OUP user menu

Assessment of optimal right ventricular pacing site using invasive measurement of left ventricular systolic and diastolic function

Vlastimil Vančura, Dan Wichterle, Vojtěch Melenovský, Josef Kautzner
DOI: http://dx.doi.org/10.1093/europace/eut068 1482-1490 First published online: 13 April 2013

Abstract

Aims Right ventricular apical pacing has a detrimental effect on left ventricular (LV) function. More optimal pacing site may be found by invasive measurement of LV mechanical performance during pacing from different RV pacing sites. We aimed to investigate the effect of RV pacing lead location on invasive indices of LV mechanical performance.

Methods and results Patients undergoing catheter ablation for persistent atrial fibrillation were enrolled. Single-site endocardial pacing from the lateral LV region was periodically switched to pacing from the mapping catheter navigated to different RV sites within the three-dimensional electroanatomical RV map. SystIndex, DiastIndex, and PPIndex were defined as the ratio of LV dP/dtmax, LV dP/dtmin, and arterial pulse pressure during RV pacing to corresponding values from adjacent periods of LV pacing. Haemodynamic data were analysed in 18 RV segments created by dividing RV horizontally (basal, mid, and apical portion), vertically (inferior, mid, and superior portion) and frontally (septum and free wall). Eight patients (58 ± 7 years; 2 females; 26 ± 4 RV pacing sites per patient) were enrolled into the study. Compared with LV pacing, the best RV pacing values of SystIndex and DiastIndex were achieved in basal-mid-septal segment (+6.9%, P = 0.02 and +3.4%, P = 0.36, respectively) while the best PPIndex was obtained in superior-mid-septal segment of RV (+4.5%, P = 0.02). All indices were fairly concordant showing significant improvement of haemodynamics during RV pacing in the direction from free wall to septum, from apex to base, and from inferior to superior segments.

Conclusion The best LV mechanical performance was achieved by RV septal pacing in the non-apical mid-to-superior segments.

  • Right ventricular pacing
  • Haemodynamics of pacing
  • Electroanatomical mapping
  • Optimal pacing site
  • Resynchronization

What's new?

  • This is the first study analysing virtually entire right ventricular (RV) endocardial surface for the impact of pacing on left ventricular (LV) mechanical performance.

  • Investigational protocol utilized: (i) high-rate ventricular pacing; (ii) three-dimensional navigation to RV pacing sites; (iii) absence of atrioventricular synchrony issues; (iv) invasively measured LV function; (v) reference LV pacing; and (vi) computerized analysis of signals.

  • Concordant gradients in improvement of LV mechanical performance were observed when pacing in the direction from free wall to septum, from apex to base, and from inferior to superior RV segments.

Introduction

Implantation of a pacemaker is the only reliable long-term treatment option in patients with significant bradyarrhythmias. Most pacing systems employ one lead in the right ventricle (RV) usually implanted into the apical region. However, RV apical pacing prolongs QRS complex duration,1 induces mechanical asynchrony,2,3 promotes atrial fibrillation (AF) and heart failure.46 Detrimental effect of RV apical pacing could be potentially diminished by the use of alternative RV pacing sites. Several studies demonstrated that septal pacing or pacing from RV outflow tract shows better results than apical pacing.3,711 However, none of them mapped systematically RV endocardium to evaluate haemodynamic impact of pacing from various locations.

We hypothesized that a distinct RV pacing region could be identified which would result in best left ventricular (LV) haemodynamics, irrespective of interindividual heterogeneity in pacing response. To test this hypothesis we investigated acute changes in invasively measured indices of LV mechanical performance by alternating single-site LV pacing with different RV pacing sites under the guidance of the three-dimensional (3D) navigation system. The study was performed in patients with persistent AF in order to avoid possible bias introduced by variable atrioventricular delay.

Methods

Patients

Patients undergoing catheter ablation for AF were eligible for the study. Exclusion criteria encompassed significant structural heart disease (LV ejection fraction <40%, LV end-diastolic diameter >60 mm, QRS ≥ 0.12), valvular heart disease, and coronary artery disease. Patients in sinus rhythm at the time of the procedure were not eligible. Only patients with strictly controlled ventricular response were enrolled in order to minimize the occurrence of fusion beats during ventricular pacing. The protocol of the study was complied with the Declaration of Helsinki, it was approved by the local Ethics Committee and all patients gave written informed consent before entering the study.

Study protocol

After a double transseptal puncture (8 Fr. Swartz SL1 introducer, 8.5 Fr. Agilis NxT steerable introducer, St Jude Medical, Inc.) under the guidance of intracardiac echocardiography (10 Fr. AcuNav, Siemens Healthcare), two catheters were transseptally inserted into the LV: Millar tipmanometric catheter (Millar Instruments, Inc.) and LV pacing catheter positioned at the lateral, basal-to-mid portion of the LV endocardial surface. A short, 5 Fr. sheath was placed into the left femoral artery and connected to standard blood pressure transducer. Left ventricular pacing with a cycle length of 430 ms (pacing output of 5 V and impulse duration of 1 ms) was used to eliminate the effect of fusion. A 3.5 mm irrigated-tip catheter (Navistar Thermocool, Biosense Webster, Inc.) was inserted through the long sheath (8.5 Fr. Swartz SR0, St Jude Medical, Inc.) into the RV. Electroanatomical RV map was obtained using the CARTO system (Biosense Webster, Inc.) with a minimum of 90 mapping points. In all patients, the location of His bundle was identified during short interruption of pacing and annotated on CARTO map while pacing was resumed to take into account possible spatial shifts associated with fast pacing.

The acquisition of electrocardiogram (ECG) and haemodynamic data was commenced after finalizing the anatomical RV map in sufficient quality. Surface ECG, LV, and femoral arterial pressure were recorded using the PowerLab system (ADInstruments Pty Ltd) at a sampling rate of 4 kHz. During continuous LV pacing, mapping catheter was consecutively positioned to specific RV sites. While maintaining stable endocardial RV contact, LV pacing was periodically switched to unipolar pacing from the mapping catheter. Positive pacing pole for unipolar pacing was connected to a proximal ring of decapolar coronary sinus catheter located at the coronary sinus ostium with negligible ventricular far-field potential in order to prevent anodal ventricular capture. Every RV pacing as well as interleaving LV pacing episodes had minimum duration of 30 s. At the end, study catheters were removed and catheter ablation for AF followed.

Data processing

Digital ECG and pressure recordings were exported and off-line analysed using a purpose-made software. QRS complexes were automatically detected employing a combination of threshold and derivative methods. Both ECG and pressure signals were carefully visually inspected and manually edited to exclude artifacts, irregularities in rhythm and morphology including premature/fusion beats, or non-capture events. Semi-automatic QRS template-matching algorithm was used to exclude all abnormal QRS morphologies. His capture was considered a reason for excluding a pacing site from the analysis because even small fluctuations in catheter position could have substantial effect on haemodynamics and permanent His-bundle pacing in clinical practice is not easy to achieve. Edited signals from individual 30-s LV or RV pacing episodes were signal-averaged. Subsequently, LV dP/dtmax, LV dP/dtmin, and femoral arterial pulse pressure were calculated (Figure 1).

Figure 1

Example of raw data and signal-averaged signals. Left panels (from top to bottom) show ECG, LV pressure, trend in LV systolic pressure, LV dP/dt, and trend in LV dP/dtmax during three consecutive 30 s episodes of RV (site X) − LV (reference) − RV (site X + 1) pacing. Two vertical red lines denote the time of switching from RV pacing to LV reference pacing and vice versa. Right panel shows (from top to bottom) the corresponding signal-averaged, single-cycle surface ECG channel, LV pressure, and LV dP/dt. Signal-templates for pacing from RV (site X), LV (reference), and RV (site X+1) are in red, green, and blue, respectively. Considerable variability of haemodynamic parameters in time can be noted even without changing of pacing site. This underscores the necessity of longer recordings for valid assessment of pacing-induced change in haemodynamics.

All indices obtained during RV pacing were adjusted for mean reference value from LV pacing periods immediately preceding and following a particular RV pacing episode. Thus for each pacing site, an index of contractility (SystIndex) was calculated as the ratio of LV dP/dtmax during RV pacing (LV dP/dtmaxRVP) and LV pacing (LV dP/dtmaxLVP):Embedded Image

Analogically, an index of relaxation (DiastIndex) was derived as the ratio of LV dP/dtmin during RV and LV pacing, and PPIndex as the ratio of pulse pressures during RV and LV pacing.

Haemodynamic indices from individual RV pacing sites from all patients were pooled and blindly grouped according to electroanatomically defined RV pacing segments. Both septum and RV free wall were divided horizontally into basal, mid, and apical portion, and vertically into inferior, mid, and superior portion, so that a total of 18 segments were defined (Figure 2).

Figure 2

Definition of RV segments. Right ventricular electroanatomical map. Left panel—free wall view (right anterior oblique projection), right panel—septum view (modified posterior anterior projection). Green and white points are markers for tricuspid and pulmonary valve annulus, respectively. Yellow point denotes the site with His bundle recording. Right ventricular segments are numbered for easier data presentation.

Statistical analysis

Haemodynamic variables were expressed as means with standard deviations and 95% confidence intervals after testing for normality of distribution (Shapiro Wilk's test). Superiority/inferiority of RV pacing in individual segments vs. LV pacing was tested by two-tailed t-test for dependent samples, specifically by comparison of indices against unity. One-way analysis of variance (ANOVA) with Newman–Keuls post hoc test was used for the analysis of differences among multiple RV pacing segments. Two-tailed t-test for independent samples was used for comparison of pairs of RV pacing regions. Pearson's correlation was applied for the analysis of relationship between haemodynamic variables. A P value ≤ 0.05 was considered statistically significant. All analyses were performed using the STATISTICA version 9 software (Statsoft, Inc.).

Results

Eight patients were included into the study (mean age 58.0 ± 6.5 years, two women). The mean LV ejection fraction was 57 ± 6% and LV end-diastolic diameter 54 ± 4 mm. Recordings from a total of 209 RV pacing sites (26 ± 4 per patient, range: 21–34) were eligible for the analysis. When assessed by individual RV segments, the highest and the lowest density of pacing sites were 2.8 ± 0.9 at segment #4 (superior midseptum) and 0.6 ± 0.7 at segment #1 (septal superior apex), respectively.

We have not observed fusion beats during postprocessing of signals. Only sporadic ventricular premature complexes induced by catheter irritation and non-capture events required manual editing together with exclusion of several subsequent normal pacing cycles with irregularity-induced transient change in haemodynamics. Data from one pacing site were excluded because of intermittent His bundle capture. The duration of measurement period did not exceed 45 min in any patient. Mean values for measured haemodynamic parameters from all episodes included into analysis are summarized in Table 1. There were no major complications related to either study protocol or subsequent left atrial ablation.

View this table:
Table 1

Global haemodynamic parameters

LVPRVP
LV dP/dtmax (mmHg/s)1245 ± 2181194 ± 222
LV dP/dtmin (mmHg/s)−1204 ± 354−1147 ± 309
Systolic BP (mmHg)112 ± 28110 ± 26
Diastolic BP (mmHg)77 ± 1576 ± 14
Pulse pressure (mmHg)36 ± 1334 ± 12
  • All differences between LVP and RVP were not significant.

  • BP, blood pressure; LV, left ventricle; LVP, left ventricular pacing; RVP, right ventricular pacing.

Haemodynamic data for individual right ventricular segments

Segment assignment as an independent categorical variable was significantly associated with SystIndex (r = 0.40, P = 0.008), DiastIndex (r = 0.47, P = 0.00007) as well as PPIndex (r = 0.53, P < 0.00001). The mean values of SystIndex, DiastIndex, and PPIndex with 95% confidence intervals for individual RV segments are shown in Figure 3. The colour-coded values of haemodynamic indices averaged per segment and projected on electroanatomical map are shown in Figure 4.

Figure 3

Indices of LV mechanical performance for individual RV pacing segments. Segments numbering is defined in Figure 2. Means and 95% confidence intervals are plotted for SystIndex (turquoise), DiastIndex (red), and PPIndex (green); FW—free wall; n, number of pacing sites in individual segment; RVP, right ventricular pacing; *P< 0.05, **P< 0.01 (statistical significance for the difference from unity).

Figure 4

Indices of LV mechanical performance projected on CARTO map. Right ventricular segments are sorted in descendent order by the magnitude of SystIndex (A, B), DiastIndex (C, D) and PPIndex (E, F), numbered accordingly (1 = maximum; 18 =minimum value), and displayed on CARTO map in free wall view (left) and septum view (right). Red, green, and magenta colouring is used for upper, mid, and lower terciles of haemodynamic indices. Tagged points as in Figure 2.

The highest SystIndex of 1.07 (P = 0.019) was found for segment #8 (septal-basal-mid) while the lowest SystIndex of 0.86 (P = 0.014) was found for segment #10 (free wall-apical-superior), both mutually different (P = 0.013). There were five more segments (septal-mid-inferior and four others at RV free wall) with significantly lower LV contractility than during LV pacing.

The highest DiastIndex of 1.03 (P = 0.36) was found for segment #8 (septal-basal-mid) while the lowest DiastIndex of 0.87 (P = 0.0004) was found for segment #18 (free wall-basal-inferior), both mutually different (P = 0.013). Superiority of DiastIndex was also found for segment #8 (septal-basal-mid) vs. segment #11 (free wall-apical-mid) (P = 0.037) as well as segment #4 (septal-mid-superior) vs. segment #18 (free wall-basal-inferior) (P = 0.039). There were six more segments (two at inferior septum and four at RV free wall) with significantly lower LV relaxation when compared to LV pacing.

The highest PPIndex of 1.04 (P = 0.02) was found for segment #4 (septal-mid-superior) while the lowest PPIndex of 0.86 (P = 0.0001) was found for segment #18 (free wall-basal-inferior), both mutually different (P = 0.008). There were six more segments (three at mid/inferior apex and three others at non-superior free wall) with significantly lower LV contractility than during LV pacing.

Table 2 shows pacing segments associated with the highest LV mechanical response (based on limited number of measurements) for individual patients and for individual indices of LV performance. Although the study was not powered to detect the optimum pacing segments in individual patients, it is obvious from the Table 2 that 88, 100, and 50% patients (depending on SystIndex, DiastIndex, and PPIndex used, respectively) had the maximum haemodynamic response if pacing at septum. On the other hand, pacing in all of six inferior segments was never associated with maximum response according to any of these three indices of LV mechanical performance. The absolute percentage differences in averaged LV haemodynamic indices between the best and the worst RV pacing segments reached 20.7, 16, and 18.9% for SystIndex, DiastIndex, and PPIndex, respectively.

View this table:
Table 2

Right ventricular pacing segments with the highest LV performance for individual patients/indices

Patient #SystIndexDiastIndexPPIndex
18 (2)8 (2)13 (1)
21 (1)7 (3)13 (2)
34 (2)5 (4)13 (2)
48 (2)4 (4)14 (2)
52 (4)2 (4)2 (4)
611 (1)8 (1)5 (5)
78 (1)8 (1)8 (1)
81 (1)8 (2)8 (2)
  • The values are segment # (numbers of corresponding pacing sites). Segments numbering is defined in Figure 2. Note that all optimum segments (except the segment #14 with PPIndex maximum in patient #4) belong to mid-to-superior septum or superior free wall area.

  • LV, left ventricle.

Correlation between SystIndex and DiastIndex analysed per individual pacing segments was significant (r = 0.53, P = 0.02, Figure 5).

Figure 5

Correlation between SystIndex and DiastIndex. SystIndex vs. DiastIndex were analysed per pacing segments. The best RV pacing site overall was the segment #8 (septal-basal-mid). Regression line with 95% confidence intervals indicates significant positive relationship between both indices.

Haemodynamic data for right ventricular pacing regions

Aggregation of neighbouring RV segments into clinically meaningful RV pacing regions showed that pacing in inferior regions (not only overall but also when restricted to septal portion) was clearly worse compared to mid/superior regions. Such significant association was consistent across all haemodynamic indices. Favourable effect of pacing in superior region compared to mid/inferior regions (again either overall or restricted to septal portion) was less evident and statistically limited to DiastIndex and PPIndex. Anticipated superiority of septal vs. free wall pacing was not particularly strong and appeared statistically significant only for SystIndex and DiastIndex. Even less prominent differences were observed in horizontal direction where PPIndex was the only significant variable demonstrating the inferiority of apical pacing while only borderline decrease of DiastIndex was found when pacing from RV apex. Detailed results are shown in Table 3.

View this table:
Table 3

Comparison of indices of LV mechanical performance in pre-specified vs. complementary RV pacing regions

Region ANSystIndexDiastIndexPPIndexRegion BNSystIndexDiastIndexPPIndexP (region A vs. B)
Mean ± SDMean ± SDSystIndexDiastIndexPPIndex
Septal1120.98 ± 0.120.98 ± 0.100.98 ± 0.12Free wall970.94 ± 0.110.94 ± 0.080.96 ± 0.110.0070.0030.11
Superior860.97 ± 0.120.99 ± 0.081.02 ± 0.10Non-superior1230.95 ± 0.120.94 ± 0.100.94 ± 0.110.420.0001<0.0001
Inferior500.91 ± 0.100.92 ± 0.090.89 ± 0.08Non-inferior1590.98 ± 0.120.98 ± 0.091.00 ± 0.110.0006<0.0001<0.0001
Apical420.95 ± 0.130.94 ± 0.090.93 ± 0.08Non-apical1670.96 ± 0.120.97 ± 0.100.98 ± 0.120.610.070.004
Basal730.97 ± 0.110.97 ± 0.100.98 ± 0.11Non-basal136096 ± 0.130.96 ± 0.100.97 ± 0.120.460.380.53
Septal superior451.00 ± 0.131.01 ± 0.091.02 ± 0.10Septal non-superior670.97 ± 0.120.97 ± 0.110.96 ± 0.120.280.030.003
Septal inferior230.91 ± 0.100.93 ± 0.080.90 ± 0.08Septal non-inferior891.00 ± 0.121.00 ± 0.111.01 ± 0.120.0040.005<0.0001
Septal non-inferior basal291.02 ± 0.121.02 ± 0.111.03 ± 0.13Septal non-inferior non-basal600.99 ± 0.130.99 ± 0.101.00 ± 0.110.280.210.27
Septal non-inferior apical170.97 ± 0.130.95 ± 0.090.92 ± 0.09Septal non-inferior non-apical721.00 ± 0.121.01 ± 0.111.02 ± 0.110.350.0540.0009
  • N, number of pacing sites in particular region; P, P value (t-test for independent samples); SD, standard deviation.

Discussion

To our knowledge, this is the first study analysing virtually entire RV endocardial surface for the impact of pacing on LV mechanical performance. The study benefited from (i) high-rate ventricular pacing; (ii) 3D navigation to RV pacing sites; (iii) absence of atrioventricular synchrony issues; (iv) invasively measured LV function; (v) reference LV pacing; and (vi) computerized signal analysis. We demonstrated significant variability of impact among RV pacing sites.

Left ventricular haemodynamics

All haemodynamic indices used in this study were moderately concordant across various pacing regions and confirmed the anticipated superiority of RV septal pacing and inferiority of apical pacing. The strongest gradient for improvement of LV haemodynamics was unexpectedly found in the direction from inferior to superior segments/regions. Such distinct vertical gradient favouring more superiorly located pacing sites was consistently observed for all three haemodynamic indices and remained significant even when limited only to the septum. In other words, the LV mechanical performance was better during RV pacing at superior segments of the septum but also at superior segments of RV free wall. Consequently, pacing at four septal superior/mid septal segments, specifically basal-mid (#8), basal-superior (#7), mid-mid (#5), and mid-superior (#4) segments, concordantly provided all three LV haemodynamic indices within the upper tercile.

The only RV pacing segment yielding consistently better LV haemodynamics than LV pacing was the septal-basal-mid segment (#8), even if the difference from LV pacing reached statistical significance only for SystIndex based on LV dP/dtmax. It should be emphasized that none of the pacing sites in this segment used for calculation demonstrated clear His capture. This finding is even more important when taking into consideration that our reference LV pacing site was within the area considered as the best LV pacing region in heart failure patients for many years12,13 and that LV endocardial pacing exerts more pronounced positive effect on haemodynamics when compared with epicardial pacing.14 On the other hand, two recent studies concluded that the best LV pacing area may be individually different and optimization during implant procedure is necessary.15,16 Moreover, optimum LV pacing site in our study group might substantially differ from candidates of cardiac resynchronization therapy. Conversely, we observed that many segments at inferior aspect of the RV (both RV free wall and inferior septum) were associated with significantly worse LV haemodynamics compared with LV pacing.

The absolute percentage differences in LV haemodynamic indices between the best and the worst RV pacing segments were obviously overestimated by an inherent bias introduced by multiple measurements. However, ANOVA post hoc comparison revealed that at least the best and the worst segments differed significantly for all three indices of LV haemodynamics.

Technical aspects

The study was performed in patients with persistent AF undergoing catheter ablation. All patients had safe LV access available because of clinically indicated double transseptal approach. Moreover, measurements during AF excluded the influence of variable mechanical atrioventricular delay due to pacing from different ventricular pacing sites. Higher pacing rate was used in order to augment the differences between better vs. worse pacing configurations17 and to eliminate potential fusion with intrinsic beats during AF. We also used higher pacing energy to reduce the influence of variable virtual electrode size depending on local pacing thresholds. We found previously that actual pacing energy decreases QRS duration in nonlinear fashion with negligible effect above the quadruple of the threshold value.18

Two out of three haemodynamic indices were based on derivatives of the LV pressure wave. The parameter of LV dP/dtmax has been widely used as surrogate of contractility. Recently, it has been proven as a predictor of mortality in paced patients with heart failure.19 Similarly, LV dP/dtmin has been found associated with diastolic dysfunction.20,21 Pulse pressure has been widely used as a surrogate for LV stroke volume.2224

Signal averaging of LV pressure data from 30-s recordings for each pacing site effectively minimized the influence of beat-by-beat variations of haemodynamics of any origin, including the influence of respiration and transients caused by alternation of pacing regimes. For each RV pacing site, the dP/dt data were corrected for corresponding values obtained during immediately preceding and following LV pacing episodes in order to eliminate the influence of low-frequency oscillations or drifts.

Comparison with previous studies

Our results are in agreement with several published papers that compared pacing from the RV outflow tract (RVOT) or high RV septum with the RV apex. In acute studies, pacing from the RVOT or high RV septum induced less LV dyssynchrony than the RV apical pacing.3,7,8 Several acute haemodynamic studies showed that both the RVOT and RV septal pacing increase cardiac index when compared with the RV apical pacing.911 In one study, better LV diastolic function (LV dP/dtmin and isovolumic relaxation time constant) was found with pacing from RVOT compared with RV apex pacing, but unlike similar studies, this study could not find any difference in systolic parameters.25

It remains questionable whether acute improvement projects into better long-term results. It has been demonstrated that the pacing site-related difference in dyssynchrony can persist for at least 12 months26 and that the rate of dyssynchrony after implantation appears to predict further improvement of LV ejection fraction.27

Long-term studies mostly showed that these alternative pacing sites are better than RV apex pacing.2832 Some authors tried to analyse the influence on mortality. In one retrospective study, patients with RVOT pacing had better survival33 whereas another randomized study could not demonstrate any difference.34

Limitations

Only patients with AF were included into the study. Such patients may differ from general population of candidates for pacemaker implantation and even more from candidates for cardiac resynchronization therapy. Therefore, the results of the study might only be applicable for single chamber pacing in patients with permanent AF.

The number of patients was low, so that we did not reach statistical significance in LV haemodynamic indices for small inter-segment differences. However, estimated sample size for the study with improved resolution power would be unacceptably high in this technically demanding investigation. Similarly, the study was not designed to detect the optimum pacing site in individual patients, so that no inference can be drawn concerning the suitability of the method for individualized optimization in clinical practice.

Our 18 RV pacing segments were defined without knowing exact location of the moderator band and/or the border between the inflow and outflow tract. This information could be obtained through image integration with computed tomography (CT) or magnetic resonance imaging (MRI) angiograms of the RV. However, the disagreement between CT or MRI imaging information obtained before the study and electroanatomical map created during rapid ventricular pacing would substantially interfere with correct anatomical allocation of pacing sites. We consider high-density electroanatomical map of the RV more than sufficient for reliable display of boundaries of true interventricular septum vs. RV free wall.

Finally, it is not known to what extent acutely assessed haemodynamic indices can predict long-term clinical impact in pacemaker candidates. Only limited number of studies investigated this issue. They differed from our study either in measured parameters or in analysed population. None of them exceeded 12 months in duration.26,27,35

Conclusion

The best LV mechanical performance was achieved when the RV was paced at the interventricular septum in the non-apical, mid-to-superior segments. Pacing at inferior regions of the RV free wall as well as at inferior regions of the septum exhibited the worst haemodynamic response.

Conflict of interest: J.K. is on the advisory boards of Biosense Webster, Boston Scientific, Siemens, Hansen Medical, and St. Jude Medical. He has received speakers' fees from Biotronik, Biosense Webster, Boston Scientific, Hansen Medical, Medtronic, St. Jude Medical, and Siemens; and his department has received research contracts from Biotronik, Boston Scientific, Endosense, Medtronic, and Rhythmia. All other authors declare no conflict of interest.

Funding

This work was supported by project of Charles University in Prague, Faculty of Medicine in Pilsen SVV 264 807/2012, Charles University Research Project PRVOUK-P35/LF1/5; by the project Ministry of Health, Czech Republic for conceptual development of research organization 00669806—Faculty Hospital in Pilsen, Czech Republic, grant from Ministry of Health Of Czech Republic (MZO 00023001); Grant Agency of Czech Republic (305/09/1390); EU Operational Program Prague—Competitiveness project ‘CEVKOON’ (CZ.2.16/3.1.00/22126).

References

View Abstract