Summary
Background
Insufficient correction of mechanical dyssynchrony is a cause of non-response to cardiac resynchronization therapy (CRT).
Aims
To determine if CRT delivery could be optimized during the implantation procedure by choosing the number and location of pacing sites using echocardiography guidance.
Methods
In patients with a QRS ≥ 150 ms or a QRS < 150 ms and criteria for mechanical dyssynchrony, the objective of the implantation procedure was to shorten the left pre-ejection interval (LPEI), measured online, by at least 10 ms compared with standard biventricular configuration, by moving the right ventricular (RV) lead at different locations and, if necessary, by adding a second RV lead.
Results
Ninety-one patients (70 men; mean age 73 ± 10 years; left ventricular [LV] ejection fraction 29 ± 10%) were included. The final pacing configuration was standard biventricular in 15 (17%) patients, optimized biventricular in 22 (24%) and triple-site ventricular in 54 (59%). LPEI was shortened by ≥ 10 ms compared with standard biventricular stimulation in 73 (80%) patients. Compared with standard biventricular pacing, the final optimized pacing configuration improved global intraventricular synchrony (decreasing LPEI from 158 ± 36 ms to 134 ± 29 ms; P < 0.001), LV systolic efficiency (decreasing LPEI/LV ejection time from 0.58 ± 0.18 to 0.46 ± 0.13; P < 0.001) and LV filling (increasing LV filling time/RR from 44 ± 8% to 47 ± 7%; P < 0.001) and decreased mitral valve regurgitation.
Conclusion
Intraoperative echocardiography-guided placement of RV lead(s) during CRT implantation is feasible and acutely improves LV synchrony compared with standard biventricular stimulation.
Résumé
Contexte
La correction insuffisante de la désynchronisation mécanique est une cause de non-réponse à la thérapie de resynchronisation cardiaque (CRT).
Objectif
Déterminer si la CRT est optimisable durant l’implantation en faisant varier le nombre et le site de stimulation par guidage échocardiographique.
Méthodes
Chez des patients avec un QRS ≥ 150 ms ou avec un QRS < 150 ms et une désynchronisation mécanique, l’objectif de la procédure d’implantation était de raccourcir le délai pré-éjectionnel gauche (LPEI), mesuré en temps réel, de ≥ 10 ms par rapport à la configuration biventriculaire standard, en déplaçant la sonde ventriculaire droite (VD) et si nécessaire en ajoutant une seconde sonde VD.
Résultats
Quatre-vingt onze patients (70 hommes ; âge 73 ± 10 ans ; FEVG 29 ± 10 %) ont été inclus. La configuration de stimulation finale était biventriculaire standard chez 15 (17 %) patients, biventriculaire optimisé chez 22 (24 %) et triple-site ventriculaire chez 54 (59 %). L’objectif a été atteint chez 73 (80 %) patients. Comparée à la stimulation biventriculaire standard, la configuration de stimulation finale, optimisée, a amélioré la synchronisation intraventriculaire globale (LPEI raccourci de 158 ± 36 ms à 134 ± 29 ms ; p < 0,001), l’efficience systolique du ventricule gauche (VG) (LPEI/temps d’éjection VG diminuant de 0,58 ± 0,18 à 0,46 ± 0,13 ; p < 0,001), le remplissage VG (remplissage VG/RR augmentant de 44 ± 8 % à 47 ± 7 % ; p < 0,001) et a diminué l’insuffisance mitrale.
Conclusion
Le guidage de la mise en place de la (des) sonde(s) VD durant l’implantation de la CRT par une échocardiographique peropératoire est faisable et améliore en aigu la synchronisation VG comparé à la stimulation biventriculaire standard.
Introduction
Cardiac resynchronization therapy (CRT) is an effective treatment for patients with symptomatic heart failure, low left ventricular ejection fraction (LVEF) and a wide QRS . In controlled clinical trials, at least one third of patients do not respond clinically to CRT . The causes of non-response are various, including improper patient selection and therapy delivery. Patient selection is presently based on QRS morphology and QRS width , but this approach could be too simplistic, as some wide QRS patients do not take advantage of CRT. Additionally, narrow QRS patients do not seem to benefit from standard implantation techniques, as demonstrated in EchoCRT trial , and alternative procedures are needed to address this particular patient population.
From a global mechanical perspective, the effect of resynchronization may be viewed as improvement of left ventricular (LV) efficiency by shortening the total systole duration without shortening the LV ejection time (LVET) , with a consequent increase in LV filling time (LVFT). In a previous study in patients with a QRS < 150 ms, we showed that selecting patients with pre-existing left atrioventricular, interventricular and/or temporal but not spatial left intraventricular dyssynchrony was associated with improved clinical response to CRT , irrespective of baseline QRS width and QRS change after CRT. However, and despite the optimized selection process, effective and/or optimal correction of mechanical dyssynchrony could not be achieved in all patients using a standard biventricular pacing configuration.
Therefore, we hypothesized that the quality of resynchronization might be influenced by the number and location of pacing sites, which are difficult to anticipate before the implantation procedure.
In this study, we selected candidates for CRT and attempted to acutely optimize CRT delivery during the procedure, with echocardiographically guided mechanistic evaluation of synchrony. As optimal placement of the LV lead is challenging and as both ventricles are not electrically independent, we deliberately choose to optimize therapy delivery by moving the right ventricular (RV) lead at different locations and, if necessary, by adding a second RV lead. The objective of the optimization process was to improve LV efficiency and to decrease the left pre-ejection interval (LPEI, defined as the time interval between the onset of QRS and the onset of LV ejection) as much as possible compared with a standard biventricular configuration.
Methods
Patient selection
Patients were included in this prospective observational single-centre study if they had New York Heart Association (NYHA) class II–IV heart failure despite optimal medical therapy, an LVEF ≤ 40% and either a QRS ≥ 150 ms or a QRS < 150 ms with criteria for mechanical dyssynchrony as defined in the DESIRE study : atrioventricular dyssynchrony was defined as an LVFT/RR interval ratio < 40%; interventricular dyssynchrony was defined as an interventricular mechanical delay (IVMD, calculated as the difference between LPEI and right pre-ejection interval [RPEI]) > 40 ms; global intraventricular dyssynchrony was defined as an LPEI > 140 ms; local (lateral or septal) temporal intraventricular dyssynchrony was defined as the presence of a ‘diastolic’ contraction of the LV lateral wall or septum occurring after the aortic valve closure, sometimes with an overlap beyond opening of the mitral valve ( Fig. 1 ).
Ninety-one patients were enrolled, including 46 patients undergoing a first CRT device implantation, 31 patients upgraded from a conventional dual-chamber (DDD) pacemaker to CRT and 14 patients reoperated on for clinical non-response (persistent signs of heart failure) after previous CRT implantation.
This study was approved by our institution’s ethics committee. All patients granted their informed consent to participate.
Echocardiographical measurements
Intraoperative transthoracic echocardiography was performed using the four-chamber apical view (Vivid 7; General Electric Healthcare, Milwaukee, WI, USA). Acquisitions were done in baseline condition (i.e. without pacing in patients undergoing their first CRT device implantation, standard DDD pacing in patients upgraded from DDD pacing to CRT and biventricular pacing in patients reoperated on after previous CRT implantation) then during RV, LV and biventricular pacing configurations in random order, in VDD mode for patients in sinus rhythm and in VVI mode for patients in atrial fibrillation. Aortic ejection flow, transmitral flow, mitral valve regurgitation area and tissue Doppler images of the LV lateral wall and septum were recorded. Due to the surgical draping acquisitions for RPEI measurement were only recorded at baseline and with the final configuration. LPEI was measured online and was used to guide the implantation procedure. All other time intervals were measured offline. In addition to the time intervals described above, QRS-E (onset of QRS to onset of mitral E wave), LVET (opening to closure of the aortic valve), total duration of the systole (TDS, equal to the sum of LPEI plus LVET) and lateral-to-septal wall motion delay (LSWMD) by tissue Doppler imaging were measured. The ratio between pre-ejection and ejection durations (LPEI/LVET) was calculated. Mitral valve regurgitation was evaluated as the maximal surface of the regurgitant flow divided by the left atrial surface (MVRS/LAS).
There was no interobserver variability because all the measurements were taken by a single experienced echocardiographer. Intraobserver variability was minimized by a mean of 3–5 measures per acquisition.
Echocardiography-guided implantation procedure
Baseline LPEI was measured at admission in the operating room in all patients ( Fig. 2A ). The predefined objective of this patient-tailored procedure was to shorten the LPEI by at least 10 ms compared with standard biventricular configuration. This value was determined based on findings from the PROSPECT trial, in which the intraobserver and interobserver coefficients of variation of LPEI measurements were 3.7% and 6.5%, respectively . As a prolonged LPEI is defined as ≥140 ms, we hypothesized that 140 ms × 6.5% (i.e. 9.1 ms) should be the minimal change not related to any intra- or interobserver variability; 10 ms is a convenient approximation of this delay.
In patients undergoing a first CRT implantation, RV and LV leads were implanted in usual positions according to operator’s habits (RV lead either at the apex or septum, LV lead classically in a lateral or posterolateral vein). After recording the echocardiographical variables in this first configuration, hereafter termed ‘standard biventricular stimulation’ ( Fig. 2B ), the RV lead was moved to a distant second position and new echocardiographical data were acquired. If LPEI reduction was satisfactory in the first position and not in the second position, standard biventricular stimulation was selected and the procedure was stopped. If both RV lead positions did not achieve sufficient LPEI reduction and/or if the second was better than the first, the RV lead was further moved until ‘optimized biventricular stimulation’ was obtained. If sufficient LPEI reduction could still not be achieved with a single RV lead, a second RV lead was inserted and the best ‘triple-site ventricular configuration’ was selected ( Fig. 2C ). In case of triple-site configuration, a Y-bifurcated adapter (2872; Medtronic, Minneapolis, MN, USA) was used to connect the LV lead and a RV lead (the pacing lead in case of a CRT defibrillator or the best RV lead in case of a CRT pacemaker) to the LV port of the CRT device ( Fig. 3 ). The defibrillation lead in case of a CRT defibrillator or the other RV lead in case of CRT pacemaker was connected alone to the RV channel.
In patients upgraded from a conventional dual-chamber pacemaker, the echocardiographical variables were recorded after implantation of the LV lead. If this standard biventricular configuration was not optimal, a new RV lead was inserted and the procedure was completed to either optimized biventricular stimulation or triple-site ventricular stimulation.
In patients reoperated on after a first CRT implant, a second RV lead was systematically implanted at a site distant from the original RV lead and moved until the best optimized biventricular or triple-site ventricular stimulation was obtained.
The procedure was stopped after 2 hours of total procedural time if sufficient LPEI reduction was still not obtained.
Statistical analysis
Data are expressed as means ± standard deviations. A paired samples t test was applied to assess the differences between baseline and final measurements. A P value < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS software, version 17.0 (SPSS Inc., Chicago, IL, USA).
Results
Patient population
Ninety-one patients (70 men; 21 women) with a mean age of 73 ± 10 years were included. Thirty-two (35%) patients had been hospitalized for congestive heart failure in the previous 6 months. Twenty-three (25%) patients were in NYHA class II, 57 (63%) in class III and 11 (12%) in class IV at the time of the procedure. Heart failure was due to ischaemic cardiomyopathy in 41 (45%) patients and dilated cardiomyopathy in 50 (55%) patients. Sixteen (18%) patients had diabetes and 37 (41%) had a history of hypertension. Permanent atrial fibrillation was present in 22 (24%) patients. Mean LVEF was 29 ± 10% and mean baseline LPEI was 154 ± 40 ms. Mean QRS width was 165 ± 37 ms (141 ± 31 ms in patients undergoing first CRT device implant, 193 ± 22 ms in patients upgraded from a conventional DDD pacemaker to CRT and 184 ± 30 ms in patients reoperated on after a previous CRT). Baseline medical therapy included beta-blockers in 67 (74%) patients, an angiotensin-converting enzyme inhibitor or an angiotensin receptor blocker in 80 (88%) patients, spironolactone in 36 (40%) patients and a loop diuretic in 80 (88%) patients.
Implantation procedure
The optimization process could not be completed in three patients because of difficult manipulation of leads in a narrow costoclavicular space. There were two perioperative complications: one pericardial effusion during coronary sinus cannulation; and one pocket haematoma that did not require evacuation. The initial RV lead positioning was apical in 45 patients and septal in 46 patients. Implantation of the LV lead in a lateral or posterolateral vein was successful in all patients. The final pacing configuration was standard biventricular in 15 (17%) patients, optimized biventricular in 22 (24%) and triple-site ventricular in 54 (59%) ( Table 1 ); it was triple-site ventricular in all non-responders with a previously implanted CRT device. At the end of the procedure, LVEF increased acutely to 37 ± 12% ( P < 0.001 compared with baseline), paced QRS width shortened to 148 ± 24 ms ( P < 0.001 compared with baseline) and the QRS axis shifted rightward to 13 ± 111° ( P = 0.007) ( Table 2 ).
