Professional Documents
Culture Documents
doi:10.1093/europace/euz126
Received 13 November 2018; editorial decision 2 April 2019; accepted 5 April 2019
Cardiac resynchronization therapy (CRT) is a cornerstone of therapy for patients with heart failure, reduced left ventricular (LV) ejection
fraction, and a wide QRS complex. However, not all patients respond to CRT: 30% of CRT implanted patients are currently considered
clinical non-responders and up to 40% do not achieve LV reverse remodelling. In order to achieve the best CRT response, appropriate
patient selection, device implantation, and programming are important factors. Optimization of CRT pacing intervals may improve results,
increasing the number of responders, and the magnitude of the response. Echocardiography is considered the reference method for
atrioventricular and interventricular (VV) intervals optimization but it is time-consuming, complex and it has a large interobserver and
intraobserver variability. Previous studies have linked QRS shortening to clinical response, echocardiographic improvement and favourable
prognosis. In this review, we describe the electrocardiographic optimization methods available: 12-lead electrocardiogram; fusion-
optimized intervals (FOI); intracardiac electrogram-based algorithms; and electrocardiographic imaging. Fusion-optimized intervals is an
electrocardiographic method of optimizing CRT based on QRS duration that combines fusion with intrinsic conduction. The FOI method
is feasible and fast, further reduces QRS duration, can be performed during implant, improves acute haemodynamic response, and achieves
greater LV remodelling compared with nominal programming of CRT.
...................................................................................................................................................................................................
Keywords Cardiac resynchronization therapy • Optimization • Fusion-optimized intervals • Reverse remodelling
* Corresponding author. Tel: þ34 932 271 778; fax: 934513045. Villarroel 170, Barcelona, Catalonia 08036, Spain. E-mail address: lmont@clinic.cat
C The Author(s) 2019. For permissions, please email: journals.permissions@oup.com.
Published on behalf of the European Society of Cardiology. All rights reserved. V
Page 2 of 11 M. Pujol-López et al.
The electrocardiogram (ECG) seems therefore a potential alterna- that FOI optimization resulted in further improvement of dP/dt. Most
tive, especially since it is not observer dependent and can be easily recently, Jastrzebski et al.19 presented a retrospective mortality study
performed and repeated. that reinforces the importance of QRS shortening in CRT. They
This review will examine the available knowledge about optimiza- reported that immediate shortening of QRS duration with initiation
tion based on electrocardiographic measurements—an alternative to of CRT in patients with left bundle branch block (LBBB) strongly pre-
QRS pattern during biventricular pacing in V1–V2 and reported the following predictors of reverse
Biventricular pacing typically results in a merging of three wave fronts remodelling27:
of activation: RV, LV pacing, and intrinsic conduction. It results in a • Two BV fusion patterns in leads V1 and V2: QRS normalization or
conformational change in the QRS complex, generating a hybrid a new or an increased R wave.
morphology with recognizable features of the patterns produced by • QRS difference <_25 ms. Remodelling probability increases as
each wave front. The QRS fusion contour is intermediate in shape QRS difference takes on larger negative values. (QRS difference =
and duration between the QRS contours of the independent wave BV paced QRS LBBB QRS duration, in ms).
fronts (Figure 1C).27
During BV pacing, the QRS axis is most often directed towards the Difference in haemodynamic response
right superior quadrant, which results in a dominant R-wave in lead
in left ventricular vs. right ventricular
aVR; however, a QRS axis in the other quadrants does not necessar-
ily indicate inappropriate programming or lead position. The QRS pre-excitation in cardiac
complex during BV pacing most often has a dominant R-wave in lead resynchronization therapy
V1–V2, suggestive of contribution from LV pacing. There is a clear difference in haemodynamic response in LV vs. RV
Sweeney et al.28 described that increasing R amplitudes in V1 pre-excitation in CRT. Left ventricular pacing alone may offer theo-
through V2, indicating ventricular fusion, were associated with in- retical advantages over conventional BV pacing, potentially averting
creased probability of reverse remodelling. More recently, they per- deleterious effects from RV pacing. Verbeek et al.29 showed that LV
formed an extensive QRS analysis of the BV paced QRS morphology pre-excitation results in a haemodynamic response that is similar to
Page 4 of 11 M. Pujol-López et al.
BV pacing; their explanation is that the benefit of LV pacing primarily The implication of that observation is that the QRS width should be
originates from fusion of the LV pacing-derived wave front with that measured from the beginning of the QRS upstroke and not from the
of intrinsic conduction. The GREATER EARTH trial30 has shown that pacing spike as previously shown.33 In the initial work by Vidal et al.,34
non-responders to BV pacing may respond favourably to LV pacing, only VV interval was optimized, based on the shortest QRS, and a
suggesting a potential role as tiered therapy. fixed AV interval was programmed (Figure 2).
Along the same lines, Lumens et al.31 reported that both LV and Vidal et al.34 found that optimizing CRT devices with interventricu-
BV pacing improve systolic function of the dyssynchronous failing lar activation delay measurement showed a good correlation with
heart, to a similar extent, in experimental animals and in patients. the results obtained via echo-guided optimization in patients with
They showed for the first time that the RV myocardium contributes LBBB. Another relevant finding of this study was that the optimal VV
significantly to the improvement of LV pump function in pacing thera- interval involved in most patients was LV preactivation. This observa-
pies, especially LV pacing. During CRT, RV myocardium contributes tion is in agreement with the fact that most patients had a LBBB,
therefore, the latest activation was located at the left lateral wall, plus
to LV pump function and this contribution differs between LV pacing
the 30 ms compensation because of the pacing from the epicardium.
and BV pacing. This may explain why some patients respond better
Perego et al.35 and van Gelder et al.36 obtained, respectively, 75% and
to LV pacing and others to BV pacing, as demonstrated in the
83% of LV preactivations as the optimal VV interval.
GREATER EARTH study.
Tamborero et al.33 found that optimization of the VV interval by
measuring the QRS width from the earliest deflection obtained a bet-
Optimization of the interventricular ter acute haemodynamic response than the other VV optimization
delay using QRS width methods. Measuring the QRS width from the earliest fast deflection
The QRS width may be used as a surrogate endpoint for LV activa- was the only method that significantly improved upon the haemody-
tion time. However, pacing from the left epicardium adds a latency namic response obtained by programming a predefined VV of 0 ms in
time of 30–40 ms, the time required for the pacing front to reach the all patients (LV dP/dt 925 6 178 vs. 906 6 183 mmHg/s; P = 0.003).
endocardium and fast conducting Purkinje fibres, as demonstrated by They observed that QRS narrowing was the best marker of improved
pacing from the endocardium and epicardium at the same spot.32 acute haemodynamic response to BV pacing. This explains previous
Electrocardiographic optimization techniques in CRT Page 5 of 11
findings of a relationship between the QRS narrowing induced by and V2 is the best predictor of LV remodelling. Therefore, program-
CRT and a positive clinical response.37 ming of the VV interval should be used to create a QRS complex
with adequate contribution from LV pacing to maintain a dominant R
Electrocardiographic vs. (R, Rs, or RS pattern) in leads V1–V2.39 Leads V1–V3 are especially
echocardiographic optimization of the useful to identify the contribution from LV pacing and show the grad-
Mechanism of fusion-optimized intervals: by nominal settings (40 6 21 ms; P < 0.001). Moreover, using FOI,
pacing with fusion 86% of the patients achieved >10% shortening of the QRS, and none
had prolonged QRS duration, compared with results using the nomi-
Fusion with intrinsic rhythm during pacing is considered when LV ac-
nal values. In addition, the FOI method is faster (the duration of the
tivation is produced at least partially by intrinsic depolarization.43
optimization of the AV and VV interval was 15.5 6 7.3 min) than
Spontaneous activation over the right bundle provides a faster and
more organized RV contraction—sometimes the left bundle is not echocardiography optimization.
Regarding the invasive haemodynamic evaluation,18 the QRS nar-
completely blocked—and part of the LV can be activated from the in-
trinsic conduction. Fusion pacing creates three activation fronts, com- rowing was correlated with an increase in LV dP/dtmax independently
pared with two during pure BV pacing (Figure 5), which nominal of the baseline QRS duration. Baseline dP/dtmax improvement was
device programming does not take into account.18,44 greater using FOI (127 6 95 mmHg/s) than in nominal settings
Several studies have pointed out the potential benefit of fusion (102 6 71 mmHg/s; P = 0.05). On the other hand, FOI significantly re-
with the intrinsic rhythm. Vatasescu et al.43 used electroanatomic duced intraventricular LV dyssynchrony compared with baseline val-
mapping in 15 patients with CRT optimized by echocardiogram; BV ues, as assessed by TDI with the septal-to-lateral method and by
pacing (from the RV apex and LV) showed fusion with spontaneous decreased presence of septal flash (78% vs. 42% of patients; P < 0.01).
depolarization in eight patients, producing a higher 6-month response
rate and a more significant LVESV reduction. In patients with normal Improvement of reverse remodelling by
AV conduction, CRT with fusion is superior to any optimized BV using fusion-optimized intervals
configuration in improving LV as well as RV systolic performance.45,46 Trucco et al.8 compared echocardiographic LV reverse remodelling
Cardiac resynchronization therapy optimization with FOI method of CRT using FOI to nominal programming in a randomized study of
creates three activation wave fronts (instead of two during pure BV 180 consecutive CRT patients. Left ventricular remodelling was de-
pacing) shortening LV activation time (Figure 5).43 fined as >15% decrease in LVESV at 12-month follow-up. At
12 months, LV reverse remodelling was achieved in a larger propor-
Benefits of fusion-optimized intervals tion of the FOI group [74% vs. 53%, respectively (OR 2.02, 95% CI
Arbelo et al.18 described the method and the benefits of FOI optimi- 1.08–3.76); P = 0.026]. There was a correlation (r = 0.23; P = 0.01) be-
zation. Baseline QRS was shortened more by FOI (59 6 19 ms) than tween response and the degree of QRS narrowing: patients with the
Electrocardiographic optimization techniques in CRT Page 7 of 11
LV, left ventricle; RV, right ventricle; VV, interventricular pacing delay.
Intracardiac electrogram-based CRT trial.9 This fusion algorithm of Medtronic seems to be promising,
algorithms but more studies are needed.
QuickOpt (St Jude Medical—Abbott) calculates the optimal AV de-
QRS shortening is not maximal with nominal programming, but it
lay and VV offset based on intracardiac electrograms measured from
could be facilitated by automatic device-based algorithms in order to
the right atrium, RV, and LV leads. It determines the optimal AV and
accommodate variability in intrinsic intervals. As we have mentioned,
few studies have targeted QRS narrowing as a goal in optimization. In VV delays that achieve the best electromechanical resynchronization
the case of automatic algorithms, Varma et al.49 used a device-based of LV myocardial segments. The VV delay is estimated using the for-
algorithm (SyncAV, St Jude Medical—Abbott) that automatically mula VV = 0.5 (D þ e) (where the conduction delay D is the differ-
adjusts paced AV delay according to intrinsic AV conduction. They ence between the time of peak intrinsic activation at the LV vs. the
compared different pacing configurations: nominal simultaneous BV, RV lead; and e is the difference in VV conduction delay between two
BV þ SyncAV, and LV-only pacing þ SyncAV. BV þ SyncAV with the off- ventricular-paced propagated waveforms).57 Validation studies have
set value individualized to minimize QRS duration yielded the short- shown conflicting results.
est overall QRS duration and was the only configuration without SmartDelay (Boston Scientific) electrogram-based method calcu-
QRS prolongation in any patient. lates sensed and paced AV delays that provide maximum haemody-
There are other intracardiac electrogram-based algorithms: namic response by accounting for three inputs: intrinsic AV intervals,
AdaptivCRT, Medtronic; QuickOpt, St Jude Medical—Abbott; interventricular timing, and LV lead location.58 Ellenbogen et al.12
SmartDelay, Boston Scientific. Nevertheless, these algorithms do not showed that optimization with the SmartDelay algorithm was not dif-
verify if the narrowest QRS has been obtained. The FOI method, al- ferent from echocardiographically determined AV interval optimiza-
though it requires manual measurements, is directly focused on tion or a fixed AV delay of 120 ms.
obtaining the narrowest QRS.
AdaptivCRT (Medtronic) algorithm adapts the sensed and paced AV Electrocardiographic imaging
intervals each minute to accommodate natural variations in intrinsic As previously commented, QRS shortening is related to clinical re-
AV over time. The VV offset is then determined based on the intrinsic sponse and echocardiographic improvement. However, this parame-
AV delay and interval from the sensed RV electrogram to the end of ter is not necessarily related to a good activation sequence. Body
the QRS complex.55 The algorithm uses intrinsic intervals to provide surface activation mapping59 is a non-invasive technology based on a
RV-synchronized LV pacing when AV conduction is normal (fusion vest (or belt) with multiple body surface electrodes (Figure 6). Body
with the intrinsic activation), or BV pacing otherwise. It has shown sig- surface activation mapping simultaneously records ECG activity and
nificant increase in effective LV pacing over echo-optimized CRT56 allows obtaining an electrocardiographic image of ventricular activa-
and better clinical outcomes in a secondary analysis of the adaptive tion. The vest measures body-surface unipolar ECGs and generates
Electrocardiographic optimization techniques in CRT Page 9 of 11
7. Jansen AH, Bracke FA, van Dantzig JM, Meijer A, van der Voort PH, Aarnoudse ventricular reverse volumetric remodeling during cardiac resynchronization ther-
W et al. Correlation of echo-Doppler optimization of atrioventricular delay in apy. Circulation 2010;121:626–34.
cardiac resynchronization therapy with invasive hemodynamics in patients with 29. Verbeek XA, Auricchio A, Yu Y, Ding J, Pochet T, Vernooy K et al. Tailoring car-
heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am J diac resynchronization therapy using interventricular asynchrony. Validation of a
Cardiol 2006;97:552–7. simple model. Am J Physiol Heart Circ Physiol 2006;290:968–77.
8. Trucco E, Tolosana JM, Arbelo E, Doltra A, Castel MA, Benito E et al. 30. Thibault B, Ducharme A, Harel F, White M, O’Meara E, Guertin MC et al.;
reaching beyond left bundle branch block and left ventricular activation delay. the Frequent Optimization Study Using the QuickOpt Method (FREEDOM) trial.
J Am Heart Assoc 2018;7:e007489. Am Heart J 2010;159:944–8.
50. Linde C, Bongiorni MG, Birgersdotter-Green U, Curtis AB, Deisenhofer I, 58. Stein KM, Ellenbogen KA, Gold MR, Lemke B, Lozano IF, Mittal S et al.
Furokawa T et al. Sex differences in cardiac arrhythmia: a consensus document SmartDelay determined AV optimization: a comparison of AV delay methods
of the European Heart Rhythm Association, endorsed by the Heart Rhythm used in cardiac resynchronization therapy (SMART-AV): rationale and design.
Society and Asia Pacific Heart Rhythm Society. Europace 2018;20:1565–1565ao. Pacing Clin Electrophysiol 2010;33:54–63.