OUP user menu

Left ventricular function improvement after prophylactic implantable cardioverter-defibrillator implantation in patients with non-ischaemic dilated cardiomyopathy

Wolfram Grimm, Nina Timmesfeld, Elena Efimova
DOI: http://dx.doi.org/10.1093/europace/eut097 1594-1600 First published online: 2 May 2013

Abstract

Aims To assess the incidence and prognostic significance of left ventricular (LV) function improvement in patients with non-ischaemic dilated cardiomyopathy (DCM) and prophylactic implantable cardioverter-defibrillator (ICD).

Methods and results A total of 123 patients with DCM and echocardiographic follow-up assessments within 1 year after prophylactic ICD implant were retrospectively studied at our institution. All patients had New York Heart Association class II or III symptoms in the presence of a LV ejection fraction of 23 ± 6% (range: 9–35%) despite optimized medical therapy for at least 3 months prior to ICD implant. Left ventricular function improvement was defined as an increase of LV ejection fraction of more than 5% to more than 35% combined with a decrease LV end-diastolic diameter of at least 5 mm. Left ventricular function improvement after prophylactic ICD implant was found in 30 of 123 patients (24%). Multivariate logistic regression revealed recent onset DCM with symptoms ≤9 months as the only significant predictor of LV function improvement [odds ratio: 6.89; 95% confidence interval (CI): 2.43–21.99, P = 0.0002]. During 74 months mean follow-up, total mortality was higher in patients without vs. with LV function improvement [hazard ratio (HR): 3.75; 95% CI: 1.14–12.31, P = 0.0034], while the incidence of appropriate ICD therapies was similar in both groups in the early phase after prophylactic ICD implant (HR: 1.15; 95% CI: 0.57–2.33, P = 0.70). The incidence of appropriate ICD therapies decreased to ∼1% per year after LV function improvement had occurred.

Conclusion Recently diagnosed DCM predicts LV function improvement after prophylactic ICD implant. Overall survival was significantly better in patients with vs. without LV function improvement, while appropriate ICD therapy rates were similar in both groups in the early phase after prophylactic ICD implantation before LV function improvement occurred.

  • Implantable cardioverter-defibrillator
  • Left ventricular function improvement

What's new?

  • Recently diagnosed dilated cardiomyopathy with a symptom duration ≤9 months predicts left ventricular (LV) function improvement after prophylactic implantable cardioverter-defibrillator (ICD) implant.

  • Overall survival is significantly better in patients with vs. without LV function improvement, while appropriate ICD therapy rates are similar in both groups in the early phase after prophylactic ICD implant. The incidence of appropriate ICD therapies decreases to ∼1% per year after LV function improvement has occurred.

Introduction

Implantable cardioverter-defibrillators (ICDs) have become therapy of first choice to prevent sudden death in patients with dilated cardiomyopathy (DCM), who meet the inclusion criteria of the sudden cardiac death–Heart Failure Trial (SCD–HeFT).1 These criteria include New York Heart Association (NYHA) class II or III heart failure in the presence of a left ventricular (LV) ejection fraction ≤35% despite optimized medical therapy for at least 3 months.2 The optimal timing of prophylactic ICD implant following the initial diagnosis of DCM, however, remains controversial, since a waiting period of 3 months on optimized medical therapy has been found to be too short in some patients to develop a marked increase of LV ejection fraction to a degree which obviates the need for prophylactic ICD implant.318 To date, only limited data are available regarding incidence and predictors of LV function improvement after primary prevention ICD implant in DCM patients who received optimal medical therapy for at least 3 months as recommended by current European Society of Cardiology (ESC) guidelines,2,4 and the prognostic significance of LV function improvement after prophylactic ICD implant in DCM is unknown. Therefore, we sought to determine the incidence and potential predictors of LV function improvement in DCM patients after prophylactic ICD implant, and to determine the prognostic significance of LV function improvement documented by echocardiographic follow-up assessment within the first year after prophylactic ICD-implant in these patients.

Methods

Patients

We retrospectively identified 123 patients with DCM from the monocentre Marburg ICD-Registry, who met the following inclusion criteria:

  1. non-ischaemic DCM with NYHA class II or III symptoms in the presence of a LV ejection fraction ≤35% despite optimized medical therapy for at least 3 months,

  2. prophylactic ICD implant at the Hospital of the University of Marburg, and

  3. repeated echocardiographic studies to assess LV end-diastolic diameter and LV ejection fraction at the time of prophylactic ICD implant and 3–12 months thereafter.

Patients with reversible causes of cardiomyopathy, such as myocarditis, tachycardiomyopathy, peripartum cardiomyopathy, excessive alcohol or drug abuse, or other conditions known to cause cardiomyopathy, were excluded from this study. All patients had undergone coronary angiography prior to ICD implant. Patients with >50% stenosis in any major coronary vessel and patients with a history of myocardial infarction were excluded from this study.

Echocardiography

Two-dimensional echocardiographic examinations were performed in all patients using a Vingmed Vivid Seven™ machine (General Electronics Medical Systems) to determine LV ejection fraction and LV size. Left ventricular ejection fraction was measured in the apical four-chamber view and orthogonal two-chamber view using the modified Simpson method. Left ventricular function improvement was defined as an increase of LV ejection fraction of more than 5% to more than 35% combined with a decrease LV end-diastolic diameter of at least 5 mm.36

Implanted defibrillators and device programming

Exclusively ICDs with nonthoracotomy lead systems, biphasic shock waveforms, and stored intracardiac electrograms in addition to beat-to-beat intervals for episodes triggering device therapy were used in this study. To minimize the number of inappropriate shocks for supraventricular tachyarrhythmias, we routinely programmed a high rate cut-off for ventricular tachycardia (VT) therapy of 327 ± 13 ms (median: 330 ms) in addition to routine use of supraventricular tachycardia discrimination algorithms and use of β-blocker therapy in the majority of patients (Table 1). One or more attempts of antitachycardia pacing were programmed in the majority of patients for VT with rates up to 250 b.p.m. to minimize the number of ICD shocks.7

View this table:
Table 1

Characteristics of 123 patients stratified for recently and remotely diagnosed DCM

All patientsDiagnosis 3–9 monthsDiagnosis >9 monthsP value
Patients, n1235073
 Age, years (range)52 ± 1448 ± 1356 ± 130.001
 Men, n (%)95 (77)41 (82)54 (74)0.38
 Atrial fibrillation, n (%)19 (15)9 (18)10 (14)0.61
 Left bundle branch block, n (%)45 (37)16 (32)29 (40)0.45
 Symptoms duration, months (range)38 ± 45 (3–224)4 ± 2 (3–9)61 ± 46 (10–224)<0.0001
NYHA heart failure class, n (%)
 Class II50 (41)15 (30)35 (48)0.06
 Class III73 (59)35 (70)38 (52)
Medication at study entry, n (%)
 Digitalis82 (67)31 (62)51 (70)0.44
 Diuretics111 (90)46 (92)65 (89)0.76
 ACE inhibitors or ARBs114 (93)49 (98)65 (89)0.08
 Aldosterone antagonists79 (64)35 (70)44 (60)0.34
 β-Blockers92 (75)38 (76)54 (74)0.84
 Sotalol1 (1)0 (0)1 (1)1.00
 Amiodarone17 (14)5 (10)12 (16)0.43
 Class I anitarrhythmics0 (0)0 (0)0 (0)
Implanted ICD and programming
 Single-chamber ICD77 (63)33 (66)44 (60)0.56
 Dual-chamber ICD8 (7)4 (8)4 (5)
 CRT-ICD38 (31)13 (26)25 (34)
 Lower VT detection interval (ms)327 ± 13325 ± 11328 ± 140.15
Echocardiography at implant
 LV ejection fraction (%)23 ± 623 ± 624 ± 60.64
 LV end-diastolic diameter (mm)69 ± 870 ± 869 ± 80.77
Echocardiography during follow-up
 Time from ICD implant to follow-up echocardiography (months)7.9 ± 3.27.4 ± 3.08.2 ± 3.40.18
 LV ejection fraction (%)32 ± 1134 ± 1230 ± 100.08
 LV end-diastolic diameter (mm)66 ± 1064 ± 967 ± 100.12
  • ACE, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; CRT, cardiac resynchronization therapy; ICD, implantable cardioverter-defibrillator; LV, left ventricular; NYHA, New York Heart Association.

Follow-up

All data of baseline clinical characteristics including the results of non-invasive and invasive cardiac evaluation were collected prospectively in the Marburg Defibrillator Database. Patients were followed in 6 months intervals or as soon as possible after spontaneous ICD shocks for device interrogation or retrieval of stored electrograms. Nine patients (7%) who underwent successful heart transplant were censored at the time of heart transplant. Endpoints were appropriate ICD therapies for ventricular tachyarrhythmias as documented by stored electrograms by the device, all-cause mortality, and heart-transplant-free survival.

Statistical analysis

Baseline clinical characteristics stratified for patients with recently and remotely diagnosed DCM with a symptom duration cut point at 9 months3,4,810 were compared using the Mann–Whitney test for continuous variables and Fisher's exact test for nominal variables (Table 1). Univariate logistic regression and multivariate stepwise logistic regression analysis with likelihood-ratio test and backward selection with a cut-off level of 0.2 was used to predict improvement of LV function after prophylactic ICD implant. Odds ratios (ORs) and corresponding 95% confidence intervals (CIs) were reported for multivariate analysis in Table 2. Event-free survival probabilities were estimated with the Kaplan–Meier method stratified for patients with and without LV function improvement. Comparison between patients with and without LV function improvement was performed using the log-rank test. Hazard ratios (HRs) and corresponding 95% CIs were obtained from multivariate Cox regression analysis in Table 3. Results are expressed as mean ± standard deviation unless specified otherwise. All probability values reported are two-sided, and a probability value of P < 0.05 was considered to indicate statistical significance. R-software version 2.15.0 (www.R-project.org) and package survival was used for all statistical analyses.

View this table:
Table 2

Predictors of LV function improvement following prophylactic ICD implantation

All patientsLV function improvementNo LV function improvementP univariateP multivariateHR (95% CI)
Patients, n1233093
 Age (years)52 ± 1453 ± 1452 ± 140.710.101.03 (0.99–1.07)
 Men, n (%)95 (77)23 (77)72 (77)0.93
 Atrial fibrillation, n (%)19 (15)4 (13)15 (16)0.710.130.37 (0.08–1.33)
 Left bundle branch block, n (%)45 (37)12 (40)33 (35)0.66
 Diagnosis 3–9 months, n (%)50 (41)19 (63)31 (33)0.00380.00026.87 (2.43–21.99)
NYHA heart failure class, n (%)
 Class II50 (41)9 (30)41 (44)0.17
 Class III73 (59)21 (70)52 (56)
Medication at study entry, n (%)
 Digitalis82 (67)17 (57)65 (70)0.19
 Diuretics111 (90)28 (93)83 (89)0.50
 ACE inhibitors or ARBs114 (93)28 (93)86 (92)0.87
 Aldosterone antagonists79 (64)24 (80)55 (59)0.032
 β-Blockers92 (75)26 (87)66 (71)0.0710.0543.21 (0.98–13.31)
 Sotalol1 (1)0 (0)1 (1)0.45
 Amiodarone17 (14)2 (7)15 (16)0.16
 Class I anitarrhythmics0 (0)0 (0)0 (0)
ICD and ICD programming
 Single-chamber ICD77 (63)17 (57)60 (65)0.38
 Dual-chamber ICD8 (7)1 (3)7 (8)
 CRT-ICD38 (31)12 (40)26 (28)
 Lower VT detection interval (ms)327 ± 13329 ± 13326 ± 120.180.0791.04 (1.00–1.08)
Echocardiography at implant
 LV ejection fraction (%)23 ± 623 ± 623 ± 60.94
 LV end-diastolic diameter (mm)69 ± 869 ± 970 ± 80.83
Follow-up echocardiography
 Time from ICD implant to follow-up Echocardiography (months)7.9 ± 3.28.2 ± 3.27.8 ± 3.30.610.151.11 (0.96–1.30)
 LV ejection fraction (%)32 ± 1147 ± 727 ± 7
 LV end-diastolic diameter (mm)66 ± 1058 ± 869 ± 9
  • ACE, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; CI, confidence interval; CRT, cardiac resynchronization therapy; HR, hazard ration, ICD, implantable cardioverter-defibrillator; LV, left ventricular; NYHA, New York Heart Association.

View this table:
Table 3

Follow-up results stratified for presence or absence of LV function improvement

All patientsNo LV function improvementLV function improvementP valueHR (95% CI)a
Patients, n1239330
Follow-up duration (months)74 ± 4672 ± 4881 ± 41
Appropriate ICD therapy, n (%)44 (36)34 (37)10 (33)0.701.15 (0.57–2.33)
Inappropriate ICD therapy, n (%)17 (14)12 (13)5 (17)0.770.85 (0.30–2.43)
Total mortality, n (%)33 (27)30 (32)3 (10)0.0193.75 (1.14–12.31)
Heart transplant, n (%)9 (7)9 (10)0 (0)0.066
  • CI, confidence interval; HR, hazard ratio; ICD, implantable cardioverter-defibrillator.

  • aFor patients without LV function improvement vs. patients with LV function improvement.

Results

Characteristics of 123 study patients

The clinical characteristics of 123 study patients stratified for recently and remotely diagnosed DCM are presented in Table 1. There were no significant differences between patients with recently and remotely diagnosed DCM with regard to gender, presence of atrial fibrillation and left bundle branch block, heart failure medication, implanted devices, and ICD programming, as well as baseline LV end-diastolic diameter and baseline LV ejection fraction. Patients with recently diagnosed DCM, however, were significantly younger than patients with remotely diagnosed DCM.

Predictors of left ventricular function improvement

Left ventricular function improvement after prophylactic ICD implant was found in 30 of 123 patients (24%) as summarized in Table 2. Multivariate logistic regression analysis revealed recently diagnosed DCM with a symptom duration of ≤9 months as the only significant predictor for LV function improvement (OR: 6.89; 95% CI: 2.43–21.99, P = 0.0002).

Appropriate implantable cardioverter-defibrillator therapy and all-cause mortality during follow-up

Follow-up results stratified for patients with and without LV function improvement at repeated echocardiographic assessment after ICD implant are summarized in Table 3. During 74 months mean follow-up, 44 of 123 patients (36%) received appropriate ICD therapies without significant differences between patients with and without LV function improvement in the early phase after prophylactic ICD implantation (Figure 1A). Kaplan–Meier curves for arrhythmia-free survival stratified for patients with and without LV function improvement after a blanking period of 12 months following initial ICD implant and after exclusion of 18 patients who received appropriate ICD therapies within the first 12 months of ICD implant are shown in Figure 2. Patients with DCM in whom LV function improvement had occurred within 12 months after initial ICD implantation and who did not receive appropriate ICD therapies within the first 12 months following ICD implantation had a low incidence of arrhythmic events of ∼1% per year during subsequent follow-up, whereas the arrhythmia-free survival curve continued to decrease in patients without LV function improvement (Figure 2). Nine of 123 patients (7%) underwent heart transplant. All nine patients with heart transplant during follow-up belonged to the group of patients without LV function improvement. Thirty-three of 123 patients (27%) died during follow-up. All-cause mortality was higher in patients without vs. with LV function improvement (HR: 3.75; 95% CI: 1.14–12.31, P = 0.019). The combined endpoint of death or heart transplant was also higher in patients without vs. with LV function improvement (HR: 4.89; 95% CI: 1.51–15.84, P = 0.0034).

Figure 1

(A) Kaplan–Meier analysis of arrhythmia-free survival with regard to appropriate ICD therapy for VT or ventricular fibrillation stratified for patients with LV function improvement (Group A) and without LV function improvement (Group B). (B) Kaplan–Meier analysis of heart-transplant-free survival for patients with LV function improvement (Group A) and patients with LV function improvement (Group B). Patients without LV function improvement had a significantly increased risk for subsequent death or heart transplant compared with patients with LV function improvement (HR 4.89; 95% CI: 1.51–15.84; P = 0.0034).

Figure 2

Kaplan–Meier analysis of arrhythmia-free survival with regard to appropriate ICD therapy for VT or ventricular fibrillation stratified for patients with LV function improvement (Group A) and without LV function improvement (Group B) after exclusion of 18 patients who had received appropriate ICD therapies during a blanking period of 12 months (dashed line) indicating a decreased arrhythmia risk in patients of Group A after LV improvement had occurred compared with patients of Group B.

Discussion

This study found a marked improvement of LV function with an increase of LV ejection to more than 35% in approximately one-quarter of DCM patients during follow-up echocardiographic assessment within the first year after prophylactic ICD implant. Recent onset DCM with a symptom duration ≤9 months was the only independent predictor of LV function improvement on multivariate regression analysis. Overall survival was significantly better in patients with vs. without LV function improvement, while appropriate ICD therapy rates were similar in both groups in the early phase after prophylactic ICD implantation before LV function improvement occurred. The incidence of appropriate ICD therapies decreased to ∼1% per year in DCM patients in whom LV function improvement had occurred.

Incidence of left ventricular function improvement in patients with implantable cardioverter-defibrillator

Previous investigators found reversed remodelling with an significant increase in LV ejection fraction within 12 months after initial diagnosis of DCM in a highly variable percentage517 ranging from 19% in an early study of Cicoira et al.11 to 67% in a recent retrospective report by Zecchin et al.3 The majority of these studies were conducted before prophylactic ICD implant became a class I indication for patients with DCM who met SCD–HeFT criteria, i.e. NYHA class II or III heart failure in the presence of a LV ejection fraction ≤35% despite optimized medical therapy for at least 3 months.1,2 The lower incidence of LV improvement in our study (24%) compared with the majority of previous reports3,5,1417 may, in part, be explained by the medical treatment phase of at least 3 months before ICD implant in our study including β-blockers, ACE inhibitors, and aldosterone antagonists in most patients. Patients who experienced a marked improvement of LV function on optimized medical therapy within 3 months after initial diagnosis of DCM had been excluded from our study according to the study design. Nevertheless, our study confirms the finding of several previous reports4,6,11,18 that recent onset cardiomyopathy with a short symptom duration is the most important predictor of LV function improvement during the first year of follow-up, whereas LV ejection fraction does not appear to predict subsequent LV function improvement.5

Predictors of left ventricular function improvement in dilated cardiomyopathy patients with prophylactic implantable cardioverter-defibrillator

Only one previous retrospective multicentre study by Verma et al.4 investigated the incidence and predictors of LV function improvement after prophylactic ICD implant in DCM patients. Repeated echocardiographic assessment of LV function within 1 year after ICD implant revealed a markedly improved LV ejection fraction >35% in 12% of patients in the study of Verma et al.4 compared with 24% of patients in our present study. The lower incidence of patients with LV function improvement in the study by Verma et al.4 compared with our study may be explained, in part, by the fact that patients with a symptom duration <9 months were excluded from study participation in the study by Verma et al.,4 whereas 41% of all patients in our study had a symptom duration of 3–9 months. Similar to our study, multivariate regression analysis identified a shorter symptom duration as the only independent predictor of LV function improvement after ICD implant in the study by Verma et al.4 In contrast to our study, however, Verma et al.4 did not provide follow-up data to determine the prognostic significance of LV function improvement after ICD implant with regard to mortality or transplant-free survival.

Prognostic significance of left ventricular function improvement after prophylactic implantable cardioverter-defibrillator implant

To the best of our knowledge, this is the first study to demonstrate that LV function improvement detected by repeated echocardiographic assessment after prophylactic ICD implant in DCM was associated with an excellent heart-transplant-free survival during follow-up. Despite the observed LV function improvement, we observed a considerable risk for ventricular tachyarrhythmias triggering ICD therapy, which occurred in a similar percentage of patients with and without LV function improvement in the early phase after ICD implantation. The subgroup of DCM patients in our study in whom LV function improvement had occurred within 12 months after initial ICD implantation and who did not receive appropriate ICD therapies within these first 12 months had a low incidence of appropriate ICD therapies of ∼1% per year during subsequent follow-up. This low incidence of arrhythmic events supports the hypothesis that discontinuation of ICD therapy might be safe in selected DCM patients with LV function improvement with an increase in LV ejection fraction >35%. Due to the retrospective nature of our study and due to the limited number of study patients, however, this hypothesis needs to be confirmed by well designed prospective studies or multicentre ICD registries incorporating serial echocardiographic follow-up assessments to determine the occurrence of LV function improvement to a degree which makes prophylactic ICD therapy unnecessary. In addition, future studies or registries need to confirm the stability of LV function improvement in DCM patients during long-term follow-up.

Another important question is the optimal timing of ICD implant for primary prevention of SCD in DCM patients. Current ESC guidelines2 for the treatment of chronic heart failure do not differentiate between recent onset and remote onset heart failure with regard to the recommendation of a prophylactic ICD. These ESC guidelines2 recommend primary prevention ICD implantation in DCM patients with NYHA class II or III heart failure in the presence of a LV ejection fraction ≤35% despite 3 months of optimal medical therapy irrespective of symptom duration. Consistent with the results of previous studies,36,11,13,19 our study suggests that this 3 months waiting period on optimal medical therapy prior to primary prevention ICD implantation may be too short in a significant number of patients with recent onset DCM because reverse remodelling of a time-dependent nature may lead to marked LV function improvement averting the need for an ICD more than 3 months after initiation of optimal medical therapy. To date, it is unknown, in whom prophylactic ICD implantation can safely be delayed for more than 3 months after initiation of medical heart failure therapy, if LV ejection fraction remains still below 35% after 3 months. Future studies are needed to identify potential predictors of LV improvement in patients with recent onset DCM including the amount of myocardial fibrosis detected by cardiac magnetic resonance imaging, the degree of LV dysfunction at initial presentation, the response to optimal medical therapy within the first 3 months, the presence and degree of right ventricular dysfunction, the response to provocative manoeuvres such as dobutamine stress testing, biomarkers including increased B-type natriuretic peptide, or elevated troponin levels, etc.1925

Three previous studies810 investigated whether recent onset DCM was associated with appropriate ICD use during follow-up without using repeated echocardiography to determine the relation between symptom duration and LV function improvement alter ICD implant. All three studies810 found a similar incidence of appropriate ICD therapies for ventricular tachyarrhythmias in patients with recently vs. remotely diagnosed DCM as a surrogate parameter for a similar ICD benefit in both groups. In contrast to our present study, however, none of the previous studies810 evaluated the incidence of appropriate ICD therapies after a blanking period in which LV function improvement had occurred. Furthermore, counting appropriate ICD therapies for ventricular tachyarrhythmias is well known to be a poor surrogate parameter for sudden death prevention and clearly overestimates the benefit of ICD therapy.26 While the incidence of appropriate ICD therapies is critically dependent on ICD programming, more restrictive ICD programming strategies with high rate cut-offs for VT detection and long detection times have recently been demonstrated to significantly reduce appropriate and inappropriate ICD therapies as well as total mortality27

Study limitations

This was a retrospective study with all inherent limitations of retrospective analyses. Our study has also a relatively small number of patients with DCM and prophylactic ICD implant, which limits our ability to detect statistically significant differences in both univariate and multivariate analyses. In addition, this study used repeated echocardiographic assessments at variable time intervals ranging from 3–12 months after initial ICD implantation to determine LV end-diastolic diameter and LV ejection fraction using the modified Simpson method, both of which are subject to inter- and intraobserver variability. Finally, ICD programming in our study did not meet the restrictive criteria used in the multicenter automatic defibrillator implantation trial-reduce inappropriate therapy trial,27 which have recently been shown to be superior compared with conventional ICD programming strategies as used in our study.

Conflict of interest: none declared.

References

View Abstract