Skip to main content

Comparative Effectiveness Research to Assess the Use of Primary ProphylacTic Implantable Cardioverter Defibrillators in EUrope

Final Report Summary - EU-CERT-ICD (Comparative Effectiveness Research to Assess the Use of Primary ProphylacTic Implantable Cardioverter Defibrillators in EUrope)

Executive Summary:
Patients for the landmark clinical trials that established the clinical treatment guidelines for primary prophylactic implantable cardioverter defibrillators (ICDs) were recruited 15-20 years ago. Over this long timespan, cardiovascular treatment significantly improved while ICD treatment guidelines remained unchanged. The randomized DANISH ICD study published in 2016 exemplified the need for overhaul of the guidelines. The European collaborative project EU-CERT-ICD (Comparative Effectiveness Research to Assess the Use of Primary ProphylacTic Implantable Cardioverter Defibrillators in Europe) aimed to analyze the effectiveness of prophylactic implantation of cardioverter defibrillators (ICDs). The project featured a non-randomized controlled cohort study with two treatment groups in candidate patients for primary prophylactic ICD therapy. A wide-ranging collection of baseline clinical characteristics and advanced electrocardiographic non-invasive diagnostics was collected to be combined in multivariate models and risk scores. These are the necessary tools in order to individually predict primary outcomes such as all-cause mortality and the occurrence of appropriate ICD therapy, and utilize this information for indication of the prophylactic ICD. Identification of patients with particularly high or low risk of death from all causes, and/or particularly high or low risk of appropriate ICD therapy is possible. Assessing the risk of these two primary outcomes allows to estimate the clinical benefit of the ICD. In addition, retrospective ICD and ECG data were collected in 14 European centers and up to 5.300 patients for comparative analyses of prophylactic ICD treatment. A meta-analysis of more recently published studies on primary prophylactic ICD treatment included data from more than 46.000 patients. From all three data sources (prospective study, retrospective registry, and meta-analysis) we estimated quality of life (QoL)-adjusted cost-effectiveness of prophylactic ICD treatment. Attention was given to individual sub-groups, including sex comparisons and regional differences. EU-CERT-ICD publications are expected to be relevant for decisions in primary prophylactic ICD treatment.
Project Context and Objectives:
The European collaborative project EU-CERT-ICD (Comparative Effectiveness Research to Assess the Use of Primary ProphylacTic Implantable Cardioverter Defibrillators in Europe) aimed to analyze the effectiveness of prophylactic implantation of cardioverter defibrillators (ICDs). Annually, more than 100.000 ICDs are implanted in the EU-28 with huge disparities between countries (see Raatikainen et al, Europace 2015). This reflects a billion € - market and considerable socio-economic and health-care inequalities across the EU.
The project featured a non-randomized controlled cohort study with two treatment groups in candidate patients for primary prophylactic ICD therapy. A wide-ranging collection of baseline clinical characteristics and advanced electrocardiographic non-invasive diagnostics was collected to be combined in multivariate models and risk scores. These are the necessary tools in order to individually predict primary outcomes such as all-cause mortality and the occurrence of appropriate ICD therapy, and utilize this information for indication of the prophylactic ICD. Identification of patients with particularly high or low risk of death from all causes, and/or particularly high or low risk of appropriate ICD therapy is possible. Assessing the risk of these two primary outcomes allows to estimate the clinical benefit of the ICD.

The objectives of the project were prospectively defined and remained unchanged to the end of the project, these are:
• To compare mortality rates between patients eligible, according to guideline-derived criteria, for primary prevention of sudden cardiac death by ICD implantation who do receive an ICD, and patients fulfilling these criteria who do not receive an ICD implantation.
• To identify important patient subgroups with particularly high or low risk of the major outcomes all-cause mortality, appropriate and inappropriate ICD shock, by using dedicated diagnostic risk methods as well as primary clinical characteristics.
• To determine socio-economic outcomes of primary preventive ICD implantation, including QoL, costs, and QALY-adjusted cost-effectiveness in the target population featuring sub-groups and regional comparisons within Europe.
• To determine any gender-related effects on the above.
• To validate or challenge current guideline indications for primary prophylactic ICD treatment.
• To establish a biobank for the purpose of future risk stratification.
Project Results:
In summary, EU-CERT-ICD was successfully completed in five years since the project start and has produced important results from the meta-analyses, registries, and more importantly its prospective study.

The Prospective Study WP01 was conducted under the leadership of UMG and networked to interact with many of the work packages (WP03 through WP12, see Figure 1). In total, 2327 patients were enrolled between May 2014 and August 2018 in 27 partner sites and 17 third party centers in 15 European countries. Two non-randomized treatment groups were formed with target sample sizes as per the initial calculation: at least 1500 analyzable ICD and 750 analyzable control. A follow-up for the pre-defined outcome events such as all-cause mortality and first appropriate ICD shock was available for up to 52 months. With regard to all-cause mortality, the number of the events exceeded the pre-calculations while the number of first appropriate shocks was below but still within the range of expectations. Propensity scoring and multivariate models for correction were used to compare the the two treatment groups ICD and non-ICD for all-cause mortality. The overall cohort had an average age of 63 years, with very small differences between ICD and control group.

The Retrospective Registry WP02 was led by UHBS and compiled data from more than 5300 patients and 14 centers. The registry presented novel scientific results starting from 2015 (presented at ESC congresses), as well as a full-size manuscript in Europace in 2017/2018 (Sticherling et al). The main finding of the manuscript was that women experience only about half the number of appropriate shocks, even after multivariate correction. Using this powerful data source, it will be statistically possible to further refine factors predicting mortality and shock risk, and assess overall event rates.
WP03 Health economics and Quality of life (QoL) collected data from the prospective study with good levels of quality and completeness (all participating countries: information on quality of life, Germany and Switzerland: resource consumption and costs). For the present report, all WP03 deliverables from the prospective study were completed. The results indicate that patients undergoing prophylactic ICD implantation had a neutral effect in their quality of life. The device was well accepted, with differences between participating countries. The implantation leads to significantly higher costs (from health payers’ as well as societies’ perspective). Cost-effectiveness of ICD implantation was assessed in the overall cohort, further analyses may reveal patient subgroups with improved or reduced cost-effectiveness. A systematic review on QoL in ICD patients was published. Further publications are planned.
WP04 Gender issues completed two gender-related analyses as original papers, a single center analysis of 1151 ICD patients (Seegers et al, Europace 2016), and a gender-specific meta-analysis of ICD studies (Conen et al, PLoS One 2016). The results confirmed that women receive 50% fewer appropriate shocks than men, but have a similar or reduced mortality risk (Sticherling et al, Europace 2018). In the prospective study, we found female patients to have 43% fewer appropriate shocks after multivariate correction.
WP05 Beat-to-beat variability had organized an open ECG conference with an emerging QT variability consensus paper (Baumert et al., Europace 2016). Several publications on the pathophysiologic value of the STV QT parameter followed. In the prospective study, STV QT at baseline was tested as a predictor of first appropriate shock or mortality in both treatment groups.
WP06 T-wave morphology compiled an ECG database of WP02 patients from which 2041 digital ECGs were available. Analyses for all-cause mortality showed that several automatically analyzable 12-lead-ECG parameters were predictive of mortality. The data were presented at the General Assembly and are currently prepared for publication. Furthermore, WP06 evaluated morphological indices and developed, tested analysis software packages and secured a common recording platform for the consortium. In the WP01 cohort, the same parameters were again used for prediction of all-cause mortality and first appropriate shock.
WP07 Fractionation and early repolarization analyzed exemplary sets of WP01 prospective study patients and the complete database of digital ECGs from WP02. Results from the latter showed that a number of the measured variables were highly predictive for endpoints. In the WP01 cohort, all intended ECG parameters were available for correlation with clinical outcomes mortality and appropriate shocks.
WP08 Heart rate variability and heart rate turbulence analyzed all incoming Holter ECGs from WP01 with high quality. All intended parameters were correlated with outcomes and several parameters attained significant results for prediction of both endpoints, all-cause mortality and first appropriate shock.
WP09 Biobanking was successfully storing biosamples for more than 94% of enrolled WP01 patients in the EU-CERT-ICD biobank, while it was eventually deemed impossible to collect samples from surviving WP02 patients. The use of the biosamples beyond the funding period of the project was organized.
WP10 Statistics and meta-analyses published two meta-analyses, one on gender as prognostic factor in cooperation with WP04 (Conen et al, PLOS One 2016) and one on other prognostic factors (Bergau et al, PLOS One 2017). WP10 was also in charge of all other statistical calculations in the consortium. WP10 supported the analyses of the registries which led to a publication on gender in cooperation with WP04 (Sticherling et al, Europace 2018). With outcome data available from WP01 at the end of the project, meaningful analyses were conducted from WP01, WP04, WP05, WP06, WP07, and WP08. Extensive analyses were done to compare the non-randomized treatment groups ICD and control.
WP11 Dissemination provided a dissemination and exploitation plan and oversaw publications and other dissemination activities. The number of publications has increased significantly over the last year. Project logo and project website were created and updated.
WP12 Prospective study management supported the prospective study with eCRFs, study database, regulatory affairs, central monitoring and on-site monitoring. WP12 was in charge of negotiation and preparation of all clinical center contracts as well as third-party contracts.
WP13 Project management was taken over by the UMG Research & EU Office with effective date 01/07/2016. They performed all consortium-wide management tasks, administrative project controlling, organized consortium finances, telephone conferences and face-to-face meetings.

Potential Impact:
With the results of the DANISH ICD study being discussed since its publication 2016, the need for additional data is even more evident. The EU-CERT-ICD project will provide these data with large sample sizes. A similar but smaller Dutch project (DO-IT, n=1500 patients) is also underway which we are already cooperating with. The observational study design of EU-CERT-ICD was correctly chosen. Carefully designed randomized ICD studies have recently become feasible, such studies should be planned with the background of EU-CERT-ICD data.

To date, the DANISH study did not change international guidelines while an update of the ESC guidelines for the diagnosis and treatment of acute and chronic heart failure is still pending. Only two randomized ICD studies are currently underway. The first is the US I-70 study funded by the Veterans Administration (PI: Singh-SN), the study cleared a pilot phase, enrollment of 1000 patients (age > 70, ICD vs. control) is planned until 2019, results will become available in mid-2020. The second is the German RESET-CRT study funded by German health insurers (PI: Hindricks-G), the study has started in 2018, enrollment of 2000 patients (ICM+DCM, CRT-D vs. CRT-P) is planned until 2020, results are planned to become available in 2021.

With the results of the aforementioned randomized studies becoming available up to 4 years later, the results of the EU-CERT-ICD study in 2300 patients (including unrandomized controls) become important data upon publication, and they can be expected to be relevant for European treatment guidelines.

Our study analyzed a possible ICD survival benefit in the overall cohort of ICM and DCM patients. In our prospective study design, important sub-cohorts were defined for which the average ICD benefit would be expected to be higher (strengthening the individual indication) or lower (eventually losing the benefit). Candidate subcohorts were elderly patients ≥ 75 years of age and patients with a high mortality rate above 15% per year due to comorbidities. Specific prediction of arrhythmias and appropriate shocks was attempted with multiple techniques and parameters and expected to be difficult. Several ECG parameters accessible from the digital 12-lead-ECG were cancidates to predict appropriate shocks (or mortality) well.

In a post-hoc publication of the DANISH study on interacting patient characteristics, they demonstrated ICD benefit for younger patients, but not for patients older than 70 years. With our prospective data, we attempt to confirm the absent ICD benefit in elderly patients ≥75 years. The EU-CERT-ICD data also give wide access to other clinical factors and ECG methods that can successfully substratify the expected patient risks.

The decision to implant prophylactic ICDs should be done more individualized and the data and results of the EU-CERT-ICD retrospective and prospective studies can provide guidance. This kind of stratification had been predefined by the EU-CERT-ICD project to improve patient selection.

The socio-economic dimensions can be determined for the EU as a whole, as well as for each individual country. The final results of WP03 demonstrate that in the overall cohort, the ICD is also cost-efficient. The effect of the ICD on quality of life is clearly neutral. When the ICD benefit is lost in a well-defined subgroup, an economical analysis is not indicated. For an assessment of intermediate risk groups and variation from country to country, we need further health economical analyses. Eventually, our health economics WP will be able to translate the clinical ICD benefit in important sub-groups into variations of cost-effectiveness and QoL-adjusted cost-effectiveness. We expect reluctance from industry to development of stratified guidelines. The EU-CERT-ICD consortium wants to assist allocating ICD implants to patients with predictable ICD benefit, and equally in all countries. Implantations in patients who do not benefit from the device should be avoided. This would mean implanting the same number of devices in better selected patients.

The established EU-CERT-ICD biobanking module has established a biobank of approximately 2300 samples for future arrhythmia risk assessments of individual patients based on genetic analyses. To regulate the use of these samples after the end of the EU-CERT-ICD project, a Use and Access Committee has been formed.
List of Websites:
http://www.eu-cert-icd.eu/
Contact:
Prof. Markus Zabel, UMG Göttingen
M.Sc. Rajeeva Sritharan, UMG Göttingen, 0551-39-10475