECGI-computed SOO and mechanisms of VT concurred with invasive electrophysiologic studies in 25 patients undergoing catheter ablation (16, 17) ECGI has been validated and can provide spatial discrimination within 10 mm.(18, 19) It has been used in longitudinal studies to map electrophysiological substrate and arrhythmia mechanisms. (20, 21)

In the present study, we used ECGI to define PVC characteristics in MR patients before and after mitral valve surgery.

Fifty-two patients with primary MR underwent ECGI. PVCs were mapped in 13/52 MR patients during pre-surgery ECGI. Fifteen age-matched volunteers (mean age: 66 ± 8 years) with normal ECGs and no history of cardiac disease were recruited as controls. The Washington University Institutional Review Board approved the protocol for this study. All participants gave written informed consent.

Inclusion and Exclusion Criteria

Patients in the study group met all the criteria to be considered eligible for study participation:

1. A diagnosis of degenerative mitral valve disease and planned surgical intervention: (a) lone mitral valve replacement; (b) lone mitral valve repair; (c) mitral valve replacement or repair with concomitant Cox Maze procedure, or tricuspid valve procedure.
2. 18 years of age or above.

Patients in the control group met the following criteria:

Normal 12 lead ECG.

Between 45 and 80 years of age.

No history of arrhythmia or valvular heart disease.

The exclusion criteria included the following: 

1. Need for concomitant aortic or pulmonic valve surgery.
2. Concomitant coronary artery bypass grafting.
3. Ejection fraction <30%.
4. Prior cardiac ablation or surgery.
5. Women who are pregnant or breastfeeding.
6. Contraindication for MRI including severe claustrophobia, permanent metal implants, pacemakers, AICD’s, joint replacements, and coronary stents placed within the last six months.
7. Prior history of kidney disease or other contraindication to gadolinium-based contrast.
8. Uncontrolled hypertension.
9. Morbid obesity (BMI >40) or inability to be placed in MRI scanner due to body mass.

Electrocardiographic Imaging

Body surface electrocardiograms were recorded using 256 electrodes placed on the torso in supine position. The recording duration ranged from 2 to 5 minutes. Patients underwent a thoracic CT or MRI scan gated at 70% of the R-R interval which provided cardiac anatomy and body-surface electrode location in the same frame of reference. Anatomical imaging data were segmented using commercially available image processing software (Amira, Thermo Fisher Scientific, Waltham, MA) to obtain patient-specific 3-dimensional biventricular geometry. The electrical data and ventricular geometry were processed using custom algorithms in MATLAB (Mathworks, Natick, MA) to reconstruct electrical potentials and unipolar electrograms on 502 locations of the biventricular epicardial surface during sinus rhythm and PVCs.(16)

Data Analysis

Local activation time (AT) was computed as the time of steepest negative deflection during the QRS complex of each epicardial electrogram. Activation times were indexed to the QRS onset and plotted on the biventricular geometry to obtain an activation isochrone map during sinus rhythm and PVCs. The total ventricular activation time (TVAT) was computed as the difference between the latest and earliest AT. A TVAT value greater than 120 ms was defined as a wide QRS PVC.

PVC burden was quantified as the rate of PVCs observed in the body surface electrocardiograms over the mapping duration. PVCs were characterized as follows: (i) the site of origin (SOO) of the PVC was determined from earliest site of activation on the isochrone map and verified using the location of earliest potential minimum in the corresponding epicardial potential map; ii) the pattern of activation (right or left bundle branch block, RBBB and LBBB respectively) was determined based on the activation isochrone map as described in our previous work.(14) The coupling interval of the PVC was computed as the time interval between the QRS of the premature beat and that of the preceding sinus rhythm beat. 

Follow-up ECGI was done in similar fashion on 22/52 patients three months after surgery. PVCs were mapped in 6/22 patients during follow-up.

Statistical Analysis

To test that the PVC rates are equally distributed in the MR and control groups, it was assumed that the distribution of all PVCs in the MR and control groups can be described as a random process in which each PVC occurs either in the MR or the control group, with probabilities pMR=tMR/(tMR+tC) for the MR group and pC=tC/(tMR+tC) for the control group, where tMR is the overall recording time in the MR group and tC is the overall recording time in the control group. For a given overall number of PVCs n, the probability of k PVCs in the control group is then given by the binomial distribution B(pC,n)(k) and the probability for k or less PVCs in the control group would be given by cumulative density function of B(pC,n)(k). A corresponding analysis was also performed to compare the pre- and post-surgery PVC burden.

Panel C shows the activation sequence during PVC mapped from the inferior basal LV in the same MR patient in panel A. The epicardial potential map (panel D) further confirms the SOO. The TVAT was 82 ms. Panel E shows the activation sequence during PVC mapped underneath the tricuspid valve in a different patient. The pattern of activation is consistent with a LBBB pattern. The epicardial potential map (panel B) further confirms the SOO. The TVAT was 115 ms. Panel G shows the activation sequence during PVC mapped from the right ventricular outflow tract (RVOT) in a different patient. This pattern of activation is LBBB pattern. The epicardial potential map (panel D) further confirms the SOO. The TVAT was 124 ms. 

Online Supplement

Electrocardiographic Imaging of Premature Ventricular Contractions in Mitral Regurgitation Patients

Ramya Vijayakumar, PhD, Tari-Ann Yates, MD, Martha McGilvray, MD,MSt, MPHS, Richard B. Schuessler, PhD, Christian Zemlin, PhD, Yoram Rudy, PhD, Ralph J. Damiano Jr., MD, Matthew R. Schill, MD

Ventricular activation in MR patient during sinus rhythm

Online Supplement Figure 1 (panel A) shows the ventricular epicardial activation isochrones for an MR patient mapped during sinus rhythm in two views. Activation originates from the anterior right ventricle (RV; black asterisk in red region) and proceeds towards the basal region of the left ventricle (blue). The total ventricular activation time (TVAT) was 52 ms. The corresponding epicardial potential map (panel B, same two views as in panel A) shows an area of early intense negative potential minimum, 25 ms after the onset of QRS, at the same location (white asterisk) in the anterior RV. This confirms the initiation site in the isochrone maps of panel A.  The electrogram from the earliest activation site is included. The electrogram shows an r wave, indicating the sub-endocardial origin of activation. This activation pattern was observed in all 52 MR patients during sinus rhythm and was consistent with previously published data on healthy adults.(1, 2) The mean TVAT for MR patients during sinus rhythm was 68 ± 12 ms, which was not significantly different than  TVAT in control patients, 71±11 ms; N = 15.

Figure Caption

Online Supplement Figure 1: Sinus rhythm activation in an MR patient. Panel A shows the sequence of ventricular epicardial activation in two views. Activation originates in the anterior RV (black asterisk in red region) and propagates across the septum. The basolateral LV is the latest region to activate (blue). Panel B shows the corresponding epicardial potential distribution at t = 25 ms from the onset of QRS. The development of the potential minimum (blue region in panel B) confirms the earliest activation site (white asterisk). Total ventricular activation time was 52 ms. The electrogram at the earliest activation size is shown in blue.

PVCs were mapped in 25% of MR patients undergoing surgery. The PVC burden in MR patients was significantly greater than in the age-matched control group. Noninvasive ECGI mapped polymorphic, wide QRS PVCs with a predominant RBBB activation pattern, consistent with other studies.(1, 2, 22)  Complex ectopy such as ventricular doublets and bigeminy were mapped in 19% of MR patients. PVCs most frequently originated from the mitral and peri-annular region. PVCs were also mapped in a subset of MR patients three months after surgery, and the overall PVC burden was found to increased slightly but significantly compared to before mitral surgery.

Among degenerative mitral valve disease, Barlow’s disease, which is found more commonly in young patients, has been known to be associated with sudden death and arrhythmias (3, 4). However, in the present study, we found that older patients with degenerative MR without evidence of Barlow’s disease had a substantially and significantly higher PVC burden than the age-matched healthy volunteers (p < 0.001).  Studies have reported similar findings while comparing patients with and without mitral valve prolapse.(23) While the present cohort did not have sustained VT/VF, the high prevalence of PVCs indicates an abnormal electrophysiological substrate, which is of potential clinical interest. 

The SOO of PVC is an important marker of sudden death, as a previous study identified PVCs from Purkinje origin to be triggers for VF in patients with bileaflet MVP and history of cardiac arrest.(22) . While Holter recordings can provide the overall PVC burden, PVC SOO and frequency are challenging to obtain noninvasively and often have required an invasive EP study. Panoramic ECGI data analyzed for several beats (QRS intervals) enabled us to image distinct PVC SOO and determine their frequency in each patient during physiological in vivo conditions. Polymorphic PVCs were mapped in 7 patients as shown in Table 4.  Most PVC locations were basal and could not be classified as Purkinje origin. The most frequent SOO were the mitral annulus and peri-annular region. Eighty eight percent of the patients in the present study had annular dilation. It can be hypothesized that annular dilation affects the stress-strain relation in this region, creating a substrate that can possibly facilitate PVCs. PVCs were also mapped in the lateral LV, inferobasal LV and inferior septum. Studies examining pathology and LGE in MVP patients have hypothesized that these areas, closely linked to the valve, are subject to greater mechanical stress, which could lead to increased fibrosis creating a substrate for PVCs.(2, 24) Future studies will be needed to define the relationship between ECGI-computed PVC SOO and anatomical fibrosis using delayed enhancement MRI. Our findings are consistent with that of a previous study, which similarly reported that an LV origin of PVCs was more common than other locations in patients with mitral valve prolapse.(25)

An interesting finding of our study was that 12/15 (80%) of patients had PVCs that were repeatedly mapped from the same location as shown in Table 2. Prior studies have shown the feasibility of stand-alone radiofrequency ablation(22) and intraoperative cryoablation(26) of PVCs in patients with bileaflet prolapse. ECGI can provide a priori patient-specific PVC SOO and frequency, which can help plan ablation.

The mean PVC coupling interval (Table 4) was greater than the normal QT interval of 460 ms, indicating that on average, PVCs occurred after the refractory period and would be less likely to cause VT/VF. Patient 7 and 15 had a PVC coupling interval of less than 460 ms, but they each had only one PVC during the mapping period. Patients with monomorphic PVCs did not exhibit variability in the coupling interval. Furthermore, none of the patients in this study had documented VT/VF. Therefore, it was not possible to assess the malignancy of these PVCs. A prospective ECGI study in a much larger cohort would be needed for further risk stratification, since sudden death and malignant ventricular arrhythmias are rare in this population (0.29% of MVP cases in US). (27) 

ECGI was successfully used as a follow-up tool to map post-surgery PVC characteristics in 6 patients. The most frequent PVC locations after surgery were the mitral and peri-annular regions. Comparison of pre- and post-surgery ECGI maps in each patient showed that post-surgery PVCs typically had different SOO compared to pre-surgery PVCs. In one patient, a PVC SOO (inferobasal LV) persisted after surgery. This patient also had LGE and aneurysm in this region, which indicates a possible role of fibrotic substrate in sustaining PVC. 

The overall PVC burden was significantly greater after surgery, but the significance is caused by a single patient (Patient 24 in the Supplemental Table 1) and therefore not robust.  PVC burden decreased in four patients and increased in three patients after surgery. The other patients did not have PVCs before or after surgery. Naksuk et al. examined the effect of mitral repair using Holter recordings and reported that surgery did not uniformly reduce the ectopy burden.(28) A more comprehensive evaluation is needed to determine both the preoperative and postoperative burdens of PVCs in order to establish the effects of surgery on PVC burden, and this would be best accomplished by more prolonged monitoring. 

Limitations

There were a number of limitations to this study. First, the MR group was remarkable for any history of AF and heart failure compared to the control group due to the inclusion and exclusion criteria. This could have played a role in the PVC burden mapped in MR. Secondly, body surface signals were recorded for a limited duration both before and after surgery, and this may not have given a comprehensive look at the PVC type and burden. The patients did not have any sustained ventricular arrhythmias at the time of ECGI. The focus of this study was to understand the PVC SOO and activation patterns. As such, repolarization parameters were not examined in this pilot study.  Also, this study included a relatively small number of patients, and none considered at high risk for sudden death or ventricular arrhythmias, such as those with Barlow’s disease or mitral annular dysjunction. Moreover, the incidence of ventricular arrhythmias and sudden death is relatively low and would require a much larger study to precisely determine electrophysiological mechanisms.

Despite these limitations, the present study is the largest study using noninvasive ECGI to define the PVC site of origin and activation patterns in MR patients. This may provide a window into understanding the phenomena and mechanisms underlying the much less common ventricular arrhythmias in this population of patients.

PVCs in MR patients originated most frequently from the peri-annular region. It can be hypothesized that the annular dilation affects the stress-strain relation in this region, creating a possible  substrate that can facilitate PVCs. Future studies will examine the presence and location of fibrosis using cardiac MRI in this region. The peri-annular region also contributed to the greatest number of PVCs in a subset of patients mapped after surgery, indicating that the valve surgery alone did not dramatically affect the PVC burden or location. Future studies are needed to determine the clinical relevance of these findings.

1. Sriram CS, Syed FF, Ferguson ME, Johnson JN, Enriquez-Sarano M, Cetta F, et al. Malignant bileaflet mitral valve prolapse syndrome in patients with otherwise idiopathic out-of-hospital cardiac arrest. Journal of the American College of Cardiology. 2013;62(3):222-30.
2. Basso C, Perazzolo Marra M, Rizzo S, De Lazzari M, Giorgi B, Cipriani A, et al. Arrhythmic Mitral Valve Prolapse and Sudden Cardiac Death. Circulation. 2015;132(7):556-66.
3. Vergara P, Altizio S, Falasconi G, Pannone L, Gulletta S, Della Bella P. Electrophysiological Substrate in Patients with Barlow's Disease. Arrhythm Electrophysiol Rev. 2021;10(1):33-7.
4. Vergara P, Scarfo I, Esposito A, Colantoni C, Palmisano A, Altizio S, et al. Characterization of the electrophysiological substrate in patients with Barlow's disease. J Cardiovasc Electrophysiol. 2021;32(12):3179-86.
5. Ahmed M, Roshdy A, Sharma R, Fletcher N. Sudden cardiac arrest and coexisting mitral valve prolapse: a case report and literature review. Echo research and practice. 2016;3(1):D1-8.
6. Turker Y, Ozaydin M, Acar G, Ozgul M, Hoscan Y, Varol E, et al. Predictors of ventricular arrhythmias in patients with mitral valve prolapse. The international journal of cardiovascular imaging. 2010;26(2):139-45.
7. Miller MA, Dukkipati SR, Turagam M, Liao SL, Adams DH, Reddy VY. Arrhythmic Mitral Valve Prolapse: JACC Review Topic of the Week. Journal of the American College of Cardiology. 2018;72(23 Pt A):2904-14.
8. Noseworthy PA, Asirvatham SJ. The Knot That Binds Mitral Valve Prolapse and Sudden Cardiac Death. Circulation. 2015;132(7):551-2.
9. Carballeira Pol L, Deyell MW, Frankel DS, Benhayon D, Squara F, Chik W, et al. Ventricular premature depolarization QRS duration as a new marker of risk for the development of ventricular premature depolarization-induced cardiomyopathy. Heart rhythm. 2014;11(2):299-306.
10. Kawamura M, Badhwar N, Vedantham V, Tseng ZH, Lee BK, Lee RJ, et al. Coupling interval dispersion and body mass index are independent predictors of idiopathic premature ventricular complex-induced cardiomyopathy. J Cardiovasc Electrophysiol. 2014;25(7):756-62.
11. Bradfield JS, Homsi M, Shivkumar K, Miller JM. Coupling interval variability differentiates ventricular ectopic complexes arising in the aortic sinus of valsalva and great cardiac vein from other sources: mechanistic and arrhythmic risk implications. Journal of the American College of Cardiology. 2014;63(20):2151-8.
12. Lee CH, Park KH, Nam JH, Lee J, Choi YJ, Kong EJ, et al. Increased variability of the coupling interval of premature ventricular contractions as a predictor of cardiac mortality in patients with left ventricular dysfunction. Circulation journal : official journal of the Japanese Circulation Society. 2015;79(11):2360-6.
13. Lin CY, Chang SL, Lin YJ, Lo LW, Chung FP, Chen YY, et al. Long-term outcome of multiform premature ventricular complexes in structurally normal heart. International journal of cardiology. 2015;180:80-5.
14. Ramanathan C, Ghanem RN, Jia P, Ryu K, Rudy Y. Noninvasive electrocardiographic imaging for cardiac electrophysiology and arrhythmia. Nat Med. 2004;10(4):422-8.
15. Ramanathan C, Jia P, Ghanem R, Ryu K, Rudy Y. Activation and repolarization of the normal human heart under complete physiological conditions. Proc Natl Acad Sci U S A. 2006;103(16):6309-14.
16. Wang Y, Cuculich PS, Zhang J, Desouza KA, Vijayakumar R, Chen J, et al. Noninvasive electroanatomic mapping of human ventricular arrhythmias with electrocardiographic imaging. Science translational medicine. 2011;3(98):98ra84.
17. Wang Y, Li L, Cuculich PS, Rudy Y. Electrocardiographic imaging of ventricular bigeminy in a human subject. Circ Arrhythm Electrophysiol. 2008;1(1):74-5.
18. Oster HS, Taccardi B, Lux RL, Ershler PR, Rudy Y. Noninvasive electrocardiographic imaging: reconstruction of epicardial potentials, electrograms, and isochrones and localization of single and multiple electrocardiac events. Circulation. 1997;96(3):1012-24.
19. Cluitmans MJM, Bonizzi P, Karel JMH, Das M, Kietselaer B, de Jong MMJ, et al. In Vivo Validation of Electrocardiographic Imaging. JACC Clinical electrophysiology. 2017;3(3):232-42.
20. Vijayakumar R, Faddis MN, Cuculich PS, Rudy Y. Mechanisms of persistent atrial fibrillation and recurrences within 12 months post-ablation: Non-invasive mapping with electrocardiographic imaging. Frontiers in cardiovascular medicine. 2022;9:1052195.
21. Ghosh S, Rhee EK, Avari JN, Woodard PK, Rudy Y. Cardiac memory in patients with Wolff-Parkinson-White syndrome: noninvasive imaging of activation and repolarization before and after catheter ablation. Circulation. 2008;118(9):907-15.
22. Syed FF, Ackerman MJ, McLeod CJ, Kapa S, Mulpuru SK, Sriram CS, et al. Sites of Successful Ventricular Fibrillation Ablation in Bileaflet Mitral Valve Prolapse Syndrome. Circulation Arrhythmia and electrophysiology. 2016;9(5).
23. Engin B, Cevik E, Deniz R, Orta H, Elitok A. Ventricular arrhythmias in mitral valve prolapse syndrome and their relationship with electrocardiographic repolarization parameters. Northern clinics of Istanbul. 2023;10(2):205-11.
24. Bui AH, Roujol S, Foppa M, Kissinger KV, Goddu B, Hauser TH, et al. Diffuse myocardial fibrosis in patients with mitral valve prolapse and ventricular arrhythmia. Heart. 2017;103(3):204-9.
25. Hong T, Yang M, Zhong L, Lee YH, Vaidya VR, Asirvatham SJ, et al. Ventricular premature contraction associated with mitral valve prolapse. International journal of cardiology. 2016;221:1144-9.
26. El-Eshmawi A, Pandis D, Miller MA, Boateng P, Dukkipati SR, Reddy VY, et al. Surgical Cryoablation of Papillary Muscle PVCs During Mitral Valve Surgery: Therapeutic Consideration for Malignant MVP. Journal of the American College of Cardiology. 2020;76(25):3061-2.
27. Zuin M, Rigatelli G, Bilato C. Sudden Cardiac Death in Patients With Mitral Valve Prolapse in US (1999-2020). Am J Cardiol. 2023;193:34-6.
28. Naksuk N, Syed FF, Krittanawong C, Anderson MJ, Ebrille E, DeSimone CV, et al. The effect of mitral valve surgery on ventricular arrhythmia in patients with bileaflet mitr
29. Ramanathan C, Jia P, Ghanem R, Ryu K, Rudy Y. Activation and repolarization of the normal human heart under complete physiological conditions. Proceedings of the National Academy of Sciences of the United States of America. 2006;103(16):6309-14.
30. Andrews C, Cupps BP, Pasque MK, Rudy Y. Electromechanics of the Normal Human Heart In Situ. Circulation Arrhythmia and electrophysiology. 2019;12(11):e007484.