OUP user menu

Response to preventive cardiac resynchronization therapy in patients with ischaemic and nonischaemic cardiomyopathy in MADIT-CRT

Alon Barsheshet, Ilan Goldenberg, Arthur J. Moss, Michael Eldar, David T. Huang, Scott McNitt, Helmut U. Klein, W. Jackson Hall, Mary W. Brown, Jeffrey J. Goldberger, Robert E. Goldstein, Claudio Schuger, Wojciech Zareba, James P. Daubert
DOI: http://dx.doi.org/10.1093/eurheartj/ehq407 1622-1630 First published online: 12 November 2010

Abstract

Aims There are no data regarding the differential response to cardiac resynchronization therapy with defibrillator (CRT-D) by the aetiology of cardiomyopathy in mildly symptomatic patients. We evaluated the outcome of patients enrolled in MADIT-CRT by ischaemic and non-ischaemic aetiology of cardiomyopathy (ICM and non-ICM, respectively).

Methods and results The clinical response to CRT-D was assessed among ICM (n = 1046) and non-ICM (n = 774) patients enrolled in MADIT-CRT during an average follow-up of 2.4 years, and echocardiographic response was assessed at 1 year. Cardiac resynchronization therapy with defibrillator vs. ICD therapy was associated with respective 34% (P = 0.001) and 44% (P = 0.002) reductions in the risk of heart failure or death among ICM and non-ICM patients (P for interaction = 0.455). In the ICM group, CRT-D was associated with mean (±SD) 29 ± 14% and 18 ± 10% reductions in left ventricular end-systolic volume (LVESV) and left ventricular end-diastolic volume (LVEDV), respectively. In the non-ICM group, CRT-D was associated with significantly greater volume reductions compared with the ICM group [37 ± 16% and 24 ± 12% reductions in LVESV and LVEDV, respectively (P < 0.001 for all)]. Risk subsets in the ICM group that showed a favourable clinical response to CRT-D included patients with QRS ≥150 ms, systolic blood pressure <115 mmHg, and left bundle branch block (LBBB), whereas in the non-ICM group females, patients with diabetes mellitus, and LBBB, displayed a favourable clinical response.

Conclusion Mildly symptomatic ICM and non-ICM patients show significant differences in the echocardiographic response to CRT-D and in the clinical benefit within risk subsets suggesting that risk assessment for CRT-D in this population should be aetiology-specific.

  • Heart failure
  • Cardiac resynchronization therapy
  • Ischaemic aetiology

See page 1580 for the editorial comment on this article (doi:10.1093/eurheartj/ehr016)

Introduction

The Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronization Therapy (MADIT-CRT) showed a significant reduction in the risk for heart failure (HF) or death during preventive cardiac resynchronization therapy with defibrillator (CRT-D) in mildly symptomatic patients with ischaemic cardiomyopathy (ICM) and non-ICM.1 However, subgroup analyses of the primary results demonstrated differing benefits of the device among important subsets in the heterogeneous population of enrolled patients.1 Prior reports have shown that HF patients with ICM exhibit important differences in their co-morbidities, risk factors, outcomes, and response to medical therapy compared with patients with non-ICM.2,3 Furthermore, it has been suggested that in patients with advanced HF (NYHA class III/IV) implanted with CRT, the magnitude of improvement in left ventricular ejection fraction (LVEF) and LV volumes among non-ICM patients was greater than in patients with ICM.46 Current information is limited,7 however, regarding differences in the outcomes between ICM and non-ICM patients with asymptomatic or mildly symptomatic (NYHA class I/II) HF following preventive CRT-D implantation. Furthermore, it is possible that there are aetiology-specific factors that may help identify differences in the response to preventive CRT-D therapy in ICM and non-ICM patients.

Accordingly, the present study was designed to evaluate the clinical and echocardiographic response to CRT-D therapy among ICM and non-ICM patients enrolled in MADIT-CRT, and to identify aetiology-specific factors that may be useful for risk assessment among candidates for preventive CRT-D therapy in these populations.

Methods

Study population

The design and results of the MADIT-CRT study have been reported previously.1 Briefly, 1820 patients with ICM [NYHA class I or II (n = 1046)] or non-ICM [NYHA class II (n = 774)], LVEF <0.30, and prolonged intraventricular conduction with QRS > 130 ms were randomized to receive either CRT-D or ICD therapy in a 3:2 ratio. The randomization process was stratified by disease aetiology (ICM vs. non-ICM). Screened patients were excluded from enrolment if they had an existing indication for CRT, NYHA class III/IV in the past 90 days before enrolment, implanted pacemaker, coronary artery bypass graft surgery, percutaneous coronary intervention, or myocardial infarction (MI) within the past 90 days before enrolment. Patients were also excluded if they had reversible causes of non-ICM such as acute viral myocarditis or alcohol-induced heart disease.

Definitions and outcome measures

For the diagnosis of ICM, study patients had to have at least one of the following: (i) a documented history of a prior MI (Q-wave or enzyme-positive); (ii) a history of a coronary revascularization procedure [one or more prior coronary artery bypass graft surgeries or percutaneous coronary interventions (balloon and/or stent angioplasty)]; (iii) documented significant coronary artery disease at coronary angiography and history of angina pectoris or other coronary-related symptoms or signs, such that the enrolling physician assessed the patient as having ICM. Non-ICM was defined as the absence of each of the above criteria for ICM, implying an abnormality of the myocardium producing systolic dysfunction and HF not due to coronary disease (or primary valvular disease). Left bundle branch block (LBBB) was defined with QRS duration ≥130 ms, QS or rS in lead V1, and broad (frequently notched or slurred) R waves in lead I, AVL, V5, V6, absent Q waves in lead I, V5, and V6.

The diagnosis of HF required signs and symptoms consistent with congestive HF that was responsive to intravenous decongestive therapy on an outpatient basis or an augmented decongestive regimen with oral or parenteral medications during an in-hospital stay. Two-dimensional echocardiography was performed at baseline and at the 1-year follow-up as previously reported,1 to assess changes in left ventricle end-systolic volume (LVESV), left ventricle end-diastolic volume (LVEDV), left atrial volume (LAV), and LVEF. The change in these echocardiographic variables was evaluated both as a continuous measure and as a dichotomous measure. We used the approximate lower quartile of change in the CRT-D arm as a threshold for the definition of echocardiographic response. Accordingly, echocardiographic response was defined as : >25% reduction in LVESV, >15% reduction in LVEDV, >20% reduction in LAV, and >8% increase in LVEF. Volumes were measured according to Simpson's method, and LVEF was calculated using a standardized protocol.8 Paired echacardiographic data were available for 814 ICM and 614 non-ICM patients. The food and drug administration initially required that CRT pacing be turned off during the 1-year echocardiography, but this requirement was subsequently reversed, and the paired echocardiographic studies from this point were obtained in the CRT-D group with resynchronization on at 1 year.

The outcome measures of the present study included: (i) clinical response to preventive CRT-D therapy, defined either as a reduction in the risk of HF or death (primary analysis), or as a first HF event (secondary analysis) when compared with ICD-only therapy; and (ii) echocardiographic response to preventive CRT-D, defined as greater than lower quartile percent reduction in LVESV, LVEDV, LAV or greater than lower quartile increase in LVEF.

Statistical analysis

Characteristics of patients categorized by ischaemic aetiology were compared by the Wilcoxon rank-sum test or Chi-square test, as appropriate. The probability of HF or death by aetiology group and by treatment group was graphically displayed according to the method of Kaplan and Meier, with comparison of cumulative events by the log-rank test. Multivariate analysis was carried out using Cox proportional hazards regression modelling using clinically selected covariates deemed important risk factors for an HF event or death. Pre-specified covariates in the multivariate model included: age ≥65 years, gender, diabetes mellitus, LVEF <25%, systolic blood pressure (SBP) <115 mmHg, QRS ≥150 ms, and estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m2. Risk factor thresholds for categorization into lower risk and higher risk subsets were pre-specified with the use of the same, accepted criteria for clinical and laboratory values employed in the primary MADIT-CRT analysis. In a further analysis, the presence of LBBB (which was not pre-specified in the primary MADIT-CRT analysis)1 was added as an additional covariate to the multivariate models. Aetiology-specific differences in the clinical benefit of CRT-D therapy were assessed by including a treatment-by-risk factor interaction term in the separate models for the ICM and non-ICM subgroups.

The change in echocardiographic measures and in echocardiographic response was compared between ICM and non-ICM groups with Student's t-test or Chi-square test, as appropriate.

Since ICM patients enrolled in MADIT-CRT comprised both NYHA classes I and II, whereas non-ICM patients comprised only NYHA class II, we carried out a secondary analysis that evaluated the consistency of the results among enrolled patients with NYHA class II. All P-values were two-sided, and a P-value <0.05 was considered significant. Analyses were conducted with SAS software (version 9.2, SAS Institute, Cary, NC, USA).

Results

The 1046 study patients with ICM exhibited significant differences in baseline characteristics when compared with 774 study patients with non-ICM (Table 1). Patients with ICM were older, displayed a higher frequency of co-morbidities, including diabetes mellitus, hypertension, and impaired renal function, and more frequently received statins when compared with non-ICM. In contrast, non-ICM patients displayed a greater proportion of females, had longer baseline QRS duration, higher frequency of LBBB, and received a higher frequency of medical therapies with beta-blockers, angiotensin converting enzyme (ACE) inhibitors or angiotensin receptor blockers, and aldosterone antagonists when compared with ICM patients. Notably, baseline LV volumes were significantly greater among non-ICM patients.

View this table:
Table 1

Patient characteristics

Ischaemic (n= 1046)Non-ischaemic (n= 774)P-value
CRT-D60%59%0.690
Age (years)68 (61–74)62 (53–69)<0.001
Age ≥65 years63%39%<0.001
Female sex13%41%<0.001
Diabetes35%24%<0.001
Hypertension70%55%<0.001
Smoking13%11%0.249
Black race5%11%<0.001
Caucasian93%87%<0.001
NYHA II75%100%<0.001
Systolic blood pressure (mmHg)122 (110–135)120 (110–132)0.154
Systolic blood pressure <115 mmHg34%33%0.901
eGFR (mL/min/1.73 m2)67 ± 1972 ± 21<0.001
eGFR <60 mL/min/1.73 m238%27%<0.001
QRS (ms)152 (140–164)160 (150–176)<0.001
QRS ≥150 ms56%76%<0.001
LBBB58%88%<0.001
LVEF (%)25 (20–29)25 (20–28)<0.001
LVEF <25%34%49%<0.001
LVEDV/BSA (mL/m2)117 (104–132)122 (107–142)<0.001
LVESV/BSA (mL/m2)83 (72–94)88 (74–104)<0.001
ACE-I/ARB94%97%0.013
Beta-blockers92%95%0.033
Diuretics74%75%0.521
Aldosterone antagonist27%38%<0.001
Statins85%44%<0.001
  • Data are presented as percentage or median (interquartile range).

  • ACE-I/ARB, angiotensin converting enzyme inhibitor or angiotensin receptor blocker; BSA, body surface area; eGFR, estimated glomerular filtration rate; LBBB, left bundle branch block; LVEF, left ventricle ejection fraction; LVEDV, left ventricle end-diastolic volume; LVESV, left ventricle end-systolic volume; MI, myocardial infarction; NYHA, New York Heart Association.

Clinical and echocardiographic response to cardiac resynchronization therapy with defibrillator in patients with ischaemic cardiomyopathy

Among the 1046 patients with ICM, CRT-D therapy was associated with a significant reduction in the rate of an HF event or death when compared with ICD-only therapy (Figure 1A). At 3 years of follow-up, the cumulative probability of HF or death was 24% among CRT-D patients when compared with 34% among ICD-only patients (P < 0.001), corresponding to a relative risk-reduction of 29%. Similarly, multivariate analysis demonstrated that ICM patients who were treated with CRT-D therapy derived a significant 34% (P = 0.001) reduction in the risk of HF or death and a significant 42% (P < 0.001) reduction in the risk of a first HF event (Table 2). Among patients with ICM and NYHA class II only, CRT-D was associated with a similar magnitude of benefit [38% (P = 0.001) reduction in the risk of HF or death and a 43% (P < 0.001) reduction in the risk of a first HF event]. Preventive CRT-D therapy was also associated with a significant echocardiographic response in the ICM group. At 1 year of follow-up, CRT-D treated patients showed significant reductions in left ventricle and LAVs, when compared with baseline values, and experienced a significant increase in ejection fraction (Figure 2 and Table 3), with a similar magnitude of echocariographic response in ICM patients with NYHA classes I and II (data not shown).

View this table:
Table 2

Clinical effect of cardiac resynchronization therapy with defibrillator vs. ICD by ischaemic aetiology

All ischaemic (n= 1046)Ischaemic NYHA II (n= 781)Non-ischaemic (n= 774)P for interaction*
Adjusted HR (95% CI)P-valueAdjusted HR (95% CI)P-valueAdjusted HR (95% CI)P-value
HF or death0.66 (0.52–0.85)0.0010.62 (0.47–0.83)0.0010.56 (0.39–0.80)0.0020.455
HF event0.58 (0.45–0.77)<0.0010.57 (0.41–0.77)<0.0010.50 (0.35–0.75)0.0010.562
Death0.99 (0.65–1.52)0.9840.96 (0.59–1.55)0.8540.87 (0.45–1.67)0.6690.728
  • Adjusted HR = CRT-D:ICD hazard ratio adjusted for gender, age ≥65 years, diabetes mellitus, QRS ≥150 ms, EF<25%, eGFR<60 mL/min/1.73 m2, and systolic blood pressure <115 mmHg.

  • *P for treatment-by-aetiology (all ischaemic vs. non-ischaemic) interaction.

View this table:
Table 3

Echocardiographic effect of cardiac resynchronization therapy with defibrillator and ICD-only by aetiology

Echo parametersIschaemicNon-ischaemicCRT-D I vs. N, P-value
CRT-D (n= 154)ICD (n= 194)CRT-D (n= 477)ICD (n= 592)
(A) Echo respondersa
 LVESV (reduction >25%)64%4%80%6%<0.001
 LVEDV (reduction >15%)62%9%76%13%0.001
 LAV (reduction >20%)70%0%77%2%<0.001
 LVEF (increase >8%)68%7%83%7%<0.001
(B) Total change
 ΔLVESV/baseline volume−29 ± 14%−10 ± 9%−37 ± 16%−11 ± 9%<0.001
 ΔLVEDV/baseline volume−18 ± 10%−5 ± 5%−24 ± 12%−7 ± 6%<0.001
 ΔLAV/baseline volume−25 ± 13%−10 ± 8%−29 ± 14%−10 ± 8%<0.001
 ΔLVEF10 ± 5%3 ± 3%12 ± 5%3 ± 3%<0.001
  • Total change data are presented as percent change ± SD.

  • ΔLVEF change = 1 year LVEF (%) − baseline LVEF (%).

  • ΔVolume/baseline volume = (1 year volume − baseline volume)/baseline volume.

  • LAV, left atrial volume, LVEF, left ventricle ejection fraction; LVEDV, left ventricle end-diastolic volume; LVESV, left ventricle end-systolic volume.

  • aEchocardiographic response was defined as a reduction in volumes greater than the lower approximate quartile change in volumes during 1-year follow-up.

  • P-value for the comparison between CRT-D ischaemic and CRT-D non-ischaemic.

Figure 1

Kaplan–Meier estimates of probability of death or HF events in (A) all ischaemic cardiomyopathy patients, (B) ischaemic cardiomyopathy patients with NYHA functional class II, and (C) patients with non-ischaemic cardiomyopathy. Unadjusted P-value for aetiology-by-treatment interaction among all patients = 0.537.

Figure 2

The effects of cardiac resynchronization therapy with defibrillator on echocardiographic measures in ischaemic and non-ischaemic cardiomyopathy patients. *P< 0.001. The bars represent median values. ΔLVEF = 1 year LVEF (%) − baseline LVEF (%). ΔVolume/baseline = (1 year volume − baseline volume)/baseline volume. LVEF, left ventricle ejection fraction; LVEDV, left ventricle end-diastolic volume; LVESV, left ventricle end-systolic volume.

Subgroup analysis in the ICM group showed two statistically significant risk factor-by-treatment interaction effects (Figure 3A). Ischaemic cardiomyopathy patients with QRS ≥150 ms derived a significantly greater benefit from CRT-D therapy [HR (95% CI) = 0.53 (0.38–0.73), P < 0.001] than patients with lower QRS durations [HR = 0.89 (0.60–1.31), P = 0.55 (P-value for QRS duration-by-treatment interaction = 0.04)], and patients with SBP <115 mmHg derived a significantly greater benefit from CRT-D therapy [HR = 0.48 (0.32–0.72), P < 0.001] than those who exhibited higher SBP levels [HR = 0.80 (0.58–1.11), P = 0.18 (P-value for SBP-by-treatment interaction = 0.05)]. The benefit of CRT-D therapy was similar in both dichotomous divisions of the five other pre-specified subgroups (Figure 3A). Consistent with the hazard ratio data, CRT-D was associated with a significant reduction in 3-year event rates relative to ICD-only therapy in ICM patients with QRS ≥ 150 ms and in patients with SBP <115 mmHg (Figure 4A and B, left)

Figure 3

Cardiac resynchronization therapy with defibrillator vs. ICD adjusted hazard ratios for HF/death by risk factors in (A) patients with ischaemic cardiomyopathy and (B) patients with non-ischaemic cardiomyopathy. Risk subsets that showed a statistically significant interaction with therapy are shown first in each aetiology group.

Figure 4

Cardiac resynchronization therapy with defibrillator vs. ICD 3-year event rate by aetiology-specific risk factors for HF or death. (A) High risk patients. (B) Low risk patients. High- and low-risk patients were categorized based on their outcome with ICD-only therapy. Risk factors thresholds for categorization into lower risk and higher risk subsets were pre-specified with the use of the same accepted criteria for clinical and laboratory values employed in the primary MADIT-CRT analysis. In patients with ischaemic cardiomyopathy, interaction P-value for SBP × treatment = 0.164 and interaction P-value for QRS × treatment = 0.020. In patients with non-ischaemic cardiomyopathy, interaction P-value for gender × treatment <0.001 and interaction P-value for diabetes × treatment = 0.019.

A significant difference in the benefit of CRT-D was also shown in a secondary analysis that included LBBB as an additional covariate in the multivariate models. Among ICM patients with LBBB (n= 601), CRT-D was associated with a significant reduction in the risk of HF or death [HR = 0.47 (95% CI 0.34–0.65), P < 0.001], whereas among ICM patients without LBBB (n= 444), no significant benefit was shown [HR = 1.09 (0.72–1.65), P = 0.68 (P-value for LBBB-by-treatment interaction = 0.002)].

Clinical and echocardiographic response to cardiac resynchronization therapy with defibrillator in patients with non-ischaemic cardiomyopathy

Cardiac resynchronization therapy with defibrillator therapy was associated with a pronounced clinical and echocardiographic response among the 774 non-ICM patients enrolled in MADIT-CRT. At 3 years of follow-up, the cumulative probability of an HF event or death was 13% among CRT-D patients when compared with 24% among ICD-only patients (P= 0.002), corresponding to a relative risk reduction of 46%. In multivariate analysis (Table 2), CRT-D was shown to be associated with a pronounced 44% (P= 0.002) reduction in the risk of HF or death and with a 50% (P= 0.001) reduction in the risk of a first HF event. Cardiac resynchronization therapy with defibrillator therapy was also associated with a pronounced echocardiographic response in the non-ICM group. At 1 year of follow-up, CRT-D treated patients showed significant reductions in left ventricle and LAVs when compared with baseline values, and experienced a significant increase in ejection fraction (Figure 2 and Table 3).

Subgroup analysis in the non-ICM group (Figure 3B) showed that the benefit of CRT-D therapy was significantly higher among females [HR (95% CI) = 0.25 (0.14–0.46), P< 0.001] than among males [HR = 0.90 (0.54–1.46), P= 0.67 (P-value for gender-by-treatment interaction = 0.001)], and among patients with diabetes mellitus [HR = 0.31 (0.16–0.61), P< 0.001] than among non-diabetic patients [HR = 0.67 (0.44–1.04), P= 0.08 (P-value for diabetes-by-treatment interaction = 0.05)]. The benefit of CRT-D therapy was similar in both dichotomous divisions of the five other pre-specified subgroups. Consistently, CRT-D vs. ICD-only therapy was associated with a significant reduction in 3-year event rates in non-ICM female and diabetic patients (Figure 4A and B, right).

In a secondary analysis that included the presence of LBBB as an additional covariate in the multivariate models, non-ICM patients with LBBB (n= 680) were shown to derive a significant reduction in the risk of HF or death with CRT-D therapy [HR = 0.42 (95% CI 0.28–0.62)], whereas among the relatively small number of non-ICM patients without LBBB (n= 93), CRT-D therapy was not associated with a significant benefit [HR = 1.57 (95% CI 0.61–4.04), P= 0.35 (P-value for LBBB-by-treatment interaction = 0.011)].

Comparison of outcomes between the two aetiology groups

Among ICD-only patients, Kaplan–Meier event analysis demonstrated significantly higher rates of HF or death at 3 years in the ICM group [34% (Figure 1A)] when compared with the non-ICM group [24%; P= 0.007 (Figure 1C)]. Consistently, multivariate analysis in the ICD-only treatment arm showed that ischaemic aetiology was associated with an adjusted 45% (P= 0.03) increased risk of HF or death when compared with non-ischaemic aetiology. [For patients with NYHA II only, there was 41% increased risk P= 0.007).]

Interaction-term analysis did not show a statistically significant difference in the clinical benefit of CRT-D therapy between ICM- and non-ICM patients (P-value for aetiology-by-treatment interaction = 0.46 for all patients, and P-value for interaction = 0.63 for patients with NYHA class II; Table 2). However, the echocardiographic response to CRT-D was significantly higher in non-ICM patients when compared with ICM patients. As presented in Table 3A, the percent of CRT-D responders according to change in LVESV was 64 vs. 80% (P< 0.001); LVEDV 62 vs. 76% (P= 0.001); LAV 70 vs. 77% (P< 0.001); and LVEF 68 vs. 83% (P< 0.001), among ischaemic vs. non-ischaemic patients with CRT-D, respectively. Consistently, CRT-D was associated with significantly greater reductions in LVESV, LVEDV, and LAV, respectively, among non-ischaemic vs. ischaemic patients with CRT-D (Table 3B and Figure 2).

Similar results were obtained when further adjustment was made for the presence of LBBB (in addition to the pre-specified covariates that included QRS ≥150 ms) or medication usage, including: ACE-inhibitors, angiotensin receptor blockers, beta-blockers, and spironolactone.

Discussion

MADIT-CRT enrolled a population with dilated cardiomyopathy, a prolonged QRS duration and mild symptoms or no symptoms of HF. However, the study included two distinct aetiologies of HF that have a different clinical course and different risk factors. Currently, there are limited data regarding the response to preventive CRT-D therapy in these two distinct aetiologic groups. The present analysis extends the primary findings of the MADIT-CRT trial and demonstrates several new findings that deserve emphasis: (i) preventive CRT-D therapy is associated with significant clinical and echocardiographic benefit among both ICM and non-ICM patients; (ii) the magnitude of the echocardiographic response to CRT-D therapy is significantly higher in the non-ICM group; and (iii) within the two aetiology groups there are risk subsets that exhibit significant differences in the clinical response to CRT-D.

Differences in clinical course and response to cardiac resynchronization therapy with defibrillator between patients with ischaemic and non-ischaemic cardiomyopathy

Epidaemiologic and clinical trial data suggest that patients with ICM differ significantly from patients with non-ICM.3,9 Several studies demonstrated that ICM is associated with poorer prognosis than non-ICM and suggested a difference in response to medical therapy.23 Our findings extend prior observations, and demonstrate, for the first time, that among mildly symptomatic patients with left ventricular dysfunction who have an ICD implanted but do not receive preventive CRT, ischaemic aetiology is associated with a significant 45% increase in the risk of HF or death when compared with non-ischaemic aetiology.

Data from recent clinical trials showed that among patients with HF and NYHA functional classes III–IV, ICM and non-ICM patients gained similar clinical benefit from CRT when compared with medical treatment, but non-ICM patients had greater reverse remodelling compared with non-ICM patients.46 In addition, the REVERSE trial showed that CRT vs. medical treatment in patients with NYHA I/II resulted in a major reverse remodelling at 1 year, with the greatest changes occurring in patients with a nonischaemic cause of HF.7 The REVERSE, however, was underpowered to assess the differential clinical benefit of CRT between the two aetiology groups. The MADIT-CRT trial was the first clinical trial that was designed to explore the clinical effect of CRT-D vs. ICD in asymptomatic or mildly symptomatic HF patients. Herein, we have consistently shown that CRT-D therapy is associated with a significant reduction in the risk of an HF event or death and with significant reverse remodelling effects in both ICM and non-ICM patients with mild symptomatic HF. However, the magnitude of the echocardiographic effects of CRT-D in this population was shown to be significantly higher among non-ICM patients when compared with ICM patients.

Aetiology specific risk assessment for preventive cardiac resynchronization therapy with defibrillator therapy

The above differences in the clinical course, outcome, and response to CRT-D therapy between the aetiology groups emphasize the need to explore aetiology-specific factors that affect response to preventive CRT.

Our study is the first to identify risk subsets who exhibit a differential response to preventive CRT-D therapy among ICM and non-ICM patients. We have shown in the ICM group that the benefit of CRT-D therapy was pronounced among patients with a prolonged QRS duration and among those with low SBP levels and attenuated among those with narrower QRS durations and those with higher SBP levels. In contrast, in the non-ICM group, CRT-D therapy was associated with a pronounced benefit among females and patients with diabetes mellitus, and with attenuated efficacy among males and non-diabetic patients. The two risk factors in each group were unique to that group: each was not a significant source of risk in the other group. Notably, the same factors in each aetiology group that was associated with a pronounced clinical response to CRT-D were also associated with increased risk for HF or death among patients treated with ICD-only therapy in each group (data not shown). These findings suggest that aetiology-specific risk factors for HF or death among patients who do not receive CRT-D therapy identify risk subsets that will derive a pronounced reduction in the risk of HF or death following CRT-D implantation.

In the non-ICM patients, diabetes and female gender may identify those with a microvascular disease risk adding to the risk of the preponderant cardiomyopathy.10,11

Alternatively, it is possible that CRT-D confers greater benefit in patients with the condition known as diabetic cardiomyopathy, which can be caused by mechanisms other than micro- or macro-vascular disease such as altered myocardial substrate, altered energy metabolism, impaired calcium homeostasis, and increased oxidative stress.12

In the ICM patients, wide QRS and low blood pressure may be markers of more advanced deterioration of the myocardium (in addition to intravascular pathology).

Low SBP is known to be associated with poor outcome in patients with chronic HF.13 Systolic blood pressure is related to cardiac output and peripheral resistance; because patients with ICM when compared with non-ICM have greater degree of vascular stiffness,14 it is possible that SBP <115 mmHg represents lower cardiac output in ICM than non-ICM patients (as the greater vascular stiffness should raise SBP) and therefore greater risk for HF event or death.

The benefit of CRT-D was pronounced for both ICM and non-ICM patients with LBBB when compared with patients without LBBB on the ECG. It is conceivable that this enhanced benefit of CRT-D is mediated by coordinating contraction of the interventricular septum and lateral left ventricle, which is typically electrically activated last in patients manifesting LBBB, and is consistent with a recent study showing greater clinical and echocardiographic response in patients with LBBB.15

The differences in the proportion of patients taking ACE-inhibitors, angiotensin receptor blockers, and beta-blockers between ICM and non-ICM subgroups (Table 1) although statistically significant were very small, and as seen in the adjusted models had relatively minor clinical effects on the aetiology-specific outcomes. A possible explanation to the difference in treatment between the ICM and non-ICM subgroups is a greater percentage of patients with contra indication to these medications.

Study limitations

A coronary angiogram was not obligatory before study entry. Although lack of any of the four criteria used for definition of ICM makes the diagnosis unlikely, ICM might not be absolutely excluded. We did not have data on myocardial viability and location of the most delayed myocardial wall. However, analysis of lead positions (799 patients had lead position data) demonstrated neither significant differences in the frequency of lead positions (anterior, lateral, posterior, basal, middle, and apical) nor significant differences in CRT-D clinical benefit for LV lead positions between ICM and non-ICM patients, suggesting that it is unlikely that a different lead position relative to the most delayed segment contributed significantly to the different effects of CRT-D among ischaemic and non-ischaemic patients. Left bundle branch block was not a pre-specified variable in the MADIT-CRT trial, and was therefore included as an additional covariate only in a secondary analysis (that did not change the primary results). Analyses in this study are all secondary to the original clinical trial, and hence all P-values and other formalities should be considered ‘nominal’. However, cutpoints such as 150 ms for QRS were pre-specified in the MADIT-CRT trial; there has been no attempt to search over alternative values. Subgroup analyses are underpowered, particularly when moving onto subgroups of subgroups; in addition, caution in the interpretation of the subgroup interactions is needed because of multiple testing.

Conclusions and clinical implications

In recent years, there has been a significant increase in the usage of CRT for the treatment of symptomatic HF. Furthermore, data from MADIT-CRT now demonstrate that CRT-D therapy can also be employed for the prevention of HF progression among asymptomatic and mildly symptomatic patients with both ischaemic- and non-ischaemic LV dysfunction. Our data demonstrate that despite the significant benefit of preventive CRT-D therapy in both ICM and non-ICM patients, the two aetiology groups exhibited significant differences in the echocardiographic response and clinical response within risk subsets. We suggest that preventive CRT-D should be considered a first-line therapy among mildly symptomatic ICM or non-ICM patients with a markedly prolonged QRS duration and LBBB. In addition, a more focused and aetiology-specific risk assessment among candidates for preventive CRT-D therapy with ICM and non-ICM will further improve patient selection and outcomes in these populations.

Funding

The MADIT-CRT study was supported by a research grant from Boston Scientific, St Paul, Minnesota, to the University of Rochester School of Medicine and Dentistry.

Conflict of interest: A.J.M. reports receiving grant support from Boston Scientific, lecture fees from Boston Scientific, Medtronic, and St Jude Medical; J.J.G., research support from Medtronic, Boston Scientific, and St Jude Medical; W.Z., grant support and lecture fees from Boston Scientific and grant support from Medtronic; J.P.D., lecture fees from Boston Scientific and consulting fees from Medtronic. No other potential conflict of interest relevant to this article was reported.

Acknowledgements

This research was carried out while A.B. was a Mirowski-Moss Career Development Awardee at the University of Rochester Medical Center, Rochester, NY.

References

View Abstract