ABC | Volume 111, Nº2, August 2018

Original Article Sampaio et al Monitoring of rhythm, atrial fibrillation and stroke Arq Bras Cardiol. 2018; 111(2):122-131 per patient (Table 2). Also, we found that after the repair of transmission towers and antennas, signal reception was changed from 2.5G (GPRS General Packet Radio Service) to 3G, which negatively affects data transmission. Updating of the technology from 3G to 4G would resolve this issue, as well as reduce the energy expenditure with data package transmission, resulting in optimization of rechargeable battery duration, reduction of charging time and improving monitoring performance. Greater data loss due to artifacts was seen in control subjects in the PoIP group, whichmay be justified by the greater freedom of movement of patients in ambulatory treatment. Arrhythmias detected by PoIP (firs 24 hours) compared with Holter-24 In the first 24 hours, no difference in arrhythmias was observed (AT, SVES, SVES + AT). Despite the longer monitoring period by Holter recordings, all AT runs and the three episodes of AF (2 in the stroke and 1 in the control group). Twenty-four hour Holter compared with prolonged monitoring Comparison between Holter and PoIP monitoring results showed a higher proportion of frequent AT and SVES detected by PoIP monitoring in both stroke/TIA and control groups, which was expected by its longer monitoring period. Comparison of arrhythmias detected in stroke group and controls No significant difference was found in the occurrence of AT or nonsustained AF, in the comparison between patients with cryptogenic stroke and a control group matched by sex, age and corrected CHADS 2 . We report a high prevalence of atrial arrhythmias in 52 patients, including 40 with AT and 7 with AF. In stroke/TIA group, proportion of AF was 23.1% in patients monitored by PoIP, and 3.8% in those monitored by Holter, which is in agreement with the literature (Tables 3, 4 and 5). 20 Some studies have suggested that and additional 24-hour period of monitoring would increase the percentage of new diagnoses of paroxysmal AF in 2-4% stroke patients. 21,22 This confirms the efficacy of prolonged ambulatory ECG in patients at risk of AF and may generate a clinically significant diagnostic yield. 23 Studies have highlighted the association of frequent SVES and AT with increased risk of stroke. 2,3,4,24-27 Studies involving long-term heart rhythm monitoring in patients with previous stroke/TIA have reported a paroxysmal AF prevalence of 5-20%. 20,28,30-33 In our study, all AF episodes lasted less than 30 seconds. Although an AF episode ≥ 30 seconds is used as a parameter for the diagnosis of AF, 7 some authors have suggested that short AF episodes have an impact on the risk of stroke/TIA or systemic thromboembolism. 10,33 One important finding was the lack of difference in the prevalence of atrial arrhythmias between patients with and without stroke or TIA, at similar risk for these conditions. This finding suggests that the atrial arrhythmias detected may be an epiphenomenon. Kottkamp and other authors 15,34 have suggested the presence of a thrombogenic fibrotic atrial cardiomyopathy, with risk for embolic events with no causal connections with atrial arrhythmias. Contractile changes would be responsible for the increased thrombogenic risk during sinus rhythm, in addition to interatrial block and sinus node dysfunction. Even ablation of AF would not be able to impede the progression of fibrotic process. 34 Factors like diabetes, hypertension, age, among others, would be involved in myocardial damage. In our sample, more than 80% of patients had arterial hypertension and more than 50% were diabetic. Non-invasive detection of atrial fibrosis is currently limited to MRI techniques, not available in clinical practice. 34 In this context, AF would be a manifestation of atrial structural changes, and thereby increasing the risk of embolic events. None of our patients with stroke/TIA had AF before or during stroke. In fact, AF may be detected in only a minority of the cases and may take months, as shown by the TRANDS, ASSERT and IMPACT studies, which included patients with implantable continuous monitoring devices. 35-37 The paradigm used in most studies is that AF detection would be just a matter of time, but even in a one-year follow up, AF is detected in less than half of patients with cryptogenic stroke. This is a pioneering study in monitoring patients at similar stroke and TIA risk, by including a group with stroke and a control group without the disease. The finding that the incidence of atrial arrhythmias was not different between both groups is consistent with the hypothesis that a factor other than arrhythmia may be involved in the risk for stroke; one possibility is fibrotic atrial cardiomyopathy. Study limitations The sample size was insufficient to evaluate individual risk factors. Discrimination between short runs of atrial tachycardia and AF may be difficult, even to an experienced electrophysiologist. P-waves in ambulatory monitoring systems may not be clearly identified as compared with conventional 12-lead ECG. Nevertheless, analysis of isolated episodes and analysis of more than one arrhythmia episode yielded similar results, since all patients that had short AF episodes also had AT. Mobile phone services currently available still have limited coverage, with absent or deficient signal strength, and unstable transmission velocity, which altogether, negatively affect PoIP data collection. Due to frequent repairs of problems caused by electrical discharges in cell phone towers, access to GPRS may be lost, thereby affecting signal reception, which may be solved by implementation of the 4G technology. Conclusions Holter and PoIP showed comparable results in the first 24 hours. The shorter monitoring period was caused by a low signal strength. Data transmission loss in hospitalized patients resulted from a mismatch between the protocol of signal transmission in the cell phone tower (3G) and the signal effectively transmitted (2.5G), which can be mitigated by the adoption of a 4G technology. The incidence of arrythmia was not different between stroke and control groups. 129