OUP user menu

Effects of mid-myocardial pacing on transmural dispersion of repolarization and arrhythmogenesis

Tao Xu, Hao Wang, Jia-You Zhang, Yu Zhang, Ran Zhang, Li-Qin Jiang, Ji-Feng Zheng, Hang Zhu, Zong-Gui Wu, De-Ning Liao
DOI: http://dx.doi.org/10.1093/europace/eus011 1363-1368 First published online: 9 February 2012


Introduction Epicardial (Epi) activation of the left ventricular (LV) wall increases transmural dispersion of repolarization (TDR), which creates a substrate for the development of ventricular arrhythmia. We hypothesize that pacing from the LV mid-myocardium may decrease the TDR and occurrence of arrhythmias.

Methods and results A transmural electrocardiogram and transmembrane action potentials were simultaneously recorded from Epi, mid-myocardial (M), and endocardial (Endo) layers of the arterially perfused canine LV wedge preparations (n= 8). Transmural dispersion of repolarization varied when the preparations were paced at each layer, respectively (Endo pacing, 35.6 ± 6.6 ms; M pacing, 34.9 ± 7.3 ms; Epi pacing, 72.4 ± 4.9 ms; P< 0.001). A significant difference was noted in TDR between M pacing and Epi pacing (P< 0.001), but not between M pacing and Endo pacing (P= 0.831). This result was reproducible in the presence of ischaemia-reperfusion experiments (n= 8). Transmural dispersion of repolarization was amplified as compared with non-ischaemic experiments and differed when preparations were paced at each layer (Endo pacing, 62.8 ± 13.8 ms; M pacing, 63.3 ± 13.3 ms; Epi pacing, 111.1 ± 17.7 ms; P< 0.001). There was again no significant difference between Endo pacing and M pacing (P= 0.948). However, as pacing was shifted from M to Epi, there was a significant increase in TDR (P< 0.001). Ventricular arrhythmias were induced in two of eight ischaemic preparations during Epi pacing, but did not occur in either M or Endo pacing.

Conclusion Mid-myocardial pacing can significantly decrease the TDR and prevent the occurrence of ventricular arrhythmias as compared with Epi pacing.

  • Electrophysiology
  • Transmural dispersion of repolarization
  • Mid-myocardial pacing
  • Action potential duration


Cardiac resynchronization therapy (CRT) is one of the most effective therapies for drug-refractory, chronic congestive heart failure in recent years. Numerous randomized clinical trials have demonstrated that CRT could improve cardiac functional status, quality of life, and exercise tolerance.1 However, this frequently performed treatment might have some disadvantages, one of which is the production of a proarrhythmic effect in a number of studies.25

Medina-Ravell et al.5 first reported the development of ventricular arrhythmia in a few of heart failure patients whenever epicardial (Epi) pacing was instituted, but not when the stimulation was delivered from the endocardium. They suggested that the Epi pacing prolonged the QT interval and increased the existing transmural dispersion of repolarization (TDR) due to a reversal of the normal activation sequence, creating the substrate and trigger for re-entrant arrhythmias, especially under long QT conditions. These may be the underlying mechanisms for the development of ventricular arrhythmia in patients undergoing resynchronization therapy. Afterwards, Medina-Ravell et al.5 and Fish et al.6 studied the cellular basis of arrhythmias and QT interval prolongation during left ventricular (LV) Epi pacing. They found that the reversal of the ventricular activation sequence (epicardium activating earlier than the mid-myocardium and endocardium) from the Epi pacing led to earlier repolarization of the epicardium and delayed activation and repolarization of the mid-myocardial (M) layer cell, which results in QT prolongation, enhanced transmural heterogeneity, and torsades de pointes (Tdp). Medina-Ravell et al.5 and Fish et al.6 suggested that the potentially arrhythmogenic substrate can be avoided by stimulation of LV endocardium or possibly by usage of a screw-in lead to stimulate the mid-myocardium of the LV free wall. The latter method is likely to reduce TDR and produce an antiarrhythmic effect.

In this study, we examine the effects of pacing in the LV mid-myocardium on TDR in canines, and provide experimental evidence for the prevention of TDR-related arrhythmia in a clinical setting.


Canine arterially perfused left ventricular wedge

Using the methods of Yan et al.,7 adult mongrel dogs weighing 15–25 kg were anticoagulated with heparin (180 IU/kg) and anaesthetized with sodium pentobarbital (35 mg/kg, intravenously). The chests were opened via a left-thoracotomy and the hearts were excised and placed in a cold cardioplegic solution (Tyrode's solution containing 8.5 mmol/L [K+], 4°C). Arterially perfused LV wedge preparations were cannulated via a small branch of the left descending coronary artery and perfused with cardioplegic solution. The unperfused region was removed using a razor blade. The preparations were then placed in a chamber and perfused again with normal Tyrode's solution (4 mmol/L [K+]). The preparations were fully immersed in extracellular solution throughout the course of the experiment. Perfusion pressure was maintained at 35–45 mmHg and the temperature at 35.7 ± 0.5°C. The total time from excision of the heart to cannulation and perfusion of the artery was <4 min in all experiments.

The composition of the Tyrode's solution was (in mmol/L) NaCl 129, KCl 4, NaH2PO4 0.9, NaHCO3 20, CaCl2 1.8, MgSO4 0.5, and glucose 5.5, buffered with 95% O2 and 5% CO2.

Electrophysiology recording and different site pacing

The wedge preparations were allowed to equilibrate in the chamber until electrically stable, usually requiring 1 h. The preparations were stimulated using bipolar silver electrodes applied to the endocardial (Endo) surface at a basic cycle length (BCL) of 2000 ms.

Transmembrane action potentials were recorded from Epi, M, and Endo cells using floating microelectrodes. A transmural pseudo-electrocardiogram (ECG) was recorded with the use of two AgCl half cells placed ∼1 cm from the Epi and Endo surfaces of the preparations along the same axis as the transmembrane recordings.

During experiment, the stimulated electrode was relocated from the endocardium to the mid-myocardium, then to the epicardium at a BCL of 2000 ms. The transmural ECG and transmembrane action potentials were simultaneously recorded.

Ischaemia-reperfusion and programmed electrical stimulation

After the electrophysiology recordings were in baseline normal condition, the perfusion of preparations was stopped for 30 min, then resumed. In the course of ischaemia-reperfusion, ST segment elevated first, then became depressed. Electrophysiology recordings were not performed until the ST segment was stable, usually 30 min after reperfusion. Arrhythmias were induced by programmed electrical stimulation. The basic stimulus (S1) was delivered to the endocardium, mid-myocardium, or epicardium. After every fifth S1 (S1–S1 cycle length = 2000 ms), a premature stimulus (S2) was delivered. The S1–S2 coupling interval was progressively reduced until the preparation reached refractory (S2 stimuli were of 2–3 ms duration with an intensity equal to two-folds over the baseline pacing threshold).

Statistics analysis

Statistical analysis was done using SPSS 17.0 software. Data of electrophysiological parameters were presented as mean and standard deviations. To evaluate difference of electrophysiological parameters among Endo, M, and Epi pacing, one-way analysis of variance was performed. The Student–Newman–Keuls (SNK test) was used as a post hoc test in the subsequent multiple comparative analysis. The paired t-test was applied in comparison of electrophysiological parameters between control and ischaemia-reperfusion group. Criterion for statistical significance was P< 0.05.


Effect of endocardial, mid-myocardial, and epicardial pacing on ventricular repolarization and transmural activation sequence

With Endo pacing, the endocardium was activated first, followed by the mid-myocardium and then the epicardium. The M cell displayed the longest action potential duration (APD), its APD90 (measured at 90% repolarization) was 307.8 ± 12.3 ms; Endo APD90 was 275.1 ± 13.8 ms; and epicardium was the last to be activated and the first to repolarize, displaying the shortest APD90 at 259.9 ± 9.6 ms. Transmural dispersion of repolarization and Tp–Te (interval from the peak to the end of the T wave in transmural ECG) were 35.6 ± 6.6 and 34.9 ± 6.9 ms, respectively. With M pacing, the mid-myocardium was the first to activate but the last to repolarize, whereas the epicardium was the last to activate but the first to repolarize. Transmural dispersion of repolarization and Tp–Te were 34.9 ± 7.3 and 34.8 ± 6.8 ms. With Epi pacing, the epicardium was the first to activate and also the first to repolarize, whereas the M cells are the last to repolarize. Transmural dispersion of repolarization and Tp–Te were 70.7 ± 6.6 and 72.4 ± 4.9 ms. Action potential duration measured at 90% repolarization in all three layers showed no significant difference under Epi, M, and Endo pacing (P> 0.05) (Table 1).

View this table:
Table 1

Effect of endocardial, mid-myocardial, and epicardial pacing on action potential duration measured at 90% repolarization in the control and ischaemia-reperfusion group

Control (n= 8)Ischaemia-reperfusion (n= 8)
Endo APD90 (ms)M APD90 (ms)Epi APD90 (ms)Endo APD90 (ms)M APD90 (ms)Epi APD90 (ms)
Endo pacing275.1 ± 13.8307.8 ± 12.3259.9 ± 9.6215.9 ± 16.9*262.6 ± 8.9*184.6 ± 10.9*
M pacing275.3 ± 13.7307.9 ± 11.8259.8 ± 9.8216.1 ± 16.9*262.6 ± 8.9*184.6 ± 10.6*
Epi pacing274.9 ± 14.2306.3 ± 12.0259.1 ± 9.6216.9 ± 17.5*263.4 ± 8.5*185.0 ± 11.3*
P value0.9980.9560.9850.9930.9810.997
  • *P<0.05 (vs. control).

However, the conduction time from the mid-myocardium to epicardium (TM–Epi) increased when the pacing site shifted from Endo to Epi (P< 0.001), but not when pacing site shifted from Endo to M (P> 0.05) (Table 4 and Figure 1). As that it also indicated an increase in TDR and Tp–Te when pacing shifted from Endo to Epi but not from Endo to M (Tables 2 and 3 and Figure 1).

View this table:
Table 2

Transmural dispersion of repolarization of endocardial, mid-myocardial, and epicardial pacing between the control and ischaemia-reperfusion group (n= 8)

Endo pacingM pacingEpi pacingP value
ANOVAEndo vs. MM vs. EpiEpi vs. Endo
Control35.6 ± 6.634.9 ± 7.370.7 ± 6.6<0.0010.831<0.001<0.001
Ischaemia-reperfusion62.8 ± 13.8*63.3 ± 13.3*111.1 ± 17.7*<0.0010.948<0.001<0.001
  • ANOVA, analysis of variance.

  • *P<0.05 (vs. control).

Figure 1

Effect of different pacing site on transmural activation sequence and ventricular repolarization. (A) Endocardial pacing (‘Endo pacing’). (B) Mid-myocardial pacing (‘M pacing’). (C) Epicardial pacing (‘Epi pacing’).

Effect of endocardial, mid-myocardial, and epicardial pacing on ventricular repolarization and transmural activation sequence in ischaemia-reperfusion

During ischaemia-reperfusion, transmural activation sequence was the same as in baseline condition. Similarly, a change of pacing mode would also influence TDR, Tp–Te, and TM–Epi in the absence of transmembrane APD prolongation in all the three layers (Tables 24 and Figure 2). Action potential duration measured at 90% repolarization of three layers was obtained after ischaemia exhibit decrease when compared with pre-ischaemic APD, especially in the epicardium (Table 1). Transmural dispersion of repolarization, Tp–Te, and TM–Epi are initially augmented in the presence of ischaemia (Tables 24). A shift from Endo or M to Epi pacing caused a further prolongation of TDR and Tp–Te (Tables 2 and 3 and Figure 2).

View this table:
Table 3

Tp–Te interval of endocardial, mid-myocardial, epicardial pacing between the control and ischaemia-reperfusion group (n= 8)

Endo pacingM pacingEpi pacingP value
ANOVAEndo vs. MM vs. EpiEpi vs. Endo
Control34.9 ± 6.934.8 ± 6.872.4 ± 4.9<0.0010.969<0.001<0.001
Ischaemia-reperfusion63.6 ± 13.2*63.0 ± 12.6*112.1 ± 17.9*<0.0010.933<0.001<0.001
  • ANOVA, analysis of variance.

  • *P<0.05 (vs. control).

View this table:
Table 4

Effect of endocardial pacing vs. mid-myocardial pacing vs. epicardial pacing on TM–Epi during the control and ischaemia-reperfusion group (n= 8)

Endo pacingM pacingEpi pacingP value
ANOVAEndo vs. MM vs. EpiEpi vs. Endo
Control12.0 ± 0.812.6 ± 0.724.3 ± 2.4<0.0010.424<0.001<0.001
Ischaemia-reperfusion15.3 ± 2.8*15.1 ± 2.9*32.3 ± 5.8*<0.0010.952<0.001<0.001
  • ANOVA, analysis of variance.

  • *P<0.05 (vs. control).

Figure 2

Effect of different pacing site on transmural activation sequence and ventricular repolarization during ischaemia-reperfusion. (A) Endocardial pacing. (B) Mid-myocardial pacing. (C) Epicardial pacing.

Pacing site-dependent arrhythmia

Although Epi pacing increases TDR via the alteration in ventricular activation sequence, it is hard to induce ventricular arrhythmias under baseline conditions. In the controls, TDRs were 35 and 36 ms, respectively, during Endo and M pacing (Figure 1). Programmed electrical stimulation failed to induce ventricular arrhythmia under these conditions. When pacing was shifted to the epicardium, TDR increased to 71 ms. Programmed electrical stimulation again failed to induce ventricular arrhythmia. After ischaemia exaggerated the effect of Epi pacing to increase TDR, an extrastimulus could successively induce ventricular arrhythmia. In a total of eight preparations, ventricular tachycardia occurred in two preparations during Epi pacing, but not M or Endo pacing. One was induced by programmed electrical stimulation at an S1–S2 interval of 250 ms. The other was spontaneous, as shown in Figures 2 and 3. In this example, ischaemia prolonged TDR from 34 to 63 ms during Endo and M pacing. Epicardial pacing caused a further prolongation of TDR to 111 ms. During M and Endo pacing, a number of extrasystoles occurred, but ventricular tachycardia could not be induced at any coupling interval tested. When basic stimulation was shifted to the epicardium, ventricular arrhythmia occurred frequently (Figure 3).

Figure 3

A spontaneous ventricular arrhythmia during Epi pacing (A) but not during M pacing (B) or endocardial pacing (C) after ischaemia-reperfusion.


Abnormal amplification of TDR plays an important role in the development of re-entrant ventricular arrhythmias that may lead to sudden cardiac death. In certain pathological conditions, such as ischaemia and heart failure, it would have a proarrhythmic effect due to the increase of TDR.8,9 For patients with refractory heart failure secondary to ischaemic cardiomyopathy, the CRT could obviously improve their heart function, quality of life, and prognosis.10 However, CRT could be proarrhythmic5,11 as Epi pacing alters the transmural activation sequence of the intrinsically heterogeneous ventricular myocardium. When pacing from the epicardium, compared with Endo pacing or M pacing, the epicardium is activated and repolarizes early whereas the M cells are activated and repolarize later. Consequently, the Epi pacing augments the TDR and QT interval that underlies the development of arrhythmia,6 this is consistent with our results.

Previous studies demonstrated that conduction during Epi pacing is substantially different from conduction during Endo pacing under identical conditions.12 Transmural conduction velocity during Epi pacing was significantly slower than that of Endo pacing (48 ± 6 vs. 37 ± 6 cm/s, P < 0.01).12 Our study also revealed that the transmural conduction time (especially the conduction time between M and Epi) varies as pacing sites shifted (Table 4). Normally, ventricular activation starts from the endocardium via the subendocardial Purkinje network and spreads across the M region to the epicardium. The epicardium, with a shorter APD, is the last to be activated but is the first to repolarize, whereas the mid-myocardium, with the longest APD, is the last to repolarize. Consequently, the value of TDR was determined not only by the difference between M APD (longer APD) and Epi APD (shorter APD) but also by the activation sequence between them. Therefore, anything that alters ventricular activation sequence, delays activation in the mid-myocardium or speeds up activation in the epicardium would prolong TDR. This is demonstrated by the results of this study. Epicardial pacing delayed activation conducting to the mid-myocardium, and the additional conduction delay encountered contributes to the amplification of TDR. M pacing does not change the normal activation sequence between the mid-myocardium and epicardium, exhibits an endo-pacing-like effect on TDR and arrhythmia induction.

Previous studies illustrated that ventricular arrhythmia, especially Tdp, is easily induced only when TDR exceeds some critical point (usually ∼80–90 ms1315). In the normal heart, Epi pacing increases TDR but it is hard to reach the pro-arrhythmic level. In the ischaemia-reperfusion model, the TDR is augmented as the ventricular transmembrane APDs in all three layers are shorter than normal (Table 1), and Epi APD decreases more significantly than M and Endo APD (it is due to the different ischaemia tolerance of three ventricular layers, the endocardium and mid-myocardium are more resistant to ischaemia8,9). With this ischaemic substrate, Epi pacing would exacerbate the TDR more and it would be easy to reach the critical pro-arrhythmic point.

In summary, M pacing can effectively reduce TDR and reentry-related arrhythmia as compared with Epi pacing. Although the LV Endo pacing is closer to physiological pacing, it may produce more complications such as pericardial tamponade due to atrial septal puncture, the risk of mitral regurgitation, thrombosis, etc. As we all know, one quarter of the patients submitted to CRT are non-responders; the reasons for this vary,16 but one of the most important is where the LV lead is positioned, not only if it is in the lateral or in the posterior left ventricular wall but also if it is Endo, Epi, or M. It is the M pacing that could improve the intraventricular conduction through a fast centre-type activation, which would theoretically enhance the clinical response to CRT, reduce the TDR, and enforce the safety of CRT. Although there are no M electrode products and matched applications available at present, stimulation of the LV mid-myocardium may be achieved by placing a screw-in lead mid-myocardially through thoracotomy or even the coronary vein in future.

Conflict of interest: none declared.


The study was supported by a grant from the National Nature Science Foundation of China.


View Abstract