Kidney Dysfunction and Deterioration of Ejection Fraction Pose Independent Risk Factors for Mortality in Implantable Cardioverter-Defibrillator Recipients for Primary Prevention




A growing number of patients with advanced heart failure fulfill a primary-prevention indication for an implantable cardioverter-defibrillator (ICD). This study seeks to identify new predictors of overall mortality in a Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT)-like collective to enhance risk stratification.


An impaired renal function and severely depressed left ventricular ejection fraction pose relevant risk factors for mortality in primary prevention ICD recipients.


Ninety-four consecutive ICD patients with New York Heart Association class II–III heart failure and depressed left ventricular function (left ventricular ejection fraction [LVEF] ≤35%) with no history of malignant ventricular arrhythmias were followed for 34 ± 20 months.


During this period, 30 patients died (32%). Deceased patients revealed a significantly worse renal function before ICD implantation (1.55 ± 0.7 mg/dL vs 1.1 ± 0.4 mg/dL; P = 0.007), suffered more often from coronary artery disease (53 vs 29; P = 0.006), and were older (69.5 ± 8 y vs 67 ± 12 y; P = 0.0002) than surviving patients. Furthermore, increased serum creatinine at baseline (2 mg/dL vs 1 mg/dL; odds ratio [OR]: 3.96, 95% confidence interval [CI]: 1.2–13.04, P = 0.02), presence of coronary artery disease (OR: 8.6, 95% CI: 1.1–65, P = 0.036), and low LVEF (OR per 5% baseline LVEF deterioration: 1.4, 95% CI: 1–1.8, P = 0.034) represented strong and independent predictors for overall mortality.


Impaired renal function, the presence of coronary artery disease, and reduced LVEF before implantation represent independent predictors for mortality in a cohort of patients with advanced systolic heart failure. These conditions still bear a high mortality risk, even if ICD implantation effectively prevents sudden arrhythmic death. Indeed, in patients suffering from several of the identified “high-risk” comorbidities, primary-prevention ICD implantation might have a limited survival benefit. The possible adverse effects of these comorbidities should be openly discussed with the potential ICD recipient and his or her close relatives. Clin. Cardiol. 2012 doi: 10.1002/clc.22018

The authors have no funding, financial relationships, or conflicts of interest to disclose.


Sudden cardiac death is known as a common cause of death in industrialized countries. Many previous studies demonstrate that low left ventricular ejection fraction (LVEF) in combination with either coronary artery disease (CAD) or nonischemic cardiomyopathy poses a relevant risk factor for sudden cardiac death1–4 and endangers the lives of patients belonging to this possibly high-risk group. The implantation of an implantable cardioverter-defibrillator (ICD) has been widely acknowledged to constitute the best-practice model to avoid such fatal events.5 Any physician dealing with primary prevention of sudden cardiac death is confronted with an increasing number of potential ICD recipients. Yet, despite primary-prevention ICD implantation, the majority of patients die from a progression of heart failure.6 Accordingly, an analysis of the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) data demonstrates that heart failure mortality is not reduced by implantation of a single-chamber ICD.7 Levy et al present a model of risk stratification for overall mortality. They identify circumstances under which patients suffering from certain comorbidities do not benefit from primary-prevention ICD implantation.8 Several previous publications also describe potential independent risk factors for the overall mortality of these endangered patients.9–11 The group of Cuculich et al suggest a limited survival benefit of primary-prevention ICD implantation in patients with chronic kidney disease.12

The current study seeks to identify independent, noninvasive risk factors for overall mortality in a cohort of high-risk, SCD-HeFT–like patients equipped with a single- or dual-chamber ICD for primary prevention. The main goal of this trial is the identification of independently predictive patient characteristics that could provide further insight into optimal utilization of a primary-prevention ICD device.


This retrospective observational study aims to identify independent risk factors for overall mortality in a primary-prevention cohort. For this purpose, the following baseline variables (collected in the days before ICD implantation) were chosen for univariate and multivariate analyses: age, LVEF, hypertension, presence of diabetes mellitus, coronary artery bypass graft (CABG), atrial fibrillation, occurrence of malignant ventricular arrhythmias, New York Heart Association (NYHA) class, underlying heart disease, serum creatinine, QRS width, prior myocardial infarction (MI) or percutaneous coronary intervention (PCI), and medication. All parameters were determined on the grounds of the patients' medical records before ICD implantation. Detailed information about the patients' demographics and baseline medication is shown in Table 1. Moreover, based on the results of the stepwise regression analysis, the authors intended to create a model to predict the probability of the included patients' deaths.

Table 1. Demographics and Results of Univariate Analysis Regarding Overall Mortality—Baseline Medication
 Surviving Patients at Follow-up (n = 64)Deceased Patients (n = 30)P Value
  • Abbreviations: ACEI, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin II receptor blocker; CABG, coronary artery bypass graft; CAD, coronary artery disease; DCM, dilated cardiomyopathy; DM, diabetes mellitus; ICD, implantable cardioverter-defibrillator; LVEF, left ventricular ejection fraction; MI, myocardial infarction; NYHA, New York Heart Association; PCI, percutaneous coronary intervention.

  • a

    Mann-Whitney Utest.

  • b

    Fisher exact test.

Age, ya67 ± 1269.5 ± 80.37
LVEF (%)a30 ± 625 ± 70.01
Hypertension, n (%)b32 (50)19 (63)0.2
DM, n (%)b9 (14)8 (27)0.17
CABG, n (%)b22 (42)9 (31)0.67
AF, n (%)b12 (19)8 (27)0.39
Appropriate ICD interventions, n (%)32 (50)18 (60)0.33
NYHA class II/III, nb31/3310/200.24
CAD/DCM, nb53/1129/10.011
Serum creatinine, mg/dLa1.1 ± 0.391.55 ± 0.670.007
Prior PCI/MI, n (%)b49 (92)25 (86)0.45
QRS width, msa115 ± 11117 ± 150.63
β-Blocker, n (%)58 (91)25 (83)0.461
ACEI/ARB, n (%)63 (98)28 (93)0.48
Aldosterone antagonist, n (%)24 (37.5)7 (23)0.17

Each patient was followed quarterly in the outpatient clinic of our university hospital. Patients under cardiac resynchronization therapy were excluded because of the potential improvement of left ventricular function during therapy, which would in turn contradict a primary-prevention ICD indication. Furthermore, a detailed analysis of the incidence of both appropriate and inappropriate ICD interventions in this cohort of 94 patients implanted with a single- or dual-chamber ICD from October 2000 to March 2006 (NYHA class II–III and LVEF ≤35%, with no history of malignant ventricular arrhythmia or syncope) is published elsewhere.13

The authors of this manuscript have certified that they comply with the Principles of Ethical Publishing in the International Journal of Cardiology.14

Statistical Analysis

Statistical analyses were performed by using SPSS software, version 12.0 (SPSS Inc., Chicago, IL). We applied a univariate analysis and a stepwise regression model to identify the mortality risk of the ICD recipients included in this study. Based on the results of the stepwise regression analysis, we calculated the effects of deteriorating LVEF and of advancing impaired renal function on mortality. Finally, the Kaplan-Meier method predicted differences in survival depending on renal function. A P value ≤0.05 was defined as statistically significant.


During a median follow-up of 1080 ± 600 days, 30 patients (32%) died. Deceased participants had a significantly higher baseline serum creatinine (1.55 ± 0.67 mg/dL vs 1.1 ± 0.39 mg/dL; P = 0.007), showed more frequent occurrences of CAD (29 [97%] vs 53 [83%] patients; P = 0.011), and lower LVEF (25% ± 7% vs 30% ± 6%; P = 0.01). None of the other investigated baseline characteristics provided any significant findings in the univariate analysis (Table 1). In fact, deceased patients underwent an equal number of CABG procedures (31% vs 42%; P = 0.67) and the rate of former PCI was identical in deceased (25; 86%) and surviving (49; 92%; P = 0.45) patients. Based on the stepwise regression model, increased baseline serum creatinine (2 mg/dL vs 1 mg/dL; odds ratio [OR]: 5.4, 95% confidence interval [CI]: 2.3-12.08), presence of CAD (OR: 8.6, 95% CI: 1.1-65), and reduced LVEF (OR: 1.4 per 5% deterioration of LVEF, 95% CI: 1-1.8) represent independent predictors for overall mortality in this endangered cohort. Patients suffering from CAD had an LVEF similar to patients with nonischemic cardiomyopathy (30% ± 6% vs 30% ± 4%; P = 0.76). Significant results of the stepwise regression analysis are illustrated in Table 2. Based on the results derived from this analysis, we created a statistic model that estimates the increase of the relative mortality risk of patients with worsening baseline serum creatinine values in connection with deteriorating LVEF. Table 3 shows the expected values for several pairs of these 2 parameters. Therefore, a patient with a serum creatinine of ≤0.9 mg/dL and a baseline LVEF of 35% was defined as risk 1. This analysis indicates a significant worsening of the patient outcome related to the combination of progressive heart and renal failure. The Figure 1 portrays the results of the Kaplan-Meier analysis, showing an immense influence of more advanced kidney dysfunction on the patient survival. In fact, patients with a baseline serum creatinine ≥2 mg/dL show a significantly reduced survival compared with patients with a baseline serum creatinine <2 mg/dL (P = 0.001). Patients with worse baseline renal function exhibit a 1-, 2-, and 3-year mortality of 15%, 32%, and 42%, respectively. Patients with creatinine <2 mg showed significantly lower mortality rates, of 0%, 10%, and 15%, respectively.

Figure 1.

Probability of survival in dependence of kidney function. The collective was divided into 2 groups defined by baseline serum creatinine ≥2 mg/dL and <2 mg/dl.

Table 2. Significant Parameters of Stepwise Regression Analysis
 OR (95% CI)P Value
  • Abbreviations: CAD, coronary artery disease; CI, confidence interval; LVEF, left ventricular ejection fraction; OR, odds ratio.

  • a

    Baseline serum creatinine 2 mg/dL compared with 1 mg/dL.

  • b

    Per 5% deterioration of baseline LVEF.

Serum creatininea5.4 (2.3–12.08)<0.001
CAD8.6 (1.1–65)0.036
LVEFb1.4 (1–1.8)0.034
Table 3. Comparison of the Increased Relative Mortality Risk Compared With a Patient With a Baseline Creatinine of ≤0.9 mg/dL and an LVEF of 35%
Creatinine, mg/dLLVEF, %
  1. Abbreviations: LVEF, left ventricular ejection fraction.



Nowadays it is widely accepted, mainly through guideline implementation, that patients with reduced left ventricular function in NYHA class II or III heart failure suffering from CAD or nonischemic cardiomyopathy profit from primary-prevention ICD implantation to prevent sudden cardiac death.5 Some previous studies, however, identified certain independent predictors of increased mortality in this endangered cohort of patients.8–11 Appropriate as well as inappropriate shock interventions, for instance, pose significant risk factors for mortality in primary-prevention patients recruited from the SCD-HeFT collective.9 Olshansky reports that syncope is associated with an increased risk of overall mortality in all 3 treatment arms of the SCD-HeFT study cohort.10 These data demonstrate that patients with advanced heart failure (NYHA class II and III) and impaired LVEF (≤35%) are not only in increased danger of suffering from malignant ventricular arrhythmias, but they are also threatened by an increased mortality risk.

The current study presents an analysis of independent predictors for overall mortality in a cohort of SCD-HeFT–like patients with an ICD for primary prevention. Preimplantation impaired renal function, presence of CAD, and impaired preimplantation LVEF were associated with a markedly increased risk of death.

By way of summary, 30 patients (32%) died during an average follow-up period of 3 years. This can be positively correlated with the original SCD-HeFT cohort. Packer et al report a comparable mortality rate of 26.4% within a slightly longer follow-up (666 deaths in 2521 patients; median follow-up, 1350 days).7 Cuculich et al analyzed a cohort of 229 primary-prevention patients. In this study, a lower mortality rate was documented (15%), which may be explained by the shorter follow-up period (35 of 229 patients died; median follow-up, 18 months).12

Impact of Coronary Artery Disease, Low Left Ventricular Ejection Fraction, and Impaired Renal Function on Mortality

The fact that the presence of CAD has such an immense impact on patient mortality is surprising. Deceased patients had a slightly, though not significantly, lower rate of CABG (31 vs 42%; P = 0.67). Moreover, the rate of former PCI was identical in deceased and surviving patients.

None of the former original SCD-HeFT publications revealed such an enormous impact of the underlying disease on the patients' survival.2,8–11 Even though this finding constitutes a pertinent result of the present investigation, it might have been influenced by the low number of enrolled patients. Therefore, the authors acknowledge that it is difficult to generalize their findings and that the comprehensive applicability of the data remains a contested issue. The association of deteriorating LVEF with increased mortality is quite logical. This phenomenon can be described as a consequence of more advanced heart disease, which naturally ends with the patient's death. This statement becomes even more persuasive when one takes into account the results of our combined mortality risk evaluation pictured in Table 3. A grave kidney dysfunction, documented by a preimplantation serum creatinine of 3.5 mg/dL, in combination with impaired LVEF of 10%, increases the mortality risk to more than the 300-fold value of patients with LVEF of 35% and serum creatinine of ≤0.9 mg/dL (“low-risk patient”). In addition, a patient with LVEF of 10% and serum creatinine of 1.0 mg/dL suffers from almost the same increase in mortality risk as a patient with 35% LVEF with creatinine of 2.0 mg/dL. Based on the results of the present investigation, both patient groups have a 6-fold increase of mortality risk compared with the “low-risk patient group.” These examples illustrate the enormous impact the results of this study could have on therapy decisions made by an implanting physician. Cuculich et al also document that an impaired renal function is associated with significantly worse outcomes in a primary-prevention collective.12 This decrease of life expectancy grows with the severity of renal failure. These findings are confirmed by the Kaplan-Meier analysis performed in the present study, in which patients with grave kidney dysfunction suffer from a significantly limited life expectancy, especially in comparison with patients with creatinine <2 mg/dL. Amin et al present a model correlating the severity of known impaired renal function and its influence on decisions about the meaningfulness of ICD implantation for primary prevention.15 Beyond the findings of these studies focusing on the impact of renal failure, our investigation reveals a combination of significant mortality risk factors. Taking these readily available parameters into account, it is possible to estimate the potential survival benefit of a likely ICD recipient. These data, if confirmed in a larger, prospectively conducted observational study, would allow healthcare providers to make more informed decisions about the utility of an ICD for primary prevention and discuss the benefit of such therapy with the family of the potential recipient.

Study Limitations

This study is of retrospective design and includes a small cohort of 94 patients. Therefore, its results cannot be used to adapt existing guidelines. Nonetheless, the results of this study represent “real-life” data in which no randomization bias is included. The main goal of this investigation was to examine whether or not readily available patient characteristics could be identified that might potentially help in better utilization of primary-prevention ICDs.


Today, the healthcare system is confronted with a growing number of patients who fulfil the criteria for a primary-prevention ICD. When one considers the economic aspects and potential adverse effects of ICD implantation, it would be ideal to identify those patients who will have a disproportionately high mortality risk despite ICD implantation. The findings of this preliminary study, if confirmed, could equip physicians with valuable, readily available markers to identify patients with an unacceptably high mortality risk who are unlikely to benefit from ICD implantation.