Journal of Cardiac Failure Vol. 17 No.

10 2011

Mortality Reduction of Cardiac Resynchronization and

Implantable Cardioverter-Defibrillator Therapy in Heart
Failure: An Updated Meta-Analysis. Does Recent Evidence
Change the Standard of Care?
EDUARDO GEHLING BERTOLDI, MD, MSc,1,2 CARISI ANNE POLANCZYK, MD, ScD,1,2,3 VIVIAN CUNHA, MS,1
PATRÍCIA KLARMANN ZIEGELMANN, ScD,2 LUIS BECK-DA-SILVA, MD, ScD,3
AND LUIS EDUARDO ROHDE, MD, ScD1,3
Porto Alegre, Brazil

ABSTRACT

Background: The recent publication of the MADIT-CRT and RAFT trials has more than doubled the
number of patients in which a direct comparison of the combination of cardiac resynchronization therapy
(CRT) and implantable cardioverter-defibrillator (ICD) versus ICD alone was carried out. The present
meta-analysis aims to assess the impact of combined CRT and ICD therapy on survival of heart failure
(HF) patients.
Methods and Results: Medline, Embase, and the Cochrane Library databases were searched, and all ran-
domized controlled trials of CRT alone or combined with ICDs in HF resulting from left ventricular sys-
tolic dysfunction were included. Main outcome was all-cause mortality. Summary relative risk (RR) and
95% confidence interval (CI) were calculated employing random-effects models. Twelve studies were in-
cluded, with a total of 8,284 randomized patients. For the comparison of CRT alone versus medical ther-
apy, pooled analysis of 5 available trials demonstrated a significant reduction in all-cause mortality with
CRT (RR 0.76, 95% CI: 0.64e0.9). Pooled analysis of 6 trials that compared the combination of CRT and
ICD therapy to ICD alone also showed a statistically significant reduction in all-cause mortality (RR 0.83,
95% CI: 0.72e0.96). Stratified analysis showed significant mortality reductions in all New York Heart As-
sociation class subgroups, with greater effect in classes IIIeIV (RR 0.70; 95% CI: 0.57e0.88). Pooled
estimates of implant-related risks were 0.6% for death and 8% for implant failure.
Conclusion: Combined CRT and ICD therapy reduces overall mortality in HF patients when compared
with ICD alone. (J Cardiac Fail 2011;17:860e866)
Key Words: Heart failure, cardiac resynchronization therapy, implantable cardioverter-defibrillator, meta-
analysis.

Heart failure (HF) is a leading cause of cardiovascular
morbidity and mortality worldwide.1 Because of its in-
creased prevalence in the elderly and the progressive aging
of the population in many countries, the incidence of HF is
increasing,1,2 despite modern diagnostic and therapeutic
options. In addition to pharmacologic therapies, device-
based treatmentsdnamely the use of implantable
cardioverter-defibrillators (ICDs) and biventricular cardiac
pacing devices capable of delivering cardiac resynchroniza-
tion therapy (CRT)dare now considered an important part of

From the 1Postgraduate Program in Cardiology and Cardiovascular Funding from the National Council for Scientific and Technological De-
Sciences, Federal University of Rio Grande do Sul, Porto Alegre, Brazil; velopment (CNPq), Brazil. The funder did not contribute to the study de-
2
National Institute for Health Technology Assessment (IATS), National sign, and had no role in the conduct of the study, collection, management,
Council for Scientific and Technological Development (CNPq), Brazil analysis, or interpretation of the data, nor in preparation, review, or ap-
and 3Cardiovascular Division of Hospital de Clınicas de Porto Alegre, proval of the manuscript.
Brazil. See page 865 for disclosure information.
Manuscript received April 4, 2011; revised manuscript received May 24, 1071-9164/$ - see front matter
2011; revised manuscript accepted June 8, 2011. Ó 2011 Elsevier Inc. All rights reserved.
Reprint requests: Luis Eduardo Rohde, MD, ScD, Serviço de Cardiolo- doi:10.1016/j.cardfail.2011.06.372
gia, Hospital de Clınicas de Porto Alegre. Rua Ramiro Barcelos 2350, Sala
2061, Porto Alegre, RS, Brazil 90035-003. E-mail: lerohde@terra.com.br

860
 CRT and ICD: An Updated Meta-Analysis
HF treatment in selected patients.3 ICDs reduce sudden car-
diac death4 and CRT has a beneficial impact on HF symp-
toms, hospitalization, and survival,5 when added to optimal
medical therapy. However, whether the combination of
both treatments would bring additional benefit over either
treatment alone is still an object of debate. Earlier trials in-
cluded patients with advanced HF (New York Heart Associ-
ation [NYHA] class IIIeIV), whereas later trials
predominantly included patients with mild to moderate HF
(NYHA class IeII). When previous meta-analyses were per-
formed, few trials had directly compared the combination of
CRT and ICD to either therapy alone. Previous investigators
had to use strategies such as exploratory meta-regression
analyses6,7 or Bayesian network meta-analysis8 to combine
the results of the available trials, most of which compared
the use of either ICD or CRT to medical therapy.
Since then, the publication of the MADIT-CRT trial9
and, more recently, the RAFT trial,10 comparing the combi-
nation of CRT and ICD to ICD alone, has more than dou-
bled the number of patients subjected to this comparison.
This allows an updated combined analysis, adding greater
reliability to the available information. We have performed
a systematic review, including the latest trials to evaluate
the effect of the combination of CRT and ICD therapy on
HF mortality, compared with optimal medical therapy alone
and with ICD therapy alone; to explore the impact of
NYHA class on the benefit of CRT therapy (alone or in
combination); and to assess the pooled estimate of major
implant-related complications with CRT-capable devices
by single-branch meta-analysis.
Methods
Search Strategy
We searched Medline, Embase, and the Cochrane Library data-
bases, using a highly sensitive strategy.11 We used keywords and
MeSH terms such as ‘‘resynchroni?ation therapy’’, ‘‘cardiac pace-
maker, artificial’’, ‘‘Defibrillators, Implantable’’, ‘‘((biventricular
or dual-chamber or single-chamber) adj1 (pacing or pacer or stim-
ulat$))’’, ‘‘heart failure, congestive’’, and several variations. The
detailed search strategy can be found on Supplementary Table A
(online only).
Additionally, to address the issue of studies not included in
these databases, we reviewed the reference lists of the identified
studies and previous meta-analyses, and consulted with experts
on the subject.
Study Selection
To be considered for inclusion, studies had to be randomized
controlled trials, either parallel or crossover, with more than 2
weeks of duration, that included patients with HF from left ven-
tricular systolic dysfunction, and evaluated CRT, either alone or
in combination to ICD therapy, versus medical therapy or ICD
therapy alone. The control group should not receive any sort of
resynchronization therapy.
Two investigators independently screened titles and abstracts to
identify eligible studies, according to the inclusion criteria. Full-
text versions of preselected studies were retrieved for detailed
 Bertoldi et al 861
analysis. Disagreements in the selection processes were resolved
by adjudication by a third investigator and final consensus.
Study Quality Assessment
Assessment of methodological quality and potential to bias was
performed by gathering information regarding the concealment of
treatment allocation, blinding, and intention to treat analysis on
each included study.
Outcomes
Primary outcome was all-cause mortality. For crossover trials,
only results from the first period were considered. Additional out-
comes assessed were failure to implant resynchronization devices,
and major implant-related complications. The authors’ definitions
for major implant-related complications were implicitly accepted.
Statistical Analysis
Because between-study differences to the extent of compromis-
ing a fixed-effect analysis were expected, independently of statis-
tical heterogeneity tests, we used random-effects models for all
comparisons as our main analysis. Fixed-effect models were
used for sensitivity analysis in the mortality estimates. For mortal-
ity and hospitalizations, pairwise comparisons were performed
separately for the combination of CRT and ICD versus ICD alone,
and for CRT alone versus medical therapy (MT). Additionally, to
evaluate the impact of NYHA class on the results of CRT, we per-
formed comparisons for CRT versus no CRT in studies that re-
stricted inclusion to patients in NYHA Class III or IV, and in
studies that restricted inclusion to patients in NYHA Class I or II.
Odds ratios from each study were combined, using the inverse
variance method, to provide a pooled risk ratio, using the Review
Manager Software package, version 5.0. For implant failure and
major implant complications, we performed single-branch meta-
analysis with random-effects model, pooling the incidence of
each outcome for patients that received CRT-capable devices
(with or without ICD capability), using the STATA software pack-
age, version 11. Tests for heterogeneity were performed for all
analyses, as well as quantification by the I2 statistic. A P value
#.05 was considered statistically significant both for the overall
effect and for the presence of heterogeneity. Funnel plots were
used to investigate the possibility of publication bias.
Results
Included studies
Results of the study search and selection are summarized
in Figure 1. Twelve randomized controlled trials met the in-
clusion criteria,9,10,12e21 4 of which had a crossover phase
(MUSTIC-SR,12 MUSTIC-AF,14 CONTAK CD,15 and
HOBIPACE20). Five trials evaluated combined CRT and
ICD therapy versus ICD only (CONTAK CD,15 MIRACLE
ICD,16 MIRACLE ICD II,17 MADIT-CRT,9 and RAFT10),
4 trials compared CRT to medical therapy alone
(MUSTIC-SR,12 MUSTIC-AF,14 MIRACLE,13 and CARE-
HF19), 1 trial had a triple comparison of CRTþICD versus
CRT versus medical therapy (COMPANION18), and 1 trial
compared CRT to medical therapy, but 85% of patients in
both groups also had ICD therapy (REVERSE21). All stud-
ies enrolled only patients with prolonged QRS interval and
862 Journal of Cardiac Failure Vol. 17 No. 10 October 2011

ventricular ejection fraction; MT, medical therapy; N, no; NA, not available; NYHA, New York Heart Association; RBBB, right bundle branch block; RCT, randomized controlled trial; RPT, randomized parallel
CA, concealed allocation; CRT, cardiac resynchronization therapy; ICD, implantable cardioverter-defibrillator; ITT, intention to treat analysis; MT, medical therapy; LBBB, left bundle branch; LVEF, left
UC/UC/UC
Blinding*

UC/N/Y
N/Y/UC
N/Y/UC

Y/Y/UC
Y/Y/UC

N/Y/UC
Y/Y/UC

Y/Y/UC
Y/Y/Y

Y/Y/Y

N/N/Y
CA/ITT

UC/UC
UC/UC
UC/Y
UC/Y

UC/Y
UC/Y

UC/Y

UC/Y
Y/Y

Y/Y

Y/Y
Y/Y
LBBB/

NA/NA
NA/NA
RBBB

NA/13
NA/17
87/NA

63/NA
(%)

70/11

70/12
44/26

72/9
NA

NA
Mean QRS

65% O 150
Duration
(ms)

N/A
176
209
166

164
165

160

160
174
154

157
Fig. 1. Flow diagram of study selection.

LVEF

21.7

22.6
(%)
23
26

22

25
26
26
24
22
24
24
HF patients were predominantly of ischemic etiology.
Mean age of participants ranged from 62 to 70 years. All
NYHA
Table 1. Characteristics of Included Studies

Class

III-IV

III-IV

III-IV
II-IV
II-IV

II-III
I-IV
I-II
I-II
III
III
but 2 of the studies clearly declared the use of intention-

II
to treat analysis. However, concealment of treatment alloca-
tion was unclear in most studies (Table 1).
Ischemic

Regarding functional class, 1 trial included only NYHA

N/A
(%)
43

54

70
57

55

38
57
55
55
67
69
Class II patients17; 2 trials included only NYHA Class III
patients12,14; 2 trials included NYHA Class II, III, and
IV patients15,18; 3 trials included NYHA Class III and IV
Mean
Age

patients13,16,19; 2 trials included NYHA Class I and II

(y)
63
66
64

67
63

67

67
70
62
65
66
66

patients9,21; 1 trial included patients in NYHA Class II

and III10; and 1 trial included patients in any NYHA class
(Table 1).20
Male

(%)
Sex

86
81
68

68

73
77
78
84
78
89

75
83
All-Cause MortalitydCRT versus MT
RPT/24 months

RPT/29 months

RPT/12 months
RPT/28 months
RPT/40 months
RCT/6 months
RCT/6 months

RCT/9 months
RCT/6 months
RPT/6 months

RPT/6 months
RPT/6 months
Follow-up

For the comparison of CRT alone versus medical therapy,

Design/

pooled analysis of 5 available trials demonstrated a signifi-

cant reduction in all-cause mortality with CRT (RR 0.76,
95% CI: 0.64e0.9) (Fig. 2A). No statistically significant 15% of patients in both groups did not receive ICD therapy.
heterogeneity was found (P 5 .71; I2 5 0%). Results
CRTþICD vs. ICDy
CRT on vs. CRT off
CRT on vs. CRT off

were not altered by removal of trials with less than 10

CRTþICD vs. ICD
CRTþICD vs. ICD
CRTþICD vs. ICD

CRTþICD vs. ICD

CRTþICD vs. ICD
CRTþMT vs. MT

CRTþMT vs. MT

*Blinding of patient/caregiver/endpoint assessment.

Intervention

events (RR 0.76, 95% CI: 0.60e0.95), or by fixed-effect

CRT vs. MT
CRTþICD vs.

analysis (RR 0.76, 95% CI: 0.64e0.9).12e14,20

CRT vs. MT

All-Cause MortalitydCRT D ICD versus ICD alone

Pooled analysis from 5 trials that compared the combina-

1520

1820
1798
453
581
369
186

813

610
58
43

33
N

tion of CRT and ICD therapy to ICD therapy alone showed

a statistically significant reduction in all-cause mortality
trial; UC, unclear; Y, yes.

(risk ratio [RR] 0.83, 95% confidence interval [CI]:

MIRACLE ICD, 200316

COMPANION, 200418
CONTAK CD, 200315
MUSTIC SR, 200112

0.72e0.96) (Fig. 2B). There was no statistically significant

MADIT-CRT, 20099
HOBIPACE, 200620
MUSTIC A, 200214

REVERSE, 200821
MIRACLE ICD II,

CARE HF, 200519

heterogeneity among trials (P 5 .91; I2 5 0%). Results re-

MIRACL, 200213
Acronym, Year

RAFT, 201010

mained similar after excluding the trial in which 15% of pa-

tients in both groups did not receive ICD therapy (RR 0.82,
200417

95% CI: 0.71e0.95),21 with fixed-effect analysis (RR 0.83,

y

95% CI: 0.72e0.96), and after restricting analysis to studies

 CRT and ICD: An Updated Meta-Analysis
Fig. 2-A. All-cause mortalityeCRT versus MT. CRT, cardiac resynchronization therapy; IV, inverse variance method; MT, medical therapy.
with more than 6 months of follow-up (RR 0.83, 95% CI
0.71e0.96).9,10,21
Figure 3 represents the results of cumulative meta-
analysis, adding studies 1 at a time in chronological order,
demonstrating that the point estimate suggested a reduction
in mortality since the earlier trials. However, statistical sig-
nificance is evident only after adding the results of the
RAFT trial.
All-Cause MortalitydImpact of NYHA Class
Stratified analyses showed that the lowest risk-ratio for
all-cause mortality was found in clinical trials that re-
stricted inclusion to patients in NYHA Class III and IV.
However, a significant mortality reduction was also present
in studies that did not make such restriction, and in those
that included only patients in NYHA Class I or II (Fig. 4).
Implant Failure and ComplicationsdCRT-Capable
Devices
Implant failure rate for CRT devices was reported by all
but one20 of the included trials. Pooled analysis showed an
8% risk of implant failure (95% CI: 6e11). Results were sim-
ilar with fixed effect analysis. The rate of any major peri-
implant complications was reported by 7 trials.9,10,16e19,21
Pooled analysis resulted in a risk of major complication of
13.2% (95% CI: 7.3e23.9). Implant-related mortality was
reported by 5 trials,10,14,15,18,19 with a pooled estimation of
Fig. 2-B. All-cause mortalityeCRT þ ICD versus ICD alone. CRT, cardiac resynchronization therapy; ICD, implantable cardioverter de-
fibrillator therapy; IV, inverse variance method.
 Bertoldi et al 863
absolute risk of 0.6% (95% CI: 0.2%e2.2%) (Table 2).
Rate of peri-implant complications related to the left ventric-
ular lead was reported by 5 trials.9,16e19 Pooled analysis
showed a combined risk of 3% (95% CI: 1e8.7).
The rate of all types of implant-related complication
showed a progressive decline from earlier to more recent
studies. Restricting analysis to trials published after 2004
reduced the pooled risks of implant failure to 6.3% (95%
CI: 4.3e9.2), any peri-implant complication to 9.8%
(95% CI: 6.3e15.4), complications related to left ventricu-
lar lead to 1.8% (95% CI: 0.7e4.5), and implant-related
mortality to 0.2% (95% CI: 0.1e0.9).
Discussion
Since the publication of the CARE-HF trial19 and of pre-
vious meta-analyses,6e8 the effectiveness of CRT in reduc-
ing mortality of HF patients is well established, when
compared to medical therapy alone. However, because cur-
rent guidelines indicate the use of ICD therapy for most HF
patients with reduced left ventricular ejection fraction,3 the
consideration of adding CRT to patients already eligible to
receive ICD therapy is a clinically relevant dilemma in ev-
eryday practice, because patients with HF frequently fulfill
the indication criteria for both devices.
In this scenario, one must consider that the addition of
CRT to ICD therapy is an expensive strategy, and device
864 Journal of Cardiac Failure Vol. 17 No. 10 October 2011

Fig. 3. Cumulative meta-analysis of all-cause mortality with CRTþICD versus ICD alone. CRT, cardiac resynchronization therapy; ICD,
implantable cardioverter defibrillator therapy; IV, inverse variance method. Note that the point estimate suggests a reduction in mortality
since the earlier trials, but statistical significance only appears after addition of the RAFT trial.

implantation is technically more difficult and risky. Com-
mon sense and rationalization of resource use indicate
that hard clinical evidence demonstrating the additional
benefit should be available before broad implementation
of combined therapy (CRT þ ICD). Earlier trials had
used surrogate endpoints such as exercise capacity, and
larger trials focused on combined endpoints such as HF
or death. In this context, meta-analysis is a useful statistic
tool, making it possible to combine the results from all
available trials to produce more robust evidence of reduc-
tion of hard endpoints.
Previously published meta-analysis seeking to answer
this question yielded conflicting results,6e8 but they were
limited to indirect comparisons (such as meta-regression
or Bayesian meta-analysis), because data on direct
assessment of the 2 strategies were scarce. Since then, the
publication of REVERSE,21 MADIT-CRT,9 and the re-
cently published RAFT trial10 have added a great number
of HF patients that were allocated to combined device ther-
apy. RAFT trial was the first individual study to show a sig-
nificant reduction in mortality with the addition of CRT to
ICD therapy, whereas the MADIT-CRT trial failed to dem-
onstrate such benefit. The absence of survival benefit in ear-
lier trials, including MADIT-CRT, may be due, in part, to
the fact that previous studies comparing CRT þ ICD to
ICD alone have included patients in less advanced stages
of HF, and follow-up periods were relatively short (most of-
ten less than 1 year). The RAFT trial, in contrast, was the
study with the longest follow-up period among all CRT tri-
als. Our data indicate that all available studies were

Fig. 4. Stratified analysis of all-cause mortality according to NYHA class and inclusion of ICD therapy. CI, confidence interval; ICD, im-
plantable cardioverter-defibrillator; NYHA, New York Heart Association; RR, relative risk.
 CRT and ICD: An Updated Meta-Analysis
Table 2. Pooled Incidence of Implant-Related Complications
Complication Rate (Absolute)
Implant failure 8%
Any major complication 13.2%
LV lead complication 3%
Implant-related mortality 0.6%
LV, left ventricular.
homogeneous and that there is a significant beneficial effect
in adding CRT to most HF patients that are eligible for ICD
therapy. It is noteworthy that the pooled model that showed
that sequentially adding trials did not lead to statistical sig-
nificance until the RAFT trial was added. This major effect
of 1 trial can be explained because the number of events
was tripled after the RAFT trial was published. Moreover,
data from cumulative meta-analysis demonstrates consis-
tency of effect size over time (Fig. 3), and make a strong
case for a significant mortality reduction with the addition
of CRT to ICD.
Another potential benefit of the addition of CRT to ICD
therapy is the reduction of HF-related symptoms. Trials that
evaluated combined device therapy have demonstrated im-
provement in exercise capacity,15e17 improvement in
NYHA class,16,17 or composite clinical endpoints including
symptoms requiring augmentation of decongestive medica-
tions.9 Reduction of HF hospitalizations is a more robust
evidence of symptomatic benefit, especially considering
the partially unblinded nature of many of the trials. Our ini-
tial intention was to pool hospitalization data from trials
that compared CRT þ ICD to ICD alone. However, three
trials reported only the number of hospitalized pa-
tients,10,15,16 1 trial reported the total number of hospitali-
zations,21 and 1 of the 2 largest trials, MADIT-CRT,9 did
not report hospitalization rates. We considered that pooling
these data would not properly represent true effects. Unfor-
tunately, we were unable to obtain raw patient data directly
from authors, which would allow us to pool incidence of
a single outcome in all trials.
Our study also provided a pooled estimate of the risk of
peri-implant complications. These risks should not be
underestimated in the process of defining CRT indication
to a broad population of HF patients. Possibly related to
progress in device design and greater operator experience,
the incidences of all complications have declined in more
recent studies, as clearly demonstrated in the decrease in
peri-implant mortality over time.
Finally, HF severity seems to be an important aspect that
influences the magnitude of benefit in the response to CRT.
Individual trials that included patients in NYHA functional
Class I have not showed a significant mortality reduction.
When we performed stratified meta-analysis of the studies
according to NYHA class, we found that trials that limited
inclusion to patients in NYHA Class III and IV had the
greatest mortality reduction, although studies with other
combinations of NYHA class also had a significant benefit.
 Bertoldi et al 865
95% CI Studies with Available Data
6%e11% 9, 10, 12, 14-17, 19, 21
7.3%e23.9% 9, 10, 16-19, 21
1%e8.7% 9, 16-19
0.2%e2.2% 10, 14, 15, 18, 19
This information may be useful for selection of patients that
benefit the most from CRT, in budget-constrained scenar-
ios. Cost-effectiveness studies of CRT that incorporate
these latest findings may provide additional information
in the future.
Our study design deserves several considerations. Firstly,
the unavailability of raw data of HF hospitalization did not
allow us to produce a reliable pooled estimate of the impact
of CRT in HF symptoms. In addition, the follow-up period
for earlier trials that compared CRT þ ICD to ICD alone
was 12 months or less, which limits their ability to detect
long-term effects on mortality. The Kaplan-Meier estimates
of death from any cause in the RAFT trial show a slight in-
crease in the benefit of CRT over time, which suggests that
longer follow-up in the earlier trials might have yielded
a mortality reduction of greater magnitude. It is worth not-
ing, however, that restricting the analysis to trials with
follow-up longer than 6 months did not significantly alter
our results.
Publication bias should always be considered a potential
limitation in meta-analysis. In the case of our results, funnel
plots showed some asymmetry, but their results would sug-
gest that studies favoring CRT were less likely to be pub-
lished (Supplementary Figures A and B [online only]).
Consequently, if publication bias did exist, the true relative
risk reduction with CRT would be even greater than ob-
served in our results.
Conclusion
The current meta-analysis included new data from the
REVERSE, MADIT-CRT, and RAFT trial, adding a total
of 4228 HF patients to previous evaluations. Our findings
support a profound reduction in all-cause mortality with
CRT when compared to optimal medical therapy in selected
patients with systolic HF and prolonged QRS duration. Ad-
ditionally, our results show compelling evidence that the
combination of CRT and ICD therapy also reduces overall
mortality, when compared to ICD alone, an effect that is
greater in patients in NYHA Class III or IV. Finally, our
data suggest that morbidity and mortality associated to im-
plant procedure is not negligible, but has decreased over
time.
Disclosures
None.
866 Journal of Cardiac Failure Vol. 17 No. 10 October 2011
Supplementary Data
Supplementary data related to this article can be found
online at doi:10.1016/j.cardfail.2011.06.372
References
1. Mathers C, Fat DM, Boerma JT. World Health Organization. The
global burden of disease: 2004 update. Geneva, Switzerland: World
Health Organization; 2008.
2. Lloyd-Jones D, Adams RJ, Brown TM, Carnethon M, Dai S, De
Simone G, et al. Heart disease and stroke statisticsd2010 update: a re-
port from the American Heart Association. Circulation 2010;121:
e46ee215.
3. Jessup M, Abraham WT, Casey DE, Feldman AM, Francis GS,
Ganiats TG, et al. 2009 focused update: ACCF/AHA Guidelines for
the Diagnosis and Management of Heart Failure in Adults: a report
of the American College of Cardiology Foundation/American Heart
Association Task Force on Practice Guidelines: developed in collabo-
ration with the International Society for Heart and Lung Transplanta-
tion. Circulation 2009;119:1977e2016.
4. Ezekowitz JA, Armstrong PW, McAlister FA. Implantable cardioverter
defibrillators in primary and secondary prevention: a systematic review
of randomized, controlled trials. Ann Intern Med 2003;138:445e52.
5. Bradley DJ, Bradley EA, Baughman KL, Berger RD, Calkins H,
Goodman SN, et al. Cardiac resynchronization and death from pro-
gressive heart failure: a meta-analysis of randomized controlled trials.
JAMA 2003;289:730e40.
6. Freemantle N, Tharmanathan P, Calvert MJ, Abraham WT, Ghosh J,
Cleland JG. Cardiac resynchronization for patients with heart failure
due to left ventricular systolic dysfunctionda systematic review and
meta-analysis. Eur J Heart Fail 2006;8:433e40.
7. McAlister FA, Ezekowitz JA, Wiebe N, Rowe B, Spooner C, Crumley E,
et al. Systematic review: cardiac resynchronization in patients with
symptomatic heart failure. Ann Intern Med 2004;141:381e90.
8. Lam SK, Owen A. Combined resynchronisation and implantable defi-
brillator therapy in left ventricular dysfunction: Bayesian network
meta-analysis of randomized controlled trials. BMJ 2007;335:925.
9. Moss AJ, Hall WJ, Cannom DS, Klein H, Brown MW, Daubert JP,
et al. Cardiac-resynchronization therapy for the prevention of heart-
failure events. N Engl J Med 2009;361:1329e38.
10. Tang ASL, Wells GA, Talajic M, Arnold MO, Sheldon R, Connolly S,
et al. Cardiac-resynchronization therapy for mild-to-moderate heart
failure. N Engl J Med 2010;363:2385e95.
11. Higgins JPT, Green S, Cochrane Collaboration. Cochrane handbook
for systematic reviews of interventions. Chichester, UK: Wiley-Black-
well; 2008.
12. Cazeau S, Leclercq C, Lavergne T, Walker S, Varma C, Linde C, et al.
Effects of multisite biventricular pacing in patients with heart failure
and intraventricular conduction delay. N Engl J Med 2001;344:
873e80.
13. Abraham WT, Fisher WG, Smith AL, Delurgio DB, Leon AR, Loh E,
et al. Cardiac resynchronization in chronic heart failure. N Engl J Med
2002;346:1845e53.
14. Leclercq C, Walker S, Linde C, Clementy J, Marshall AJ, Ritter P,
et al. Comparative effects of permanent biventricular and right-
univentricular pacing in heart failure patients with chronic atrial fibril-
lation. Eur Heart J 2002;23:1780e7.
15. Higgins SL, Hummel JD, Niazi IK, Giudici MC, Worley SJ,
Saxon LA, et al. Cardiac resynchronization therapy for the treatment
of heart failure in patients with intraventricular conduction delay
and malignant ventricular tachyarrhythmias. J Am Coll Cardiol
2003;42:1454e9.
16. Young JB, Abraham WT, Smith AL, Leon AR, Lieberman R,
Wilkoff B, et al. Combined Cardiac Resynchronization and Implant-
able Cardioversion Defibrillation in Advanced Chronic Heart Failure:
the MIRACLE ICD Trial. JAMA 2003;289:2685e94.
17. Abraham WT, Young JB, Leon AR, Adler S, Bank AJ, Hall SA, et al.
Effects of cardiac resynchronization on disease progression in patients
with left ventricular systolic dysfunction, an indication for an implant-
able cardioverter-defibrillator, and mildly symptomatic chronic heart
failure. Circulation 2004;110:2864e8.
18. Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, De
Marco T, et al. Cardiac-resynchronization therapy with or without
an Implantable Defibrillator in Advanced Chronic Heart Failure. N
Engl J Med 2004;350:2140e50.
19. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D,
Kappenberger L, Tavazzi L. The effect of cardiac resynchronization
on morbidity and mortality in heart failure. N Engl J Med 2005;352:
1539e49.
20. Kindermann M, Hennen B, Jung J, Geisel J, Bohm M, Frohlig G. Bi-
ventricular versus conventional right ventricular stimulation for pa-
tients with standard pacing indication and left ventricular
dysfunction: the Homburg Biventricular Pacing Evaluation (HOBI-
PACE). J Am Coll Cardiol 2006;47:1927e37.
21. Linde C, Abraham WT, Gold MR, St John Sutton M, Ghio S,
Daubert C, Group RS. Randomized trial of cardiac resynchronization
in mildly symptomatic heart failure patients and in asymptomatic pa-
tients with left ventricular dysfunction and previous heart failure
symptoms. J Am Coll Cardiol 2008;52:1834e43.