Stroke has consistently been the 2^nd leading cause of death in the age group of 60 plus worldwide [1]. Recently, it has become the 4^th leading cause of death and the 1^st leading cause of serious long-term disability in the USA [2]. Accurate definition of stroke etiology is crucial for the most effective management of the disease.

Cryptogenic stroke (CS), according to the TOAST classification, is defined as an ischemic stroke with undeterminable etiology, despite extensive evaluation [3]. Stroke might be cryptogenic for several reasons: 1) evaluation is not carried out at the time of event; 2) investigation is not sufficiently extensive; or 3) there are other unknown causes. Therefore the proportion of CS depends on the speed and wider-than-routine range of diagnostic evaluation. Epidemiological observational trials using standardized stroke etiological classification report 20–40 %, in some cases up to 50 % incidence of unrevealed causes [4‐6].

Cardio-embolic stroke is the most frequently predicted cause of CS in studies (58 %) [7, 8]. The most frequent cause of cerebral cardio-embolism, as well as the most common cardiac arrhythmia, is atrial fibrilation (AF). The incidence of stroke in people with non-valvular AF is estimated to be 5 times higher than in people without AF [9]. It has been suggested that paroxysmal AF (PAF) is more prevalent than chronic AF in stroke and TIA patients, and they both appear to carry very similar stroke risk [10].

Episodes of PAF may reoccur with variable frequency and their detection has been made possible by ECG monitoring. Short and usually asymptomatic presentations of PAF (also occult PAF) may remain undetected by routine arrhythmia screening methods. Incidence of newly identified PAF varies widely depending on the type of cardiac monitoring devices, the delay in monitoring after stroke onset and the monitoring duration. Detection rate of PAF with external ECG monitoring in patients after CS has been recorded between 5 and 20 % [11‐13]. An insertable cardiac monitor (ICM) automatically provides non-stop long-term ECG recording with very high sensitivity and specificity for PAF detection. Studies using ICM have suggested a detection rate of approximately 25 % [14‐16].

In our study we are focusing on a possible correlation between detection of atrial fibrillation by implantable ECG loop, and PITX2 and/or ZFHX3 gene mutations in cryptogenic stroke/TIA patients. A correlation could lead to implementation of this genomic approach to cryptogenic stroke/TIA diagnostics and management [17, 18].

Unrecognized PAF and the consequent lack of prophylactic anticoagulant treatment may result in significant morbidity first presenting as a heart failure, stroke, systemic embolism, and occasionally death. It is generally accepted that intermittent ECG monitoring (Holter ECG) has low sensitivity to identify patients with PAF. The commonly used 24-h Holter ECG monitoring has been found to reach PAF diagnostic sensitivity of 2.4–9.4 % [19‐21]. Longer continuous non-stop ECG monitoring has been shown to increase PAF detection. In a sub-study of TRENDS, a single 24-h Holter ECG diagnosed PAF in 3 %, whereas 7, 21, or 30 days of monitoring diagnosed PAF in 6, 9, or 11 %, respectively [22]. The EMBRACE study (The 30-Day Cardiac Event Monitor Belt for Recording Atrial Fibrillation After a Cerebral Ischemic Event) used noninvasive ambulatory ECG monitoring in patients with cryptogenic stroke or TIA for 30 days. Detection of atrial fibrillation improved fivefold and indication of anticoagulant treatment nearly doubled when compared with standard practice of short-duration ECG monitoring. AF was detected in 16.1 % of patients in the intervention group, compared with 3.2 % in the control group [23]. ICM devices for long-term ECG monitoring have been used primarily for differential diagnosis of syncope [24]. Cotter et al. reported “Reveal XT” PAF identification in 25.5 % (13 patients) of 51 patients with unexplained stroke [14]. Sensitivity and specificity of “Reveal XT” device for PAF was detected to be 96.1 and 85.4 % respectively (3-year non-stop monitoring) [25]. A large randomized clinical trial The Cryptogenic Stroke and Underlying Atrial Fibrillation trial (CRYSTAL AF) compared long-term ECG monitoring by ICM with conventional ECG Holter monitoring in patients with recent CS. After 12 months ICM monitoring detected over 7 times more patients with PAF than ECG Holter monitoring. Three-year follow-up detected 30 % PAF occurrence in the ICM group compared to 3 % in the Holter group [26]. Choe evaluated sensitivity and negative predictive value (NPV) of various external monitoring techniques in a cryptogenic stroke cohort. Simulated intermittent monitoring strategies were compared with continuous rhythm monitoring in 168 ICM patients of the CRYSTAL AF trial. Short-term monitoring included a single 24-h, 48-h and 7-day Holter, and 21-day and 30-day event recorders. Periodic monitoring consisted of monthly monitoring by 24-h Holters and quarterly monitoring by 24-h, 48-h and 7-day Holters. For a single monitoring period, sensitivity for AF diagnosis was lowest with 24-h Holters (1.3 %) and highest with 30-day event recorders (22.8 %). NPV ranged from 82.3 to 85.6 % for all single external monitoring strategies. Quarterly monitoring with 24-h Holters had 3.1 % sensitivity, whereas quarterly 7-day monitors increased it to 20.8 %. NPV of various types of repetitive periodic monitoring were similar at 82.6 to 85.3 %. Long-term continuous monitoring was superior in detecting AF compared to all types of intermittent monitoring evaluated (p <0.001) [27]. Ritter et al. presented 17.0 % PAF detection in ICM group in comparison with 1.7 % in 7-day ECG Holter group (p = 0.0077) [28].

Finding a new algorithm for detecting PAF in standard clinical practice should be of great interest as it would allow early indication of anticoagulant therapy in stroke/TIA patients. Additional information that would help to characterize stroke subtype could be provided by genetic data. Study of gene polymorphisms associated with an increased risk of AF suggests that subtle alterations in developmental factors, cell signalling, extracellular matrix regulation, and ion channels function are involved in arrhythmia pathogenesis. Large genome-wide association studies (GWAS) have found that the genetic markers close to the PITX2 gene on chromosome 4q25 and to the ZFHX3 gene on 16q22 are associated with both AF and cardio-embolic stroke [17, 18]. The PITX2 gene plays a critical role in development of left atrium and pulmonary vein myocardium. This particular region has been implicated to initiate and maintain AF. The expression patterns of the ZFHX3 gene in human cardiac and pulmonary tissue are not clear, nor are the mechanisms by which variants in this gene may predispose to AF.