The present study included patients referred for defibrillator implantation for primary prevention with or without resynchronization therapy according to current recommendations.7

The following patients were excluded: those with channelopathies, arrhythmogenic right ventricular dysplasia, hypertrophic cardiomyopathy, and congenital heart disease.

Cardiovascular mortality included sudden death and death attributed to myocardial infarction or heart failure. Sudden death was defined as death in the first hour after symptom onset or unexpected death in a person who lived alone and was feeling well within the previous 24hours. Death due to refractory heart failure was defined as death in a patient with decompensated heart failure who failed to respond to medical treatment, in the absence of any other cause of death. Death due to myocardial infarction was considered to have occurred when the infarction caused electrical or mechanical complications leading to early death. The remaining deaths were considered noncardiac deaths.

Finally, valvular cardiomyopathy was defined as cardiomyopathy secondary to replacement or repair of the aortic or mitral valve or severe valve disease causing ventricular dysfunction in the absence of other causes explaining the dysfunction before device implantation.8

The MADIT9 and SHOCKED10 scales were used to stratify the baseline risk of the population. The variables included in these scales and their scores are summarized in Table 1.

Gaussian continuous variables in an initial descriptive analysis are expressed as mean and standard deviation and non-Gaussian continuous variables as median (minimum and maximum). To identify differences among the study groups (valvular, ischemic, and dilated cardiomyopathy), the parametric analysis of variance test was used for Gaussian variables and the Kruskal-Wallis test was used for non-Gaussian variables. Categorical variables were compared with a chi-square test. Kaplan-Meier life table and competitive risk analyses were performed. All-cause mortality and cardiovascular mortality were calculated for each type of cardiomyopathy using Cox regression models. In these models, p-spline smoothing was used for continuous variables to determine the linearity of the variables. Various multivariate Cox models were used to identify the best explanatory model and the model with the highest plausibility or Atkinson R2 test and best c-statistic was selected. Cox regression models were used to analyze appropriate and inappropriate ICD interventions (presence or absence). The following variables were included: age, sex, MADIT and SHOCKED scales, peripheral arterial disease, diabetes mellitus, left ventricular ejection fraction, New York Heart Association functional class, chronic obstructive pulmonary disease, dyslipidemia, obstructive sleep apnea syndrome, history of atrial fibrillation, QRS width, creatinine clearance, heart rate at the time of implantation, blood urea nitrogen, bundle branch block at the time of implantation, hypertension, ischemic cardiomyopathy, incomplete revascularization, body mass index, and use of beta-blockers, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, statins, antialdosterone agents, and digoxin. P values were obtained with a 2-tailed test, and P<.05 was considered statistically significant. All statistical analyses were performed using SPSS 19 and Stata 13 (stcompet) software and the free software R (survival package).

For the univariate analysis, the category “ischemic cardiomyopathy” was considered a reference subcategory for the explanatory variable “type of cardiomyopathy”.

Of the 1174 patients included in the study, 73 had valvular dilated cardiomyopathy (10 patients also had coronary disease with consequent complete revascularization but the dysfunction was attributed to valve disease).

The baseline characteristics of the overall population and cardiomyopathy subtypes are shown in Table 2; notably, patients with valvular cardiomyopathy showed worse functional class, wider QRS, and a history of atrial fibrillation.

Significant drug treatment-related differences were found according to cardiomyopathy type. Overall, 85.9% of the patients received beta-blockers (86.5% of those with ischemic cardiomyopathy, 87.1% with dilated cardiomyopathy, and 75.3% with valvular cardiomyopathy; P=.02) and 4.8% were treated with calcium antagonists (6.2% of those with ischemic cardiomyopathy, 3.2% with dilated cardiomyopathy, and 1.4% with valvular cardiomyopathy; P=.02). Digoxin was taken by 13.9% (27.4% of those with valvular cardiomyopathy, 19.0% with dilated cardiomyopathy, and 8.9% with ischemic cardiomyopathy; P<.01). Significant differences were seen in antialdosterone use among the 3 groups (63,0% of those with valvular cardiomyopathy, 55.0% with dilated cardiomyopathy, and 45.9% with ischemic cardiomyopathy; P<.01). Acetylsalicylic acid was taken by 54.0% of patients (80.1% of those with ischemic cardiomyopathy, 16.4% with valvular cardiomyopathy, and 21.5% with dilated cardiomyopathy; P<.01). Oral anticoagulants were taken by 34.7% and their use was significantly higher in those with valvular cardiomyopathy (80.1% vs 28.6% with ischemic cardiomyopathy and 36.2% with dilated cardiomyopathy; P<.01 (Table 2).

Patients with valvular cardiomyopathy had a worse risk profile according to the MADIT (mean and standard deviation scores of the valves) and SHOCKED (idem) scales than the ischemic cardiomyopathy or dilated cardiomyopathy groups.

During a mean follow-up of 38.1±21.3 months, 197 patients (16.7%) died: 14 (19.2%) of those with valvular cardiomyopathy, 104 (15.8%) with ischemic cardiomyopathy, and 79 (17.9%) with dilated cardiomyopathy (P=.2) (Figure).

Figure.

Kaplan-Meier survival curves for all-cause events according to subgroup.

(0.08MB).

The cause of death was cardiovascular in 136 of the 197 deaths (11.6%): 9 patients (12.3%) with valvular cardiomyopathy, 69 (10.5%) with ischemic cardiomyopathy, and 58 (13.1%) with dilated cardiomyopathy (P=.1).

The crude effect of the analyzed variables on the event “all-cause mortality” is shown in Table 3.

In the adjusted analysis, the independent predictors of all-cause mortality were left ventricular ejection fraction on admission (hazard ratio [HR]=0.96; 95% confidence interval [95%CI], 0.93-0.97), age (HR=1.03; 95%CI, 1.01-1.06), serum creatinine (HR=1.66; 95%CI, 1.36-2.02), male sex (HR=2.5; 95%CI, 1.27-4.95), chronic obstructive pulmonary disease (HR=1.73; 95%CI, 1.09-2.75), digoxin therapy (HR=1.61; 95%CI, 1.02-2.53), with a c-statistic for this model of 0.71. No relevant changes were seen with a specific analysis of cardiovascular mortality, considering the left ventricular ejection fraction on admission (HR=0.96; 95%CI, 0.94-0.98), age (HR=1.02; 95%CI, 1.01-1.04), serum creatinine (HR=1.58; 95%CI, 1.30-1.91), male sex (HR=2.35; 95%CI, 1.26-4.38), obstructive pulmonary disease (HR=1.64; 95%CI, 1.08-2.48), and digoxin therapy (HR=1.84; 95%CI, 1.23-2.76). The type of cardiomyopathy was not significant in the univariate analysis or in any of the multivariate analyses performed, whether for all-cause mortality or cardiovascular mortality. Regression analysis ruled out the presence of competitive risks.

During the study period, 211 patients (18%) received an appropriate device intervention (antitachycardia pacing, shock, or both): 10 patients (13.7%) with valvular cardiomyopathy, 118 (17.9%) with ischemic cardiomyopathy, and 83 (18.7%) with dilated cardiomyopathy (P=.4). In the crude analysis, the only variable associated with a lower risk of appropriate therapy was beta-blockers (HR=0.63; 95CI%, 0.42-0.93; P=.02). However, for the type of cardiomyopathy variable, after adjustment for age, sex, and beta-blocker therapy, there was no significantly increased risk of appropriate therapy in patients with valvular or nonischemic cardiomyopathy (HR=0.76; 95CI%, 0.37-1.54; P=.4) vs those with ischemic cardiomyopathy (HR=1.17; 95CI%, 0.85-1.62; P=.3) (Table 4).

Inappropriate ICD interventions were seen in 106 patients (9%): 6 (8.2%) with valvular cardiomyopathy, 47 (7.1%) with ischemic cardiomyopathy, and 53 (12.0%) with dilated cardiomyopathy (P=.4). In the multivariate Cox analysis, the variables associated with inappropriate therapies were age < 65 years (HR=2.58; 95%CI, 1.28-5.22; P<.01), history of atrial fibrillation (HR=2.25; 95%CI, 1.23-4.11; P<.01), and cardiac resynchronization therapy (HR=0.38; 95%CI, 0.20-0.74; P<.01). No significantly higher risk of inappropriate therapy was found in patients with valvular cardiomyopathy (HR=2.36; 95%CI, 0.81-6.90; P=.12), but a significantly higher risk was seen in those with nonischemic cardiomyopathy (HR=1.99; 95%CI, 1.02-3.88; P=.01) vs the reference group of patients with ischemic cardiomyopathy after adjustment for age, sex, left ventricular ejection fraction, heart rate, hypertension, chronic obstructive pulmonary disease, and peripheral arterial disease (Table 4).

This study should be interpreted with the limitations inherent to its retrospective and nonrandomized design. The conclusions should take into account the particularly high percentage of patients in the valvular cardiomyopathy group who were also implanted with a left ventricular electrode for ventricular resynchronization. Therefore, the survival and device discharge rates were probably strongly influenced by resynchronization therapy. However, the proportion of patients with added resynchronization therapy is similar to that in the studies that helped to inform the current guidelines.

Finally, one important limitation is the lack of information on the type of valve disease (aortic, mitral), surgical intervention (valve replacement vs repair), and associated disease (eg, pulmonary hypertension, right ventricular dysfunction), which would surely affect survival and device interventions. Thus, the present work is an initial approach to the study of these patients. These limitations should be taken into account when interpreting its conclusions and by future studies aimed at optimizing the selection criteria of patients with valve diseases referred for an ICD implant.