Author + information
- Usha Tedrow, MD,
- William H. Maisel, MD, MPH,
- Laurence M. Epstein, MD,
- Kyoko Soejima, MD and
- William G. Stevenson, MD*,1 ()
- ↵*Brigham and Women's Hospital, Cardiovascular Division, Tower 3-B, 75 Francis Street, Boston, MA 02115,
To the Editor:Cardiac resynchronization therapy (CRT) improves symptoms and decreases hospitalizations in selected patients with heart failure (1,2). With CRT, pacing from the left ventricle (LV) alters the ventricular activation (VA) sequence, changing the QRS morphology. In responders, change in VA improves cardiac filling and ejection (3). However, many patients are nonresponders. The LV lead location is the major determinant of paced activation but is often constrained by the location of coronary sinus vein branches. During LV mapping for ventricular tachycardia, increased pacing stimulus strength (SS) has been demonstrated to capture an enlarged myocardial area, producing a larger “virtual electrode” (4). If the enlarged area of myocardial capture extends beyond a discrete region of conduction block, the SS increase could result in VA change and more rapid conduction to a remote location, such as the right ventricular (RV) apex.
We hypothesized that paced VA can be manipulated by increasing SS. We expected activation changes to be more marked when pacing near areas of scarred or infarcted myocardium.
Ten patients with New York Heart Association functional class II and III congestive heart failure, ejection fraction <40%, referred for ablation were enrolled according to protocols approved by the Brigham and Women's Hospital human subject protection committee.
Using an electroanatomic mapping system (CARTO, Biosense-Webster Inc., Diamond Bar, California), three-dimensional plots (mean of 136 ± 47 points per patient) of bipolar electrogram amplitude were created (5). Endocardial LV mapping was performed with a 7-F steerable catheter (Biosense-Webster Inc.). Electrograms were recorded on a separate digital system (Prucka Engineering Inc., Houston, Texas). Epicardial mapping by percutaneous subxiphoid approach was performed in one patient. Catheter stability was confirmed by biplane fluoroscopy, the mapping system, and continuous monitoring of electrogram morphology and timing.
Pacing at outputs of threshold, 5 mA, and 10 mA was performed at 6 to 10 separated LV sites. In one case, epicardial pacing at 20 mA was performed because of high pacing threshold (10 mA). The anode was in the inferior vena cava (IVC). A quadripolar catheter at the RV apex was the reference catheter. Atrial, ventricular, and His bundle electrograms were assessed to exclude antegrade atrial conduction during pacing.
Stimulus to RV conduction times and stimulus to QRS conduction times (QRS latency) were measured from the stimulus artifact to the first bipolar peak of the electrogram from the RV apical catheter and to the earliest VA in the surface electrocardiogram (ECG). Changes in QRS morphology with pacing were defined as any of the following: QRS width change of >40 ms; new Q-, R-, or S-wave of >25% of the total QRS amplitude; >50% change in ratio of R or S components of QRS complex; precordial transition change of more than one lead; >30° change in QRS axis (6).
Ventricular activation indices were analyzed at threshold and at 10 mA of current output. Comparisons of continuous, paired data were made using paired ttests. Where appropriate, results were adjusted using repeated measures analysis of variance to account for multiple observations in individual patients. A p value of <0.05 was considered significant. Analysis was conducted using SAS software (Version 8.2, SAS Institute Inc., Cary, North Carolina).
Electroanatomic LV endocardial voltage maps were analyzed in 10 patients with ventricular dysfunction. One patient had an additional epicardial map. The mean ejection fraction was 25 ± 9%, mean QRS duration was 151 ± 37 ms, and 80% had coronary artery disease.
Pacing was performed at 95 LV sites (Table 1).Increasing SS from threshold (3.31 ± 1.5 mA) to 10 mA reduced QRS latency (47 ± 52 ms to 32 ± 34 ms; p < 0.001) and conduction time from LV pacing site to RV apex (118 ± 62 ms to 95 ± 43 ms; p < 0.0001).
At 17 of 95 (18%) LV sites in 6 of 10 patients, a significant change in paced QRS morphology was observed. The QRS morphology changes (Fig. 1)were associated with decreases in conduction time to the RV (62 ± 76 ms vs. 14 ± 25 ms at sites with changes; p < 0.0001), QRS latency (45 ± 41 ms vs. 9 ± 26 ms; p < 0.0001), and QRS duration (18 ± 22 ms vs. 4 ± 19 ms; p = 0.03). All ventricles had abnormal low amplitude (<1.5 mV) areas. Pacing sites exhibiting QRS morphology changes with increasing SS had a lower mean electrogram amplitude of 0.8 ± 1.1 mV, compared with sites where SS did not influence QRS morphology (2.3 ± 2.1 mV; p = 0.007).
Surface QRS morphology changes identified pacing sites where VA parameters can be significantly altered with increasing SS, likely reflecting the proximity of pacing sites to discrete areas of conduction block, a more probable scenario in low-voltage areas.
Increasing SS during LV pacing reduces conduction time from the LV to the RV and can produce a sufficient change in VA to produce a change in QRS morphology. These findings are consistent with an increase in virtual electrode size as SS is increased. The effect appears most marked in low-amplitude regions, such as infarct borders (5). Although QRS duration and magnitude of reduction in QRS duration may not reliably correlate with response to CRT (7), VA change is necessary for successful resynchronization. Anatomic constraints in lead position may result in failure to respond to CRT (3). Our findings indicate that SS can be manipulated to influence VA. Potentially, resultant changes in QRS latency and morphology could be used to influence paced interventricular and intraventricular delays during CRT.
Potential limitations of application to CRT include phrenic nerve capture at high SS and shortened device battery life. Study limitations include a small patient number and a potential selection bias of electroanatomic mapping points during ablation. Pacing was endocardial in all but one patient, but epicardial pacing in one patient demonstrated similar VA parameter decreases with QRS morphology changes at three of eight pacing sites. We used unipolar pacing with an anode in the IVC to eliminate the confounding effects of anodal capture, whereas the anode in CRT devices is at the ring of the RV apical lead. The impact of SS when using current CRT systems requires further investigation with hemodynamic correlation.
Increases in SS can change paced VA. The changes can be marked and are more likely to occur in abnormal areas associated with scar or infarction that are marked by low electrogram voltage. Manipulation of the virtual electrode by altering SS may provide a simple means of adjusting CRT.
↵1 Please note: Drs. Stevenson and Soejima have participated in corporate-sponsored research for Biosense-Webster, which initially developed the electroanatomic mapping system which was used in this study.
- American College of Cardiology Foundation
- St. John Sutton M.G.,
- Plappert T.,
- Abraham W.T.,
- et al.
- Soejima K.,
- Stevenson W.G.,
- Maisel W.H.,
- et al.
- Marchlinski F.E.,
- Callans D.J.,
- Gottlieb C.D.,
- et al.
- Leclercq C.,
- Faris O.,
- Tunin R.,
- et al.