OUP user menu

Independent predictors of mortality in patients with advanced heart failure treated by cardiac resynchronization therapy

Jens Kreuz, Fritz Horlbeck, Markus Linhart, Fritz Mellert, Rolf Fimmers, Jan Schrickel, Georg Nickenig, Jörg Otto Schwab
DOI: http://dx.doi.org/10.1093/europace/eus152 1596-1601 First published online: 18 May 2012


Aims The current study includes all consecutive patients with advanced heart failure and cardiac resynchronization therapy (CRT) with an implantable cardioverter defibrillator (ICD) over a 10-year period in a tertiary referral centre. It aims at identifying independent risk factors for mortality during CRT-defibrillator (CRT-D) treatment.

Methods and results This study includes 239 consecutive patients who had undergone implantation of a CRT-D system (ejection fraction 25.9 ± 8%; 139 patients with ischaemic, 100 patients with non-ischaemic cardiomyopathy). Enrolment took place between 2001 and 2010, resulting in a median follow-up of 43 ± 30 months. During follow-up, 59 patients (25%) died. An impaired baseline kidney function [hazard ratio (HR) 1.98; 95% confidence interval (CI) 1.7–3; P< 0.0001], appropriate ICD therapy during follow-up (HR 2.1; CI 1.1–3.4; P= 0.001), lack of beta-blocker therapy (HR 2.3; CI 1.6–3.8; P= 0.004), and intake of amiodarone (HR 2; CI 1.8–4.1; P< 0.0001) were identified as predictors of overall mortality.

Conclusion This study demonstrates the benefit of beta-blocker therapy also in patients on long-term CRT-D treatment. It confirms the prognostic significance of impaired renal function and the occurrence of appropriate ICD therapies also in CRT-D patients. It argues for an intensified follow-up regimen and adjustment of heart failure treatment whenever these prognostic markers are identified in a patient treated with CRT-D.

  • Cardiac resynchronization therapy
  • Overall mortality
  • Risk stratification
  • Renal dysfunction


An implantable cardioverter defibrillator (ICD) represents the gold standard in primary and secondary prevention of sudden cardiac death.14 In addition, several previous randomized studies demonstrated not only the reduction of heart failure mortality, but even a substantial survival benefit for patients with moderate-to-severe chronic heart failure, who underwent cardiac resynchronization therapy (CRT).57 Two recent milestone trials further showed that the implantation of a CRT-defibrillator (CRT-D) as opposed to an isolated single- or dual-chamber ICD8,9 engendered a significant survival benefit and, in fact, the reduction of heart failure hospitalization. These two studies enrolled patients with primary prevention ICD indication with mild-to-moderate chronic heart failure [New York Heart Association (NYHA) stages I–III], impaired left ventricular ejection fraction (EF < 30%), and a wide QRS complex.

The current retrospective study represents all consecutive CRT-D patients at a single university centre over a 10-year period. It seeks to identify non-invasive, independent risk factors for overall mortality in this cohort of patients, who suffer from relevant chronic congestive heart failure.


This study includes 239 consecutive patients, who had undergone an implantation of a CRT-D device. Its primary goal revolves around discovering independent predictors of overall mortality. The enrolment took place between 2001 and 2010, resulting in a median follow-up period of 43 ± 30 months. Detailed information about the patients' demographics, baseline medication, and the results of the univariate mortality analysis are shown in Table 1.

View this table:
Table 1

Demographics and results of univariate analysis regarding overall mortality (baseline medication)

Surviving patients at follow up (n= 180)Deceased patients (n= 59)P value
Age (years)a66 ± 1069 ± 80.003
Genderb (men)141510.13
LVEF (%)a26 ± 1025 ± 90.15
Hypertensionb n (%)95 (53)22 (37)0.1
Diabetes mellitusb n (%)61 (34)14 (24)0.48
CABGb n40160.35
Atrial fibrillationb n (%)71 (39)23 (39)0.78
Appropriate ICD interventionsb n (%)47 (26)32 (54)0.002
NYHA class II/IIIb n54/12612/470.34
ICM/NICMb n103/7736/230.31
Serum creatininea (mg/dL)1.47 ± 0.61.9 ± 0.8<0.0001
MIb n74 (41)23 (39)0.14
QRS widtha (ms)145 ± 16150 ± 230.9
Beta-blocker n (%)171 (95)50 (85)0.006
ACE-inhibitors/ARB n (%)171 (95)57 (97)0.79
Aldosteron-antagonists n (%)97 (54)33 (56)0.66
Amiodarone n (%)28 (15)24 (41)0.0001
Loop diuretics n (%)165 (91)53 (60)0.62
Phenprocoumon n (%)92 (51)27 (46)0.33
Statin n (%)119 (66)36 (61)0.73
Digitalis n (%)50 (28)19 (32)0.95
  • ARB, angiotensin receptor blockers; MI, myocardial infarction; CABG, coronary artery bypass graft; ICM, ischaemic cardiomyopathy; NICM, non-ischaemic cardiomyopathy.

  • aUsage of Mann–Whitney test.

  • bUsage of Fisher's exact test.

Each patient was followed in the outpatient clinic of our university hospital quarterly. All parameters used in the risk analysis for overall mortality were determined on the grounds of the patients' medical records before CRT-D implantation, except from the incidence of appropriate ICD therapies during follow-up. Each visit then included a thorough device interrogation with special focus on the incidence of ventricular tachyarrhythmias. Every intracardiac electrocardiogram was reviewed by two independent cardiologists to validate the malignant nature of the episode. Appropriate therapies were applied for either sustained episodes of ventricular tachycardia (VT; malignant arrhythmias up to 240 b.p.m.) or sustained episodes of ventricular fibrillation (VF; malignant arrhythmias exceeding 240 b.p.m.).

Statistical analysis

Statistical analyses were applied using the SPSS 12.0 system (SPSS Inc., Chicago, IL, USA). We conducted both univariate and multivariate analyses (stepwise regression model) to identify independent risk factors for overall mortality. Finally, a Kaplan–Meier analysis was performed to detect significant differences in overall mortality in dependence of several baseline parameters. Moreover, the Fisher's exact and Wilcoxon test were used to determine the association between baseline characteristics and amiodarone medication. A P value ≤0.05 was defined as statistically significant.


During a median follow-up of 43 ± 30 months, 79 patients (33%) suffered from sustained malignant arrhythmias treated by the ICD and 59 patients (25%) died. Table 1 displays the results of the univariate mortality analysis. Applying the Kaplan–Meier method showed striking differences in survival of patients with a baseline serum creatinine <1.2 mg/dL in comparison to those with a baseline serum creatinine ≥1.2 mg/dL (Figure 1). Figures 24 respectively display the significant impact on overall mortality caused by the incidence of appropriate ICD therapies during follow-up (Figure 2), lack of beta-blocker therapy (Figure 3), and amiodarone intake (Figure 4).

Figure 1

Survival of patients with a baseline serum creatinine <1.2 vs. ≥1.2 mg/dL. Patients with a creatinine ≥1.2 mg/dL show a poorer outcome.

Figure 2

Survival of patients with appropriate implantable cardioverter defibrillator therapies for malignant arrhythmias during follow in comparison with patients without these interventions. VT, ventricular tachycardia; VF, ventricular fibrillation.

Figure 3

An overview regarding the survival of patients with and without beta-blocker therapy.

Figure 4

An overview regarding the survival of patients with and without amiodarone therapy.

The multivariate analysis revealed an independent mortality association for those patients who did not take beta-blockers, suffered from impaired renal function (measured by an increased pre-operative serum creatinine), and took amiodarone (Table 2). Furthermore, appropriate ICD therapies for either VT or VF during follow-up could be identified as independent predictors of the patients' mortality (Table 2). The majority of patients who took amiodarone had a secondary prevention ICD indication (40 vs. 24%; P= 0.02), and a higher baseline serum creatinine (1.67 vs. 1.29 mg/dL; P= 0.0009). Moreover, patients under amiodarone medication demonstrated a co-medication with phenprocoumon more frequently (70 vs. 40%; P= 0.001). However, patients receiving amiodarone revealed merely a slight trend towards a more frequent history of atrial fibrillation (50 vs. 37% P= 0.08).

View this table:
Table 2

Multivariate analysis of risk factors for overall mortality

Hazard ratio (95% confidence interval)P value
Serum creatininea1.98 (1.7–3)0.0007
Age at implantb1.6 (1.1–3.4)0.01
Amiodarone intake2 (1.6–3.8)0.02
VT/VFc during follow-up2.1 (1.8–4.1)0.006
Absence of beta-blocker therapy2.3 (1.6–3.8)0.003
  • aPer baseline serum creatinine increase of 0.2 mg/dL.

  • bPer 10 years of age.

  • cOccurrence of appropriate implantable cardioverter defibrillator therapies for either ventricular tachycardia (VT) or ventricular fibrillation (VF) during follow-up.


Cardiac resynchronization therapy in connection with ICD therapy is an established means to improve the life expectancy of patients suffering from an advanced stage of congestive heart failure.10,11 Several previous studies evaluated the significance of various co-morbidities, such as age or baseline medication, for the mortality rate in CRT-D recipients.1218 The main objective of this study was to discover independent predictors of mortality in a cohort of patients with CRT-D implantation.

Impact of impaired renal function

The results of the present investigation, in line with those of several previous studies, indicate that renal dysfunction poses a strong and independent risk factor for overall mortality despite CRT-D implantation and optimized medical treatment of congestive heart failure. As a matter of fact, the results of this study indicate a two-fold increase in the documented mortality rate in connection with a 0.2 mg/dL increase of baseline serum creatinine. In addition, the authors could demonstrate that a cut-off value of 1.2 mg/dL baseline serum creatinine provided a significant predictor of the patients' mortality. The serum creatinine cut-off value for the Kaplan–Meier analysis (1.2 mg/dL) was chosen because values exceeding this limit are defined as pathological in our local laboratory. This decision supports the work of Bai et al., who define chronic kidney disease by a constant serum creatinine >1.4 mg/dL. This value further entails the same negative survival effect in a similar cohort of patients.12 Accordingly, patients with chronic kidney disease revealed a significantly lower, 2-year survival rate (<50 vs. 90% in patients without chronic kidney disease). This mortality rate is somewhat higher than the one documented in our investigation (75 vs. 92%), as might be explained by the higher creatinine cut-off value Bai et al. used (1.4 vs. 1.2 mg/dL in our study). In addition, in the collective of Bai et al., the patients demonstrated a markedly lower EF (19.9 vs. 25%) than in the present trial. Another recently published trial and, in addition, the investigations of the COMPANION trial report an independent mortality risk in connection with renal dysfunction in CRT-D patients.13,19

Taking the findings of the current study and of these previous trials into account, a grave, pre-operative impaired renal dysfunction in an individual suffering from advanced chronic congestive heart failure and the possible adverse effects of this co-morbidity should be openly discussed with the potential recipient of the device and his close relatives.

Impact of a lack of beta-blocker therapy

Beta-blocker therapy should be, unless contraindicated, an essential component of the medical treatment of any patient who suffers from an advanced congestive heart failure and who may therefore be a suitable CRT recipient (symptomatic heart failure and left ventricular EF≤40%).20 Voigt et al.15 report that lack of beta-blocker therapy in a smaller cohort of CRT patients (n= 177) was associated with a three-fold increase in the risk of death or heart transplantation. The current study, including 239 participants and thus 35% more CRT-D patients than Vogt et al.'s preceding trial, confirms these results. What is more, it underlines the unavoidability of beta-blocker therapy in this endangered cohort of patients. Lately, the group of Mullens et al. published a study in which they investigated whether a protocol-driven approach incorporated in a management strategy of heart failure immediately after implantation would be able to provide benefits beyond usual care after CRT-D implantation. They concluded, that a protocol-driven approach, e.g. containing the up-titration of neurohormonal beta-blockers, had incremental favourable effects on reverse remodelling and resulted in fewer adverse events compared with usual care after implantation.21 This study demonstrates that not only the presence of a beta-blocker therapy but even its correct up-titration and administration to the patient is of crucial importance for the extraordinary survival benefit of this drug. The prognostic importance of the maximization of beta-blocker therapy in heart failure patients has also been highlighted in the above-mentioned European Society of Cardiology guidelines.20

Amiodarone intake and mortality

The antiarrhythmic potential of amiodarone is undisputable, as are its potentially life-threatening side effects.2225 The data of the SCD-HeFT trial demonstrated impressively that amiodarone cannot compete with an ICD in terms of the reduction of overall mortality in a collective of primary prevention patients suffering from chronic congestive heart failure and a reduced left ventricular EF (≤35%).3 In fact, patients in the amiodarone arm revealed the same survival benefit as the placebo group in the SCD-HeFT collective. In 2006, the OPTIC investigators reported that a combination of beta-blocker and amiodarone medication resulted in a significant reduction of appropriate ICD shocks for malignant arrhythmias. This benefit was associated with an increased rate of side effects of the combined medical treatment. In this study, no increased mortality was documented.26 Ho et al.27 and Singh et al.28 showed a neutral effect on the survival of ICD patients. Unfortunately, these studies do not report on the composition of single-, dual- and triple-chamber devices.

In the current trial, the use of the Fisher's exact and Wilcoxon test provided some significant differences in patients who receive or do not receive amiodarone. For instance, amiodarone intake was associated with secondary prevention ICD indication, phenprocoumon co-medication, and an impaired renal baseline function. It was particularly the significantly worse renal function of patients under amiodarone medication that implied a plausible explanation for the poorer outcome of these patients in the current study. These results clearly demonstrate that an impaired renal function constitutes a relevant life-threatening co-morbidity, which, especially in combination with grave heart failure and amiodarone intake, implicates an increased mortality risk. At any rate, the findings of this study have to be interpreted with due caution because the trial includes only a limited number of patients and was not designed to test amiodarone and its influence on mortality. In addition, it is of a retrospective nature. Of course, it needs to be taken into account that some patients need to take amiodarone despite an ICD implantation, e.g. to control rapidly conducted atrial fibrillation or recurrent ventricular arrhythmias. Hence, the authors of this manuscript conclude that the administration of amiodarone is still appropriate in patients with a meaningful indication for this drug.

Appropriate ICD therapies during follow-up and mortality

Appropriate ICD therapies, especially shock interventions, seem to be associated with a poorer prognosis in patients with ICD and heart failure.2932 In detail, the study of Poole et al.29 included no CRT-D patients. Dichtl et al.30 enrolled a comparable number of these CRT-D patients in comparison with the current investigation. The PREPARE trial enrolled the greatest number of CRT-D patients (n= 662).32 In the present investigation the application of an appropriate ICD therapy for either VT or VF was associated with a two-fold increase in mortality. This fact can be interpreted as the sign of an advanced stage of heart failure in which, as a natural course, the patient dies.

These findings in a cohort of patients undergoing CRT therapy, once again, stress the importance of a CRT-D device in contrast to a mere biventricular pacemaker (CRT-P). Reacting to the findings of the current study, any physician entrusted with the treatment of CRT-D patients should be alerted when appropriate ICD therapies occur. In this case, any potential optimization of medical and supportive heart failure treatment should be initiated and an intensification of follow-up intervals should be considered. In patients with coronary artery disease, an ischaemic trigger of the malignant arrhythmias should be evaluated. If an ischaemic problem seems to be evident, the indication for an invasive coronary diagnostic and therapy should be considered generously. Furthermore, it should be evaluated whether the quality of CRT treatment could be improved, e.g. by means of echocardiographic optimization. The importance of this strategy is substantiated by a recent report of the MADIT CRT investigators.33 Barsheshet et al. showed that reverse remodelling, achieved by an effective CRT therapy, was associated with a significant reduction in the risk of subsequent life-threatening arrhythmias. In addition, innovative heart failure treatment, such as chronic vagal nerve stimulation or cardiac contractility modulation should be considered.34,35


This study has several limitations. First, this observational study is not randomized and is of a retrospective nature. The association of the identified risk factors with increased mortality suggests but does not necessarily indicate that they are predictors of mortality. Therefore, prospective confirmation would be required. In addition, their relationship, although important to identify, does not indicate that their modification would influence the outcome. On the other hand one has to keep in mind that this retrospective analysis identified multiple significant predictors of mortality in a great CRT cohort in a long-term follow up. In comparison with various prospective randomized trials, which mostly explore only a few parameters, this trial identified various parameters associated with an increased mortality without any randomization bias, reflecting a ‘real world scenario’. Secondly, there exists no unified device programming concerning arrhythmia detection or antitachycardia pacing. Because of this, results concerning ATP success were deliberately not demonstrated. As a matter of fact, this investigation, which only includes a limited number of patients with a very different follow-up duration, includes more than one generation of CRT-D devices. On the other hand, one has to state that no past or current trial was able to show any survival benefit using a certain CRT-D device and its diagnostic capacities (e.g. fluid monitoring for heart failure monitoring) so that this fact should not have a great impact on the survival of the patients participating in this trial. On the other hand, one has to state that no past or current trial was able to show any survival benefit using a certain CRT-D device and its diagnostic capacities (e.g. fluid monitoring for heart failure monitoring etc.) so that this fact should not have a great impact on the survival of the patients participating in this trial. The currently ongoing OtiLink-HF study (NCT00769457) is the first major randomized prospective trial intended to investigate the potential extent of new technologies—like intrathoracic fluid monitoring—to prevent potentially adverse cardiac situations and/or hospitalizations and if there is an influence on the duration of patient's lives'.


This study urges the uncompromising use of beta-blocker therapy in patients undergoing CRT-D implantations, with the exception of absolute contraindications. It also underlines the prognostic relevance of a pre-implantation impaired renal function. The possible adverse effects of this co-morbidity should be openly discussed with the patient. Moreover, the termination of malignant arrhythmias during the course of CRT therapy should be interpreted as the harbinger of a worsened prognosis. A more intensive follow-up regimen, thorough evaluation, and, if necessary, adaptation of the patient's heart failure treatment is mandatory.

Conflict of interest: none declared.


View Abstract