Silent atrial fibrillation in patients with an implantable cardioverter defibrillator and coronary artery disease (INDICO AF) trial: study rationale and design

The initial hypothesis of this study is that patients with CAD and a primary indication for a single chamber ICD are at high risk for SCAF. The primary endpoint is the incidence of SCAF in single chamber ICD patients with CAD at 1 year. If this is less than 20% after 1 year, we will continue the study until 20% of patients have had a first episode of SCAF. Further secondary endpoints include the burden of SCAF, the occurrence of transient ischaemic attack (TIA), stroke or systemic embolism and the incidence of appropriate or inappropriate ICD therapy. All events will be adjudicated by an independent event committee, whose members are not investigators in the INDICO AF study.

To relate clinical characteristics of SCAF with biochemical changes, additional biomarkers will be sampled in the blood of patients included in this study. Specifically, analysis of biomarkers associated with myocardial fibrosis, inflammation, oxidative stress and heart failure will be assessed. In a prespecified sub-analysis, blood, ECG and ICD data in combination with the actual documentation of SCAF will be used to evaluate treatment decisions, and attempts will be made to construct a decision model guiding such treatment decisions.

Patients with a class IA indication for ICD therapy as primary prevention for SCD and age between 65 and 80 years may be eligible for this study. Main exclusion criteria are current AF or a history of AF and the use of vitamin K antagonists or non-vitamin K antagonists (NOACs). Complete inclusion and exclusion criteria are shown in Tab. 1.

Included patients will receive a single chamber ICD (VISIA AF™, Medtronic) with an integrated AF detection algorithm as part of standard of care ICD therapy. The VISIA AF™ is CE marked and commercially available in the Netherlands. It enables continuous monitoring for SCAF in single chamber ICD patients, by detecting RR interval based SCAF episodes of ≥6 min [6]. Episodes will be classified as SCAF at the end of the third 2‑min period of SCAF (Fig. 2b). The label of each 2‑min block is the result of a newly created Lorentz plot: a scatterplot of RR interval versus the preceding RR interval (Fig. 2a). As a standard setting, the ICD will only store the cumulative duration and the number of separate SCAF episodes. However, the ICD contains a dedicated Holter that communicates with the ICD and allows recording of the duration of individual AF episodes, thereby providing insight into the temporary distribution of AF episodes. The Holter monitor will only be enabled in patients in whom SCAF is detected during the study. All patients will be equipped with a remote monitoring system.

The occurrence of SCAF will be monitored via remote monitoring on a monthly basis as described previously [8, 9], for 1 year or until 20% of patients have had a first SCAF episode. All subjects will be followed at least until the last subject completes their 12-month follow-up visit. Patients will be seen at the outpatient clinic at 6 and 12 months for repeated blood testing and ECG.

A total of 50 patients will be included. This sample size is powered for the primary endpoint, namely the incidence of SCAF in patients with CAD and a primary indication for a single chamber ICD. As this is an exploratory analysis, and data on the incidence of SCAF in patients with CAD and a reduced LVEF are inconclusive, we powered the study on an estimated incidence of AF of 8.3% in patients with CAD and a reduced LVEF without the use of a cardiovascular implantable electronic device (CIED) to detect AF, and an expected event rate of 20% of the primary endpoint detected by a CIED in a high-risk population [5, 10, 11]. Using a one-group, one-sided χ² test, with a significance level of 0.05 alpha and 80% power, 46 patients would be required. Taking into consideration a loss of follow-up of approximately 10%, this would result in a total number of 50 subjects to be included. If the power is not achieved after 1 year, patients will be followed on an event-driven basis until 20% of patients demonstrate SCAF.

Of all strokes, 25% are of unknown cause and undiagnosed AF may be the aetiological factor [12]. Continuous monitoring of AF is superior to conventional follow-up visits or intermittent rhythm monitoring to identify AF after a cryptogenic stroke or detect AF recurrence after therapeutic interventions [9, 13]. In patients with a pacemaker, the incidence of SCAF is high and associated with an increased risk of stroke [14]. A relation between the duration of device-detected AF and stroke was demonstrated in a subanalysis of the ASSERT trial, although the event rate was lower than in patients with overt AF [15]. Not surprisingly, CIEDs with the ability to detect SCAF are of great interest to timely detect AF and prevent AF-related complications, as the population receiving ICD is generally older and has comorbidities increasing the risk for both AF and stroke. Indeed, the incidence of SCAF detected by an implantable loop recorder in patients without a history of AF but with a CHADS score ≥2 was 34% in 1 year [16].

Whether or not the implications with regard to oral anticoagulation are the same in patients with device-detected AHRE is subject of two ongoing clinical trials [17, 18]. There is no consensus either on whether AHRE detected from an atrial electrogram should be valued the same as SCAF detection based on irregularity of the RR intervals, or that stroke risk is merely driven by the duration of SCAF episodes [15]. Irrespective of whether anticoagulation should be started in all patients, establishing the diagnosis of AF is critical to determine increased stroke risk. This also applies for patients with an embolic stroke of undetermined source, as it was shown that starting rivaroxaban without documentation of AF in those patients was found to be futile compared with standard of care with aspirin [19].

CIEDs with an atrial lead, such as conventional dual and triple chamber devices, detect SCAF by recording episodes of AHRE in combination with or without the use of pre-programmed algorithms. The incidence of AHRE is high in patients with CIEDs, and CIED-detected AHRE lasting longer than 6 min is indeed associated with an increased risk of stroke [20, 21]. However, it is worth noting that not all AHRE equal atrial arrhythmias or atrial fibrillation. In addition, the high rate of SCAF in patients detected by AHRE may be overestimated as the atrial lead needed for the detection of AHRE may cause transient atrial pro-arrhythmia [22].

CIEDs without an atrial electrode such as implantable loop recorders, the VISIA AF™ and S‑ICD mostly use RR interval based algorithms for the detection of SCAF [6, 12, 23]. The AF interval-based algorithm in the VISIA AF™ ICD is based on the same RR interval detection and P wave evidence algorithms that are integrated in implantable loop recorders for the detection of AF (e. g. Reveal XT, Reveal Linq) [7, 12]. The VISIA AF™ uses an intracardiac ventricular ECG instead of a subcutaneous ECG, and episodes of SCAF are defined as the result of three 2‑min blocks of AF instead of one (Fig. 2). These extensions have resulted in an increased positive predictive value compared with strategies used in implantable loop recorders. Contrary to AF detection using atrial electrograms, the RR interval based algorithm is AF specific, and will thereby not identify atrial flutter or atrial tachycardia. High-risk patients with RR interval based implantable loop recorders show a substantial incidence of SCAF [10], but data on the clinical importance of SCAF detected by single chamber ICDs are inconclusive. The INDICO AF study will not only investigate the incidence SCAF in high-risk patients with a single chamber ICD, but will also determine the burden of SCAF. The combination of these device-detected SCAF episodes and the documentation of clinical characteristics, clinical follow-up and plasma biomarkers for fibrosis, inflammation, oxidative stress and heart failure will be introduced in a decision model that may allow preterm identification of those patients who will develop AF or AF-related complications.

Advances in CIED technology have made it possible to continuously monitor SCAF in different patient populations, leading to an increasing number of patients diagnosed with SCAF. SCAF is associated with an increased risk of stroke; however, there are no guidelines on oral anticoagulant (OAC) therapy yet as data on the benefit of OAC in this population is lacking. Ongoing trials will evaluate the value of OAC in patients with detected devices [17, 18]. A subanalysis from the ASSERT trial showed that only patients with SCAF duration longer than 24 h significantly increase the risk of stroke [15]. It is therefore debatable if—and if so when—SCAF truly represents clinical AF, and whether SCAF only mirrors the high-risk population. Future studies are needed to standardise the definition of SCAF and the considerations in therapy.