Heart failure with reduced ejection fraction is associated with an increased risk of sudden cardiac death (SCD). Implantable cardioverter-defibrillators (ICD) have reduced mortality and become the standard of care for these patients, as recommended by current international guidelines   . In a pooled analysis of 10 different studies on primary preventive defibrillators in patients with heart failure, all-cause mortality with ICD therapy was reduced about by 7.9% compared with optimal medical treatment only  . All ICD patients are at an increased risk of perioperative complications, inappropriate shocks, and secondary arrhythmias  . The reported complication rates vary in different studies, but in a recent review of 11 well-known randomized ICD studies, the overall lead dislodgement rate was 1.8%  . A much higher complication rate was reported in a real-life survey of 440 ICD patients in Germany, where 31% of the patients experienced some type of complication  .
Previous studies have tried to identify predictors of appropriate ICD therapies, including atrial fibrillation (AF)   , renal dysfunction  , advanced New York Heart Association (NYHA) class  , age  , gender   and parameters of electrical instability or arrhythmic substrate  . However, all those trials included different patient populations, so it is very difficult to draw conclusions for clinical decision. The aim of this study was to investigate, in real- life, the net benefit of primary preventive defibrillators as regards the incidence of appropriate ICD therapies, inappropriate shocks, complications, and identify clinical parameters and comorbidities associated with appropriated ICD therapies for a better selection of ICD candidates.
2. Materials and Methods
We scrutinized 875 consecutive patients undergoing ICD implantation from June 2002 to May 2015 in Cardiology Department (Novara); we excluded patients with a history of previous sustained ventricular arrhythmias and cardiac arrest (209 patients), arrhytmogenic syndrome as Brugada and Long QT syndromes, arrhythmogenic right ventricular dysplasia and hypertrophic cardiomyopathy (53 patients). We enrolled 613 consecutive patients receiving an ICD for primary prevention of SCD; all patients underwent ICD implantation according to current international guidelines   , defined as LVEF ≤ 35%, NYHA class II-IV, optimal medical therapy, life expectancy > 1-year; cardiac resynchronization therapy was performed if indicated. We evaluated in ischemic and non ischemic cardiomyopathy appropriate and inappropriate ICD therapies (Figure 1). All patients underwent before the ICD implantation 12-leads elec-
Figure 1. Study design and appropriate therapy in ischemic and non ischemic cardiomyopathy.
trocardiography, transthoracic echocardiography, laboratory test and evaluation of quality of life and NYHA class.
Transthoracic echocardiography data were acquired using the available system (Vivid-7 and Vivid-9 General Electric Vingmed, Milwaukee, USA). Left ventricular end-diastolic and end-systolic volumes were assessed from the apical two- and four-chamber images, and left ventricle ejection fraction (LV-EF) was calculated according to the monoplane or biplane Simpson method; diastolic function, left ventricle and left atrial M-mode parameters, color-Doppler parameter were evaluated according to guidelines    .
Quality of life and NYHA class were evaluated by questionnaire (Kansas  and Minnesota living with heart failure questionnaire  ) and functional test (6 minutes walking test  ).
Continuous variables are reported as mean ± standard deviation and categorical variables as frequencies. Continuous variables were compared using Student’s t-test and ANOVA, while categorical variables were compared using χ2 test and Fisher’s exact test. To determine the predictive factors of device therapy, univariate and multivariate analyses were performed. Variables with a P-value of ≤0.1 in univariate analysis were then included in the multivariate regression analysis for the determination of odds ratio (OR) and its 95% confidence interval (CI). A P-value of ≤0.05 was considered statistically significant. Survival and event-free rates from ICD intervention were calculated and depicted with the Kaplan-Meier method. All these analyses were performed using SPSS v20.0 (SPSS Inc., Chicago, IL, USA). Logistic regression, using Akaike Information Criterion  , and ROC curves were performed using R 3.3.0 (R Development Core Team).
Patient data at baseline are summarized in Table 1. Patients had a mean age of 67.4 ± 10.1 years, 80.9% male, 19.1% female, with LV-EF of 26% ± 8.4%, congestive heart failure of NYHA II (52.2%), due to ischemic aetiology (n = 360, 58.7%), valvular aetiology (n = 57, 9.3%) or idiopathic cardiomyopathy (n = 196, 32%); atrial fibrillation (n = 271, 44.2%) and cardiac resynchronization therapy (n = 300, 48.9%) were common. Mean follow up was 50.7 ± 31.5 months, 41 (6.7%) patients were lost at follow up and 4 (0.7%) underwent heart transplantation. There were no significant differences in comorbidities between male and female.
During follow-up 198 patients (32%) received appropriate ICD therapy, 93 patients (15%) had inappropriate shocks, 53 patients (8%) had at least one complication (electrode dysfunction, infection and pocket related) and 191 patients (33%) died. The annual rate of patients who received their first appropriate ICD interventions was 7.5%. The median time to first appropriate therapy after ICD implantation was 24 months (range 1 - 107). Differences in baseline characteristic between patients with and without appropriate ICD therapy are listed in Table 2.
In patients who received appropriate therapy were present significantly lower level of sodium (137.8 ± 3.7ml vs 138.9 ± 3.5 ml; p < 0.01) and atrial fibrillation (n = 107, 54% vs n = 164, 39.5%; p = 0.001). Indeed, among echo parameters, they had significantly reduced ejection fraction (25% ± 6.5 vs 26.5 ± 9.1; p =
Table 1. Baseline characteristics before implantation.
Table 2. Difference in baseline characteristic between patients with and without appropriate ICD therapy.
0.035), higher LV end systolic volume and diameter (respectively 147.4 ± 50.2 mm vs 138.8 ± 51.7 mm; p = 0.051 and 56.1 ± 9.5 mm vs 54.1 ± 9.0mm; p = 0.015) and higher left atrial diameter (48.7 ± 8.0 mm vs 46.9 ± 7.5 mm; p = 0.007).
A stepwise multivariate analysis was performed and 3 significant factors were correlated to appropriate therapy: atrial fibrillation (OR = 1.8, CI = 1.27 - 2.53; p < 0.01), diabetes (OR = 1.8, CI = 1.27 - 2.53; p = 0.041) and vasculopathy (OR = 1.8, CI = 1.27 - 2.53; p = 0.031).
Finally, we tried to identify a predictive score of appropriate therapy based on clinical parameters. We used logistic regression, including all variables as covariates and we screened all possible models including almost one of these. We found 220 models and to determine the best we used Akaike Information Criterion. The best model considered seven parameters that were strongly associated with appropriate therapy (atrial fibrillation, diabetes, vasculopathy, LVEF, NYHA class, left atrial diameter and natremia) and identified a difference in SCD risk probability statistically significant between patients with (APP+) and without (APP-) appropriate therapy. Probability < 0.1872 corresponded to 5th percentile of group receiving appropriate therapy; using this cut-off we identified a population at low risk of ICD treatment (only 10/102 patients; of these 10 patient 8 received only one ATP therapy and 2 patients received only one shock on sustained TV). Probability < 0.1258 corresponded to lower value of group receiving appropriate therapy; using this cut-off we identified patients who never received appropriate therapy (65 patients) (Figure 2).
The total number of patients who had complications, including inappropriate shock, during follow-up was 138 (22.5%). Of these, 93 patients (15.2%) received inappropriate shock, in most cases on sopraventricular tachiarrhythmia (prevalent atrial fibrillation); 53 patients (8.6%) had complications that required reintervention (electrode dysfunction, infection and pocket related). The types of complications are listed in Table 3; in summary the most common problems were inappropriate shocks (n = 93, 15.2%) and dislocation or dysfunction of the ICD electrode (n = 33, 5.4%).
Forty one of 613 patients (6.7%) were lost during follow up; 191 of 572 patients (33.4%) died during follow-up. 73 of 191 patients (38.2%) received an appropriate therapy on a ventricular arrhythmia. The median time from first appropriate therapy to death was 20 months (range 1 - 105).
Cardiovascular mortality was observed in 68 patients (11.9%), 87 patients (15.2%) died for non cardiovascular reasons and we had no data about 36 patients (6.3%). All causes mortality are presented in Figure 3.
Figure 2. SCD risk in patients with (APP+) and without (APP−) appropriate therapy.
Figure 3. All causes mortality.
Table 3. Device-related complication including inappropriate shocks.
Our study in a real-life cohort of primary prophylactic ICD-treated patients demonstrates that 8% of our patients annually receive presumably life-saving treatment for ventricular arrhythmias. This is almost the same proportion of patients as seen in the MADIT II  , DEFINITE  , and SCD-Heft  studies. The follow-up time in our study was also relatively long and complication rates increased over time. In our population, 22.5% of patients population had complications, including 15.2% of patients that received inappropriate therapies, principally caused by supraventricular arrhythmia. The main reason for inappropriate therapy in our study, as in many others    , was atrial fibrillation. Better knowledge and more ‘conservative’ programming (i.e. higher VT zones with longer detection intervals and more ATP therapy attempts before shock therapy) may have contributed to the reduced number of unnecessary shocks  . However, no complications in our study were lethal and there was no association between complications and increased mortality.
This study demonstrates that baseline characteristics of ICD recipients can be useful for identifying patients at high risk for appropriate therapies and that these parameters are different according to their type of cardiomyopathy. Atrial fibrillation, NYHA class III-IV, comorbidities such as vasculopathy and diabetes, hyponatremia, may be interpreted as an epiphenomenon or as a surrogate of advanced heart failure that leads to more ICD therapies. Other previously described predictors for ICD therapies were not confirmed by our findings, as renal dysfunction  , age  and gender   . This could be explained by the fact that we included a homogenous population and optimal medical therapy prior to reference in our tertiary centre. Renal dysfunction, obesity and advanced age were not as prevalent in our study population as in other studies and thus failed to reach a significant predictive value.
Our logistic regression using seven clinical parameters-atrial fibrillation, diabetes, vasculopathy, NYHA class, left atrial diameter, ejection fraction, natremia?identified a population with a higher risk of appropriate therapy.
Using a cut off of probability < 0.1872 we identified a population at low SCD risk: only 9.8% patient had appropriate ICD therapy during the follow-up (about 2.45% per year) and the ICD treatment was only on sustained VT (in most cases only one ATP was necessary). Excluding these population we could have avoided ICD implantation in about 15% that is a significant number on a large scale. Only few patients under this threshold received appropriate therapy, but only on one episode of VT.
If we considered a less tighten cut off including all patient that received appropriate therapy, we were sure not to lose any patient because of SCD: a probability <0.1258 identifies a subgroup at a very low risk (without any ICD therapy) and so we should potentially reduce implantation of 10%. It is an easy score because it is based on clinical features that are routinely evaluated before implantation. Obviously, before extended clinical implementation, it could be important to test this score on a larger population to confirm accuracy and reproducibility. These findings reveal the most important predictors of ICD therapies in a real- world experience and are valuable for their ease use in clinical practice.
This is not a randomized clinical trial but a retrospective study based on the detailed analysis of medical records from patients with primary preventive ICD treatment in a long period of follow up. However, during the last 10 years, there were not so many change in primary prevention indications and pharmacological treatment. This is a single-centre study with the limitations of a post hoc analysis. However, we included consecutive enrolled patients with regular and thorough device interrogations to obtain a comprehensive dataset.
In our real-life population of primary prevention ICD patients ventricular arrhythmias necessitating ICD therapy are common (8% annually), but complications are frequent (4% annually). Comorbidities, echo parameters, laboratory tests and functional class should be taken into consideration for a better selection of population candidate to ICD therapy, to reduce ineffective implantation. Our proposed score may eventually reduce from 10% to 15% future ICD implantation.