Verbesserung der Detektion von Vorhofflimmern durch die automatische Herzrhythmusanalyse SRAclinic(R) im klinischen Alltag einer deutschen Stroke Unit

Thromboembolic events due to paroxysmal and persistent atrial fibrillation (AF) are among the main reasons for ischemic stroke and associated morbidity and mortality in Germany. Therefore, it is very important to detect AF at an early stage to start oral anticoagulation as soon as possible, especially after an ischemic stroke or transient ischemic attack. There are 86 European institutions that currently (September 2018) use the automatic stroke risk analysis software SRAclinic (apoplex medical technologies GmbH) in the daily routine of a stroke unit. The automated screening algorithm searches for abnormalities concerning the heart rhythm while the patients are connected to the continuous bedside monitoring. Besides the detection of current episodes of AF (“indication of present AF”) the system indicates a prognosis concerning paroxysmal AF: “no risk for paroxysmal AF” or “risk for paroxysmal AF”. The purpose of this study is to capture and interpret the changes that occur due to the use of SRAclinic in daily routine of a German stroke unit. Therefore, the data of 1008 patients with ischemic stroke or transient ischemic attack that were treated in the stroke unit of Kreisklinikum Siegen during a 4-month period was analyzed. 501 patients were retrospectively identified before the establishment of SRAclinic (11/2012 – 03/2013). Usually, these patients received a standard admission-ECG plus one or more 24-hours Holter ECGs. Three years after SRAclinic was established on the stroke unit, 507 patients were identified and observed (11/2016 – 03/2017). In addition to the admission-ECG, they received continuous bedside monitoring with the automated screening algorithm plus optional Holter ECGs. Both cohorts show a very good comparability regarding their regular characteristics as well as the patients’ risk factors and comorbidities. There is no significant difference between the cohorts in matters of known AF before admission (18.7% vs. 20.3%; p=0.534) as well as new diagnosed AF in the admission-ECG (4.4% vs. 3.9%; p=0.723). Furthermore, these results correspond to the publications of recent studies. Nevertheless, there are main differences in the duration of ECG diagnostics. The patients of the retrospective cohort without AF before admission to stroke unit reached a median ECG duration of 21 hours and 44 minutes per patient by Holter ECGs. Using SRAclinic, these results changed significantly: The prospective patients received a median time of 36 hours of ECG diagnostics per patient, consisting of continuous bedside monitoring analyzed by the automated screening algorithm plus additional Holter ECGs (if used). The number of performed Holter ECGs decreased from 79.9% to 43.4%, while the percentage of patients without sufficient ECG diagnostics could be minimized (20.3% vs. 0.3%; p<0.001). The time period that patients spent on the stroke unit did not change between the cohorts. But even in the prospective cohort there are patients that did not receive an adequate duration of ECG diagnostics over 20 hours in total (11.5%). Furthermore, there is no improvement referring to the detection rate of AF after admission to the stroke unit and during the hospital stay using SRAclinic (3.0% vs. 2.0%; p=0.297). The recommended monitoring duration of 72 hours for patients with acute ischemic stroke or transient ischemic attack could not be reached in any of the subgroups of the retrospective or the prospective cohort. Additionally, there are difficulties in the assessment of SRAclinic since the risk analysis turned out to be not reliable to predict the appearance of paroxysmal AF (“risk for paroxysmal AF”) during daily routine. There is also a need for a more convenient structure of the reports that are generated by SRAclinic. In summary, the use of the resources was improved after the installation of SRAclinic while the duration of ECG diagnostics per patient increased significantly. But even despite the aforementioned development, there are still patients in the prospective cohort without adequate ECG monitoring, which leads to a lack of improvement concerning the detection rate of AF during the stroke unit and hospital stay. Consequently, there is a need for further changes in the process of a German stroke unit, especially in terms of ECG monitoring duration, to reproduce the results of recent studies in a daily routine. The purpose of future studies should be to improve the detection rate and therapy of AF to effectively decrease morbidity and mortality in Germany.