Clin Res Cardiol DOI 10.1007/s00392-014-0672-8

ORIGINAL PAPER

Not left ventricular lead position, but the extent of immediate asynchrony reduction predicts long-term response to cardiac resynchronization therapy Wolfram C. Poller • Henryk Dreger • Marius Schwerg • Hansju¨rgen Bondke Christoph Melzer



Received: 4 June 2013 / Accepted: 14 January 2014 Ó Springer-Verlag Berlin Heidelberg 2014

Abstract Cardiac resynchronization therapy (CRT) is an effective treatment for a large subgroup of chronic heart failure patients. Various attempts to improve the high nonresponder rate of 30 % by preoperative asynchrony assessment have failed. We hypothesized that superior response to CRT is correlated with greater acute reduction of asynchrony and that a concordant left ventricular (LV) lead is beneficial compared to a discordant lead. Hundred and eight consecutive CRT patients from our center were prospectively included. Clinical status and asynchrony parameters were assessed before, 1 day and 6 months after CRT implantation. Super-response was defined as an increase of the LV ejection fraction by C15 % and a decrease in LV end systolic volume (LVESV) by C30 %. When the criteria for super-response were not met, average response was given with a decrease of baseline LVESV C15 %. Sixty eight patients were classified as responders (63 %). Comparing super- (n = 19) and average (n = 49) responders, we found that greater acute reduction of LV asynchrony (change of asynchronous segments under CRT: -1.3 vs. -0.4, p \ 0.05; decrease of LV intraventricular delay: -34 ms vs. -16 ms, p \ 0.05) is associated with superior reverse remodeling after 6 months. Importantly, asynchrony parameters of super-, average and nonresponders were almost identical at baseline. A concordant LV lead (n = 63) was not associated with improved LV reverse remodeling compared to a discordant lead (n = 28): LVEF: ?8.6 % vs. ?7.8 %, p = 0.91; LVESV:

-30.5 ml vs. -23.8 mL, p = 0.84. A greater immediate reduction of LV asynchrony predicts superior response. Preoperative asynchrony parameters do not correlate with outcome. A concordant LV lead is not superior to a discordant lead. Keywords Cardiac resynchronization therapy  Asynchrony assessment  CRT super-responders  Left ventricular intramural delay  LV lead position Abbreviations AVO Aortic valve opening CRT Cardiac resynchronization therapy IVMD Interventricular mechanical delay LAO Left anterior oblique LV Left ventricle LVEF Left ventricular ejection fraction LVESV Left ventricular end systolic volume NYHA New York Heart Association RAO Right anterior oblique RV Right ventricle SD Standard deviation TDI Tissue Doppler imaging TSI Tissue synchronization imaging

Introduction W. C. Poller (&)  H. Dreger  M. Schwerg  H. Bondke  C. Melzer Medizinische Klinik mit Schwerpunkt Kardiologie und Angiologie (CCM), Charite´ Universita¨tsmedizin Berlin, Charite´platz 1, 10117 Berlin, Germany e-mail: [email protected]

Cardiac resynchronization therapy (CRT) is an established treatment for an increasing number of chronic heart failure patients [1–5]. A major unresolved issue of CRT is the high rate of non-responders (around 30 %) [1–5]. Since CRT is an expensive treatment with relevant complication rates,

123

Clin Res Cardiol

the identification of potential responders is of great importance. Current guidelines recommend only NYHA class, ejection fraction as well as QRS morphology and duration to be considered in the selection of patients for CRT [6, 7]. The idea of preselecting patients by echocardiographic determination of asynchrony has not yet been included. This is mainly based on the PROSPECT trial, a large multi-center study analyzing twelve echocardiographic parameters for asynchrony detection that entirely failed to predict response to CRT [8]. Furthermore, studies assessing the left ventricular (LV) segment with the latest mechanical activation and analyzing a concordant versus discordant LV lead placement reported controversial results [9–11]. Consequently, a recommendation to implant the LV lead concordantly in all patients has not yet been included into the guidelines [6, 7]. From these data one might question whether asynchrony assessment by echocardiography is per se unreliable or whether the relevance of mechanical asynchrony for CRT response is overestimated. Taken the basic principle of CRT, which is reduction of asynchrony, we think that changes in pre- and postoperative asynchrony are yet important determinants of CRT response. Currently, data on the impact of acute postoperative asynchrony reduction and concordant vs. discordant LV lead positions on CRT response are sparse. Here, we hypothesize that superior response to CRT is correlated with a greater acute reduction of asynchrony and that a concordant LV lead is beneficial compared to a discordant lead. To analyze these issues we initiated the Charite´ CRT registry.

Table 1 Baseline characteristics (super- vs. average vs. nonresponders) Superresponders

Average responders

Nonresponders

n (%)

19 (17.6)

49 (45.4)

40 (37)

Males, n (%)

12 (63.2)

39 (79.6)

37 (92.5)*

Age, years Ischemic cardiomyopathy, n (%)

65.5 ± 7.8 7 (36.8)

69.5 ± 8.6 12 (24.5)

69.5 ± 9.8 15 (37.5)

Sinus rhythm, n (%)

16 (84.2)

33 (67.3)

30 (75.0)

When appropriate, data are given as mean ± standard deviation * p \ 0.05 vs. super-responders

medication, acute decompensated heart failure, life expectancy below 1 year, pregnancy, or inability to give written consent were excluded [6, 7]. According to their echocardiographic 6 months outcome, patients were grouped into super-, average and non-responders. Super-response was defined as an absolute increase of ejection fraction C15 % and a relative decrease of baseline LV end systolic volume (LVESV) C30 % 6 months after implantation. Average response was given with a decrease of LVESV C15 % when the criteria for super-response were not met. Non-response was defined as a relative decrease of LVESV \15 %. The study is in agreement with local university ethic guidelines and the Declaration of Helsinki. Baseline characteristics of the three groups of patients are shown in Tables 1 and 3.

Methods Echocardiography Hypotheses The aim of our study was to evaluate two hypotheses. First: a greater acute reduction of asynchrony is associated with superior response to CRT. Second: a concordant LV lead placement improves asynchrony reduction and reverse remodeling.

All patients were examined on a Vivid 7 Ultrasound System (GE Medical Systems, Norway). Left ventricular ejection fraction (LVEF) was determined using the Simpsons biplane approach. To exclude interobserver variability, all examinations were performed by one physician at the following time points: before, 1 day, and 6 months after CRT implantation.

Study population Asynchrony assessment Within a period of 1.5 years (2010–2011), 108 CRT devices were implanted in our center. Data from all 108 consecutive patients were collected prospectively in the Charite´ CRT registry. Inclusion criteria for the study were symptomatic heart failure (NYHA class II– IV) despite optimal medical therapy, a reduced LV ejection fraction B35 % and a wide QRS C120 ms [6, 7]. Patients with recent myocardial infarction (\3 months), \3 months on optimal heart failure

123

Inter- and intraventricular asynchrony were determined as described previously [12, 13]. Analyses were performed before and 1 day after CRT implantation. As per definition, asynchrony was present when the interventricular mechanical delay (IVMD) was longer than 40 ms. To identify this delay, the right and LV delays were determined separately by measuring the interval between the beginning of the QRS complex

Clin Res Cardiol

and the onset of the pulmonary and aortic outflow, respectively. Intraventricular asynchrony was assessed by analyzing the interval between the aortic valve opening (AVO) and the peak systolic velocity (S0 ) measured by tissue Doppler imaging (TDI) in the six basal LV segments (anterior, anteroseptal, septal, inferior, posterior, lateral). For plausibility control, tissue synchronization imaging (TSI) was performed to confirm correct determination of S0 . The segment with the shortest AVO-S0 interval was chosen as a reference for intact myocardium. To quantify the asynchrony of each segment, we calculated the time difference between the AVO-S0 interval of the individual segment and the AVO-S0 interval of the reference segment. As a global indicator for the severity of LV intraventricular asynchrony, the longest intraventricular delay (i.e. the difference between the shortest and the longest AVO-S0 intervals) was calculated [12, 13]. AV- and VV-delay optimization According to the guidelines, we performed atrioventricular (AV)- and interventricular (VV)-delay optimization in all patients [6, 7, 14]. Initially, VV optimization was performed. Measurements were first taken under synchronous biventricular pacing. If residual asynchrony was present, LV and RV pacings were optimized to the VV delay resulting in the smallest inter- and intraventricular asynchrony. In a second step, the AV delay was optimized using a method published by Cleland et al. [3]. First, a long AV interval (e. g., 75 % of the intrinsic AV interval) was programmed. Then, the AV interval was decreased by 20-ms steps until an A-wave truncation occurred. Subsequently, the AV interval was increased again in 10-ms steps to optimize AV hemodynamics.

Lead position Left ventricular lead positions were fluoroscopically determined as suggested by the 2007 CRT guidelines in left anterior oblique (LAO) 40° and right anterior oblique (RAO) 30° [6]. In LAO 40° the LV lead position was classified as posterior, posterolateral, lateral, anterolateral, and anterior. In RAO 30°, the position was classified as basal, midventricular or apical. Right ventricular lead position was classified as septal or apical. The lead positions were confirmed by subsequent postoperative chest X-rays in p.a. and lateral projections. Concordant versus discordant lead placement A concordant versus discordant lead placement was determined by comparing the position of the LV lead with the location of the latest activated LV segment (longest AVO-S0 interval). A concordant lead position is given when the LV lead is located directly within the latest activated LV segment (either basal, midventricular or apical). An adjacent position is given when the lead is located in the segment directly beside the latest activated LV segment. For the present analyses, patients with adjacent leads were analyzed in the concordant group. Discordant positions are characterized by at least one segment between the LV lead position and the latest activated LV segment. Determination of the latest activated LV segment and the LV lead positions is described above. Patients with septal or anteroseptal asynchrony and a septal RV lead were also classified as concordant. Patients without clearly asynchronous segments were excluded from concordance analyses. Their data are reported separately in the ‘‘Results’’ section. Statistical analyses

CRT implantation CRT devices (Biotronik, Germany and Medtronic, USA) were implanted either subcutaneously or submuscularly in the region of the left or right pectoralis major muscle. Both single- and dual-coil defibrillator leads were used, as there is not yet a clear recommendation [15]. Electrodes were introduced via the axillary and subclavian vein and the LV lead was placed in an anatomically suitable cardiac vein with a good capture threshold and without phrenic nerve stimulation. Primary target was a lateral or posterolateral branch in the midventricular area, however, if technically necessary other cardiac veins were accepted. The operators were aware of the pre-OP asynchrony assessment. Two experienced board certified cardiologists performed implantations.

When appropriate, data are expressed as mean ± standard deviation (SD). Statistical tests were performed using SigmaStat (SigmaStat 3.0, SPSS, Inc.). To calculate significance the following tests were performed as appropriate: V2 tests, t tests, paired t tests, and Mann–Whitney Rank Sum Test (for comparison of two groups) as well as V2 tests and one-way ANOVA (for comparison of three groups). An error probability of p \ 0.05 was considered statistically significant. Diagrams depict box plots with the lower boundary indicating the 25th percentile, the upper boundary indicating the 75th percentile and a line within the box indicating the median. Whiskers above and below the box represent the 90th and 10th percentiles, respectively. Statistical outliers are marked by dots (SigmaPlot 9.0, SPSS, Inc.).

123

Clin Res Cardiol

Results

A 4

Table 2 Clinical status and asynchrony at baseline (super- vs. average vs. non-responders) Superresponders (n = 19)

Average responders (n = 49)

Nonresponders (n = 40)

NYHA class

2.7 ± 0.5

2.6 ± 0.5

2.9 ± 0.6

LVEF (%)

23.7 ± 6.0

26.4 ± 6.5

24.4 ± 5.8

LVESV (ml)

128 ± 40

144 ± 51

137 ± 57

QRS (ms)

154 ± 18

161 ± 22

167 ± 24

LBBB (%) RBBB (%)

16 (84.3) 1 (5.2)

34 (69.4) 1 (2.0)

28 (70.0) 2 (5.0)

IVCD (%)

2 (10.5)

14 (28.6)

10 (25.0)

IVMD (ms)

56 ± 23

51 ± 24

49 ± 20

Asynchronous segments, n

1.6 ± 1.0

1.8 ± 1.1

1.8 ± 1.1

Intraventricular delay (ms)

77 ± 29

84 ± 31

76 ± 29

Data are given as mean ± standard deviation LVEF left ventricular ejection fraction, LVESV left ventricular end systolic volume, IVMD interventricular mechanical delay, LBBB left bundle branch block, RBBB right bundle branch block, IVCD intraventricular conduction delay * p \ 0.05 vs. super-responders

123

NYHA class

3

2

1

super responders

B

average responders

nonresponders

70

baseline

60

follow-up 50

LVEF [%]

In our registry, 19 patients (17.6 %) fulfilled the criteria for super-response while 49 patients (45.3 %) were classified as average responders. We observed a non-responder rate of 37 % (40 patients). Baseline characteristics of super-, average, and non-responders are presented in Table 1. The three groups show only one statistically significant baseline difference (Table 1): the percentage of female patients among the super-responders was significantly higher than the percentage among the non-responders (36.8 vs. 7.5 %; p = 0.023). In addition, there was a non-significant trend towards a higher percentage of patients in sinus rhythm in the super-response group. NYHA class at baseline was similar in all three groups (2.7 in super-responders vs. 2.6 in average responders vs. 2.9 in non-responders, Table 2). When focusing on LV function and geometry, no preexisting differences could be identified. The LV ejection fraction at baseline was similar (23.7 % in super-responders vs. 26.4 % in average responders vs. 24.4 % in nonresponders, Table 2). The LVESV was also not significantly different (super-responders 128 mL vs. average responders 144 mL vs. non-responders 137 mL, Table 2). Importantly, the main asynchrony parameters, QRS duration (154 ms in super-responders vs. 161 ms in average

follow-up

40 30 20 10 0

super responders

C

400

average responders

nonresponders

average responders

nonresponders

baseline follow-up

LVESV [mL]

Clinical status and asynchrony of super-, average, and non-responders at baseline

baseline

300

200

100

0

super responders

Fig. 1 Changes in clinical status of super-, average and nonresponders from baseline to 6-month follow-up. LVEF left ventricular ejection fraction, LVESV left ventricular end systolic volume

responders vs. 167 ms in non-responders), IVMD (56 ms in super-responders vs. 51 ms in average responders vs. 49 ms in non-responders), number of asynchronous LV segments (1.6 in super-responders vs. 1.8 in average responders vs. 1.8 in non-responders) and longest intraventricular delay (77 ms in super-responders vs. 84 ms in average responders vs. 76 ms in non-responders) did not differ significantly at baseline (Table 2). Regarding QRS morphology, the majority of patients in our registry had a left bundle branch block (LBBB). An unspecific

Clin Res Cardiol

intraventricular conduction delay (IVCD) was observed in about one quarter of the patients, while only four patients had a right bundle branch block (RBBB) (LBBB 72.2 %; IVCD 24.1 %; RBBB 3.7 %). Between the three response groups, we found a non-significant trend towards a higher prevalence of LBBB in the super-response groups (84.2 % in super-responders; 69.4 % in average responders and 70.0 % in non-responders) (Table 2). Response of super-, average and non-responders at 6-month follow-up

250 200

To test our first hypothesis, we compared the three groups regarding changes in the four described asynchrony parameters. Accordingly, these parameters were measured before and 1 day after CRT implantation using our protocol as published previously [13]. Regarding the global asynchrony indicator QRS duration, the reduction after CRT was not significantly different between the groups (-32 ms in superresponders vs. -31 ms in average responders vs. -33 ms in non-responders, Fig. 2a). The IVMD decreased strongly in all groups but also without significant differences (-37 ms in super-responders vs. -35 ms in average responders vs. -35 ms in non-responders; Fig. 2b). In contrast to these rather global parameters, the reduction of LV intraventricular asynchrony showed relevant differences between the groups. While the super-responders had a significantly greater reduction of LV asynchrony, the reduction was almost similar between average and non-

C

150 100 50

Acute postoperative asynchrony reduction in super-, average and non-responders

asynchronous segments [n ± SEM]

A QRS duration [ms]

Six months after implantation of the CRT devices, superand average responders showed a significant improvement of NYHA class, LVEF and LVESV (Fig. 1). The nonresponders only had a slight improvement of NYHA class, while LVEF and LVESV remained unchanged (Fig. 1). As expected, the absolute change in LV ejection fraction was the largest in the super-response group (?24.0 % in superresponders vs. ?9.3 % in average responders vs. ?0.4 % in non-responders; p \ 0.05, Fig. 1b). The main indicator for reverse remodeling, the LVESV, decreased dramatically and much stronger in super- compared to average responders (-68 mL in super-responders vs. -42.0 mL in average responders, ?3 mL in non-responders; p \ 0.05, Fig. 1c). NYHA class improved significantly in all three groups. The improvement was significantly greater in super- compared to average responders (-1.3 NYHA

classes in super-responders vs. -0.6 NYHA classes in average responders, p \ 0.05, Fig. 1a). Even among the echocardiographic non-responders, there was a small but significant improvement of -0.4 NYHA classes (p \ 0.05, Fig. 1a). In summary, super- and average responders showed a positive response to CRT but with relevant differences in the achieved improvement (Fig. 1).

baseline follow-up

0

average responders

nonresponders

D

B

IVMD [ms]

intraventricular delay [ms]

baseline follow-up

50

0

super responders

average responders

nonresponders

baseline follow-up

1

0

super responders

100

2

super responders

average responders

nonresponders

200 baseline follow-up

150

100

50

0

super responders

average responders

nonresponders

Fig. 2 Acute postoperative asynchrony reduction in super-, average and non-responders. IVMD interventricular mechanical delay

123

Clin Res Cardiol

responders. The number of asynchronous segments decreased much stronger in the super-response group (-1.3 segments in super-responders vs. -0.4 segments in average responders vs. -0.4 segments in non-responders, p \ 0.05; Fig. 2c). This led to an almost complete LV resynchronization in super-responders (mean residual number of asynchronous segments: 0.3; Fig. 2c). In addition, the decrease of the longest intraventricular delay was likewise significantly greater among super-responders (-33 ms in superresponders vs. -16 ms in average responders vs. -15 ms in non-responders, p \ 0.05; Fig. 2d). Clinical status and asynchrony of patients with concordant and discordant leads at baseline To test our second hypothesis, we analyzed the effect of a concordant versus discordant lead placement on acute asynchrony reduction and LV reverse remodeling at 6 months follow-up. Among all patients in the registry (n = 108), 17 patients were excluded from these analyses because they did not have a clearly localized intraventricular asynchrony. The results of these 17 patients are reported separately in the last paragraph of this section. The remaining 91 patients, including super-, average, and non-responders, had at least one asynchronous LV segment. A concordant lead placement was found in 63 patients (69.2 %), while 28 patients (30.8 %) had a discordant lead position. In the short axis, the position of the LV lead was identified as posterior (6 %), posterolateral (40 %), lateral (47 %), anterolateral (5 %) and anterior (2 %). In the long axis, the LV lead was located basal (11 %), midventricular (72 %) and apical (17 %). The RV lead was in a septal position in 45.1 % and in an apical position in 53.8 % of the patients. The segment with the latest contraction was most frequently lateral (27.9 %), followed by septal (26.7 %), posterior (18.6 %), inferior (17.4 %) and anteroseptal (9.3 %). The two groups showed no statistically significant baseline differences regarding age, gender, prevalence of ischemic cardiomyopathy, and sinus rhythm (Table 3). The functional parameters including NYHA class, LV ejection fraction and LVESV as well as most asynchrony parameters (IVMD, number of asynchronous segments and longest intraventricular delay) were not significantly different at baseline (Table 4). The only baseline difference existed regarding the QRS duration, with a slightly longer QRS in the concordant group (165.2 vs. 155.7 ms, p = 0.04; Table 4). Response of patients with concordant and discordant leads at 6 months follows up At 6-month follow-up, the improvement in NYHA class was almost identical in both groups (-0.6 NYHA classes

123

Table 3 Baseline characteristics (concordant vs. discordant LV lead) Concordant

Discordant

p value

n (%)

63 (69.2)

28 (30.8)



Males, n (%)

51 (80.0)

24 (85.7)

0.78

Age, years

69.2 ± 9.5

67.3 ± 9.3

0.35

Ischemic cardiomyopathy, n (%)

19 (30.2)

11 (39.3)

0.91

Sinus rhythm, n (%)

46 (73.0)

22 (78.6)

0.84

When appropriate, data are given as mean ± standard deviation * p \ 0.05 vs. concordant patients

Table 4 Clinical status and asynchrony at baseline (concordant vs. discordant lead) Concordant

Discordant

p value

NYHA class LVEF (%)

2.7 ± 0.6 25.3 ± 6.1

2.7 ± 0.5 26.6 ± 5.8

0.74 0.32

LVESV (ml)

137.1 ± 57.3

132.8 ± 42.8

0.72

QRS (ms)

165.2 ± 23.0

155.7 ± 15.3*

0.04

IVMD (ms)

48.9 ± 23.9

55.8 ± 20.2

0.20

Asynchronous segments, n

2.14 ± 0.91

2.04 ± 0.69

0.50

Intraventricular delay (ms)

90.3 ± 26.7

83.9 ± 21.3

0.43

Data are given as mean ± standard deviation LVEF left ventricular ejection fraction, LVESV left ventricular end systolic volume, IVMD interventricular mechanical delay * p \ 0.05 vs. concordant patients

in the concordant group vs. -0.7 NYHA classes in the discordant group, p = 0.55; Fig. 3a). The increase of ejection fraction was also similar (?8.6 % in the concordant group vs. ?7.8 % in the discordant group, p = 0.91, Fig. 3b). The decrease in LVESV was not significantly greater in the concordant group (-30.5 mL in the concordant group vs. -23.8 mL in the discordant group, p = 0.84, Fig. 3c). Acute postoperative asynchrony reduction in patients with concordant and discordant leads The global asynchrony parameters QRS duration and IVMD decreased to a similar extent (Fig. 4a, b). When focusing on intraventricular asynchrony reduction, we found that the mean reduction in the number of asynchronous segments was not significantly different between the two groups (-0.90 segments in the concordant group vs. -0.87 segments in the discordant group, p = 0.80; Fig. 4c). In line with this, the reduction of the longest intraventricular delay did also not differ significantly

Clin Res Cardiol

A

NYHA class

4

3

2

baseline 1

follow-up concordant

B

discordant

70

baseline 60

follow-up

LVEF [%]

50

remaining patients enrolled in the study (5.9 % vs. 32.9 %, p \ 0.05). Interestingly, LV synchronous patients had a similar extent of interventricular asynchrony compared to the LV asynchronous patients. At 6-month follow-up, patients without intraventricular asynchrony at baseline had improved similarly compared to the baseline asynchronous patients (NYHA class -0.8, LVEF ?10.0 %, LVESV -34.5 mL). The decrease of QRS duration was also similar (-30.6 ms in LV synchronous vs. -33 ms in LV asynchronous patients), while interventricular asynchrony decreased significantly stronger in LV synchronous compared to LV asynchronous patients (IVMD -40.0 ms vs. -34.0 ms, p = 0.014). Among the 17 patients without LV asynchrony at baseline, 17.6 % were classified as super-responders, 47.1 % were average responders, while 35.3 % were non-responders. This resembles the overall distribution in our registry.

40 30

Discussion

20 10 0

concordant

C

discordant

baseline

400

follow-up LVESV [mL]

300

200

100

0

concordant

discordant

Fig. 3 Response of patients with concordant and discordant lead positions at 6-month follow-up. LVEF left ventricular ejection fraction, LVESV left ventricular end systolic volume

(-28.0 ms in the concordant group vs. -26.8 ms in the discordant group, p = 0.49; Fig. 4d). Patients without relevant mechanical LV asynchrony at baseline In our study, 17 patients had no significant intraventricular asynchrony at baseline (number of asynchronous segments 0, longest intraventricular delay 34 ms). In regard to baseline parameters, these patients suffered significantly less from ischemic heart disease as compared to the

In our study, 17.6 % of the patients were identified as super-responders 6 months after device implantation. This is in the middle of the range of previously reported rates of super-response (9.7–37.8 %) [16–20]. The rate of nonresponders (37 %) is also in the range known from other CRT studies [1–5, 19, 21]. The definitions of the response groups by echocardiographic changes of LVEF and LVESV are widely used and accepted in studies analyzing CRT response [1–5, 16–18]. Of note, the percentage of patients with ischemic cardiomyopathy was comparatively low among our patients (31.5 %) [19–25]. When we compared asynchrony parameters of super-, average and non-responders at baseline, we found no significant differences in any asynchrony parameter even between super- and non-responders (Table 2). The observation of similar baseline asynchrony is important as it readily explains the problem of response prediction by preoperative asynchrony assessment [8]. In contrast to that great differences existed regarding the extent of asynchrony reduction immediately after CRT implantation, with a much stronger reduction in the superresponse group (Fig. 2). Thus, the present study clearly shows that a greater acute reduction of LV asynchrony measured by TDI is correlated with a superior reverse remodeling after 6 months (Figs. 1, 2). Long-term CRT response depends on the immediate reduction of asynchrony. This is interesting as most studies on response prediction assessed asynchrony reduction only 6 months after implantation—at which point it is also correlated with superior response [26–28]. In line with our data, one previous study also analyzed asynchrony reduction immediately after CRT implantation and found that an

123

Clin Res Cardiol

C

250 200 150 100 50

3

asynchronous segments [n ± SEM]

QRS duration [ms]

A

baseline

baseline follow-up 2

1

follow-up 0

0

concordant

D

150

baseline

IVMD [ms]

follow-up 100

50

0

concordant

concordant

intraventricular delay [ms]

B

discordant

discordant

discordant

200

baseline follow-up 150

100

50

0

concordant

discordant

Fig. 4 Acute postoperative asynchrony reduction in patients with concordant and discordant lead positions. IVMD interventricular mechanical delay

acute LV resynchronization above 20 % predicts superior response to CRT at 6-month follow-up [29]. We conclude that our first hypothesis is well confirmed by the presented data. While our findings do also not help to pre-identify responders, they allow a better prediction of long-time response immediately after CRT implantation. However, things are more complicated as we observed that the 17 patients without any baseline LV asynchrony improved to a similar extent compared to the patients with baseline LV asynchrony. This illustrates that both inter- and intraventricular asynchronies have an important impact on CRT response. This makes response prediction even more difficult. To test our second hypothesis, LV reverse remodeling and asynchrony reduction of patients with concordant and discordant leads were compared. At 6-month follow-up, improvement in NYHA class, increase of LVEF and decrease in LVESV were similar between the concordant and discordant group (Fig. 3). In line with that, the immediate reduction of inter- and intraventricular asynchrony also showed no significant differences (Fig. 4a, b). The literature on the effect of a concordant versus discordant lead placement is controversial. Most recently, two randomized single-center studies analyzing the effect of an echo-guided LV lead positioning to the most delayed segment described an outcome benefit in the echo-guided group [10, 11]. Based on the TARGET study the new 2013

123

ESC Guidelines on cardiac pacing and CRT include a class IIb recommendation that LV lead placement may be targeted at the latest activated LV segment. However, in contrast to that but in good agreement with our data, other studies found no benefit of a concordant LV lead placement [9]. Although there are now data from two randomized single-center trials favoring a concordant LV lead placement [10, 11], we still think that various concerns need to be raised against an echo-guided concordant approach for all patients. A concordant placement in all patients would surely increase CRT implantation times [30]. In addition, it would require a detailed pre-evaluation by challenging echocardiographic techniques to determine the latest activated LV segment [30]. Problems regarding these evaluations when performed in clinical routine became obvious in the PROSPECT study [8]. The results of our and many previous studies raise the principle question why left ventricles with very similar baseline asynchrony parameters show such a varying response to CRT. Furthermore, it remains unclear why concordant LV leads are not in general superior to discordant leads. On the one hand, there are the better understood and obvious ‘‘pure’’ electrical delay problems, including LBBB, RBBB and unspecific slowing within the conduction system. These types of electrical delays can be easily detected in the surface ECG. On the other hand, there seem to be local effects at tissue, cellular and

Clin Res Cardiol

subcellular levels that modify the effects of electrical resynchronization on the homogeneity of mechanical contraction. These effects are numerous and by far not fully understood. They include prolongation of action potential duration, potassium channel down-regulation [33–38], and reduction of wall stress leading to downregulation of stress response kinases, changes in calcium homeostasis, improvement of mitochondrial efficiency, decreased apoptosis, and increased beta-receptor sensitivity [31, 40–42] but also alterations of the sympathetic nervous systems [39] and many more. This complex pathophysiology might explain why reliable predictors of CRT response are currently missing and can probably not just be found in the surface ECG and echocardiogram. If not even the response can be predicted in general, it is understandable that the determination of the optimal right and LV pacing sites for an individual heart is even more challenging. Moving in the right direction, one promising method to visualize electromechanical delay is a modified NOGA system [32]. NOGA is a catheter-based, non-fluoroscopic, three-dimensional endocardial mapping system. This technique allows simultaneous assessment of both local electrical activation and regional contractility [32]. However, its usefulness in predicting CRT response, guiding LV lead positioning and fine tuning of VV interval settings remains to be demonstrated [32]. Limitations In our study, TDI was used to assess asynchrony and to detect the latest activated LV segment. Recently, large studies including TARGET and STARTER [10, 11] evaluated the intraventricular asynchrony using 2D speckle tracking. Thus, a direct comparison with these studies is difficult. The current guidelines do not favor one of the methods. In our study, 95 % of the patients were successfully analyzed with TDI [13]. Another limitation is that the current study did not analyze functional capacity, quality of life scores and hard endpoints including death and hospitalization for heart failure. This is based on the intention of the study to analyze the mechanisms of CRT regarding asynchrony reduction and the stagnating rate of non-responders. Moreover, multivariate analysis would be the more comprehensive approach to study CRT response in continuous variables, not categories. However, this would have required much more patients.

Conclusion In the current study, like in many previous, we were unable to identify parameters to increase the rate of CRT responders. However, we clearly demonstrate that positive

response depends on asynchrony reduction. Furthermore, we found that the acute postoperative asynchrony reduction predicts long-term response to CRT. Unfortunately, the success of CRT cannot be predicted by preoperative asynchrony assessment since baseline asynchrony parameters are similar between super-, average, and nonresponders. A concordant LV lead placement does not always lead to a greater reduction of asynchrony and a discordant lead can sometimes resynchronize the entire ventricle. This is probably based on the individual intramural delay, which seems to be one reason for the stagnating rate of CRT responders. We hypothesize that we will not get significantly better in CRT patient selection until methods to measure the individual intramural delay in clinical routine are established. So far, the rate of responders to this nevertheless amazing therapy for chronic heart failure has to be accepted. Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest.

References 1. Abraham WT, Fisher WG, Smith AL et al (2002) Cardiac resynchronization in chronic heart failure. N Engl J Med 346:1845–1853 2. Bristow MR, Saxon LA, Boehmer J et al (2004) Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med 350:2140–2150 3. Cleland JG, Daubert JC, Erdmann E et al (2005) The effect of cardiac resynchronization on morbidity and mortality in heart failure. N Engl J Med 352:1539–1549 4. Linde C, Leclercq C, Rex S et al (2002) Long-term benefits of biventricular pacing in congestive heart failure: results from the MUltisite STimulation in cardiomyopathy (MUSTIC) study. J Am Coll Cardiol 40:111–118 5. Moss AJ, Hall WJ, Cannom DS et al (2009) Cardiac-resynchronization therapy for the prevention of heart-failure events. N Engl J Med 361:1329–1338 6. Vardas PE, Auricchio A, Blanc JJ et al (2007) European practice guidelines on cardiac pacemakers and cardiac resynchronization therapy. Working Group of the European Society of Cardiology (ESC) on cardiac pacemakers and cardiac resynchronization therapy. Developed in collaboration with the European Heart Rhythm Association. Rev Esp Cardiol 60:1272.e1–1272.e51 7. McMurray JJ, Adamopoulos S, Anker SD et al (2012) ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2012: The Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association (HFA) of the ESC. Eur J Heart Fail 14:803–869 8. Chung ES, Leon AR, Tavazzi L et al (2008) Results of the Predictors of Response to CRT (PROSPECT) trial. Circulation 117:2608–2616 9. Fung JW, Lam YY, Zhang Q et al (2009) Effect of left ventricular lead concordance to the delayed contraction segment on echocardiographic and clinical outcomes after cardiac resynchronization therapy. J Cardiovasc Electrophysiol 20:530–535

123

Clin Res Cardiol 10. Khan FZ, Virdee MS, Palmer CR et al (2012) Targeted left ventricular lead placement to guide cardiac resynchronization therapy: the TARGET study: a randomized, controlled trial. J Am Coll Cardiol 59:1509–1518 11. Saba S, Marek J, Schwartzman D, et al. (2013) Echocardiography-Guided Left Ventricular Lead Placement for Cardiac Resynchronization Therapy: Results of the Speckle Tracking Assisted Resynchronization Therapy for Electrode Region (STARTER) Trial. Circ Heart Fail 12. Cazeau S, Bordachar P, Jauvert G et al (2003) Echocardiographic modeling of cardiac dyssynchrony before and during multisite stimulation: a prospective study. Pacing Clin Electrophysiol 26:137–143 13. Dreger H, Borges AC, Ismer B et al (2009) A modified echocardiographic protocol with intrinsic plausibility control to determine intraventricular asynchrony based on TDI and TSI. Cardiovasc Ultrasound 7:46 14. Eberhardt F, Hanke T, Fitschen J et al (2012) AV interval optimization using pressure volume loops in dual chamber pacemaker patients with maintained systolic left ventricular function. Clin Res Cardiol 101(8):647–653 15. Neuzner J (2012) Carlsson J Dual- versus single-coil implantable defibrillator leads: review of the literature. Clin Res Cardiol 101(4):239–245 16. Rickard J, Kumbhani DJ, Popovic Z et al (2010) Characterization of super-response to cardiac resynchronization therapy. Heart Rhythm 7(7):885–889 17. Reant P, Zaroui A, Donal E et al (2010) Identification and characterization of super-responders after cardiac resynchronization therapy. Am J Cardiol 105:1327–1335 18. Antonio N, Teixeira R, Coelho L et al (2009) Identification of ‘super-responders’ to cardiac esynchronization therapy: the importance of symptom duration and left ventricular geometry. Europace 11:343–349 19. Castellant P, Fatemi M, Bertault-Valls V et al (2008) Cardiac resynchronization therapy: ‘‘nonresponders’’ and ‘‘hyperresponders’’. Heart Rhythm 2:193–197 20. Gasparini M, Regoli F, Ceriotti C et al (2008) Remission of left ventricular systolic dysfunction and of heart failure symptoms after cardiac resynchronization therapy: temporal pattern and clinical predictors. Am Heart J 155(3):507–514 21. Reithmann C, Herkommer B, Huemmer A et al (2013) The risk of delayed atrioventricular and intraventricular conduction block following ablation of bundle branch reentry. Clin Res Cardiol 102(2):145–153 22. Wasmer K, Kobe J, Andresen D et al (2013) Comparing outcome of patients with coronary artery disease and dilated cardiomyopathy in ICD and CRT recipients: data from the German DEVICE-registry. Clin Res Cardiol 102(7):513–521 23. Schau T, Koglek W, Brandl J et al (2013) Baseline vectorcardiography as a predictor of invasively determined acute hemodynamic response to cardiac resynchronization therapy. Clin Res Cardiol 102(2):129–138 24. Berger T, Zwick RH, Stuehlinger M et al (2011) Impact of oxygen uptake efficiency slope as a marker of cardiorespiratory reserve on response to cardiac resynchronization therapy. Clin Res Cardiol 100(2):159–166 25. Padeletti L, Fantappie C, Perrotta L et al (2011) Cardiac memory in humans: vectocardiographic quantification in cardiac resynchronization therapy. Clin Res Cardiol 100(1):51–56 26. Delgado V, van Bommel RJ, Bertini M et al (2011) Relative merits of left ventricular dyssynchrony, left ventricular lead position, and myocardial scar to predict long-term survival of

123

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

ischemic heart failure patients undergoing cardiac resynchronization therapy. Circulation 123:70–78 Delgado V, Ypenburg C, van Bommel RJ et al (2008) Assessment of left ventricular dyssynchrony by speckle tracking strain imaging comparison between longitudinal, circumferential, and radial strain in cardiac resynchronization therapy. J Am Coll Cardiol 51(20):1944–1952 Van de Veire NR, Bleeker GB, De Sutter J et al (2007) Tissue synchronisation imaging accurately measures left ventricular dyssynchrony and predicts response to cardiac resynchronisation therapy. Heart 93:1034–1039 Bleeker GB, Bax JJ, Fung JW et al (2006) Clinical versus echocardiographic parameters to assess response to cardiac resynchronization therapy. Am J Cardiol 97:260–263 Van Bommel RJ, Schalij MJ, Bax JJ (2009) Should the left ventricular pacing lead be positioned at the site of latest mechanical activation in cardiac resynchronization therapy? J Cardiovasc Electrophysiol 20(5):536–538 Aiba T, Barth A, Tomaselli GF (2008) Deciphering gene expression profiling in cardiac resynchronization therapy. J Am Coll Cardiol 52(14):1177 author reply 1177–1178 Klemm HU, Krause KT, Ventura R et al (2010) Slow wall motion rather than electrical conduction delay underlies mechanical dyssynchrony in postinfarction patients with narrow QRS complex. J Cardiovasc Electrophysiol 21(1):70–77 Rose J, Armoundas AA, Tian Y et al (2005) Molecular correlates of altered expression of potassium currents in failing rabbit myocardium. Am J Physiol Heart Circ Physiol 288:H2077– H2087 Akar FG, Rosenbaum DS (2003) Transmural electrophysiological heterogeneities underlying arrhythmogenesis in heart failure. Circ Res 93:638–645 Kaab S, Nuss HB, Chiamvimonvat N et al (1996) Ionic mechanism of action potential prolongation in ventricular myocytes from dogs with pacing-induced heart failure. Circ Res 78: 262–273 Tsuji Y, Zicha S, Qi XY et al (2006) Potassium channel subunit remodeling in rabbits exposed to long-term bradycardia or tachycardia: discrete arrhythmogenic consequences related to differential delayed-rectifier changes. Circulation 113:345–355 Beuckelmann DJ, Nabauer M, Erdmann E (1993) Alterations of K? currents in isolated human ventricular myocytes from patients with terminal heart failure. Circ Res 73:379–385 Aiba T, Hesketh GG, Barth AS et al (2009) Electrophysiological consequences of dyssynchronous heart failure and its restoration by resynchronization therapy. Circulation 119:1220–1230 Ukena C, Bauer A, Mahfoud F et al (2012) Renal sympathetic denervation for treatment of electrical storm: first-in-man experience. Clin Res Cardiol 101(1):63–67 Aiba T, Tomaselli G (2012) Electrical remodeling in dyssynchrony and resynchronization. J Cardiovasc Transl Res 5(2): 170–179 Sachse FB, Torres NS, Savio-Galimberti E et al (2012) Subcellular structures and function of myocytes impaired during heart failure are restored by cardiac resynchronization therapy. Circ Res 110:588–597 Yilmaz A, Gdynia HJ, Ponfick M et al (2012) Cardiovascular magnetic resonance imaging (CMR) reveals characteristic pattern of myocardial damage in patients with mitochondrial myopathy. Clin Res Cardiol 101(4):255–261

Not left ventricular lead position, but the extent of immediate asynchrony reduction predicts long-term response to cardiac resynchronization therapy.

Cardiac resynchronization therapy (CRT) is an effective treatment for a large subgroup of chronic heart failure patients. Various attempts to improve ...
355KB Sizes 0 Downloads 0 Views