European Journal of Preventive Cardiology http://cpr.sagepub.com/
Electrocardiographic monitoring during marathon running: a proof of feasibility for a new telemedical approach Sebastian Spethmann, Sandra Prescher, Henryk Dreger, Herbert Nettlau, Gert Baumann, Fabian Knebel and Friedrich Koehler European Journal of Preventive Cardiology 2014 21: 32 DOI: 10.1177/2047487314553736 The online version of this article can be found at: http://cpr.sagepub.com/content/21/2_suppl/32
Published by: http://www.sagepublications.com
On behalf of: European Society of Cardiology
European Association for Cardiovascular Prevention and Rehabilitation
Additional services and information for European Journal of Preventive Cardiology can be found at: Email Alerts: http://cpr.sagepub.com/cgi/alerts Subscriptions: http://cpr.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Oct 29, 2014 What is This?
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
EURO PEAN SO CIETY O F CARDIOLOGY ®
Original scientific paper
Electrocardiographic monitoring during marathon running: a proof of feasibility for a new telemedical approach
European Journal of Preventive Cardiology 2014, Vol. 21(2S) 32–37 ! The European Society of Cardiology 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/2047487314553736 ejpc.sagepub.com
Sebastian Spethmann1,2, Sandra Prescher1, Henryk Dreger1, Herbert Nettlau3, Gert Baumann1, Fabian Knebel1 and Friedrich Koehler1,3
Abstract Objective: There is a risk for sudden cardiac death and nonfatal arrhythmias for marathon runners. A new telemedical approach to prevent sudden cardiac death could be online electrocardiogram monitoring during endurance sport events, which would allow the emergency services located along the running track to initiate instantaneous therapy. In a first proof-of-concept study we evaluate the feasibility of recording, transferring and analysing an electrocardiogram via a mobile phone (electrocardiogram streaming) and compare the quality to a conventional Holter electrocardiogram during marathon running. Methods: A total of 10 recreational endurance runners are equipped with a Holter Tele-electrocardiogram and a standard smartphone connected via BluetoothÕ to each other in order to continuously record an electrocardiogram during a first marathon event (five runners) and a second marathon event (five runners). All streaming electrocardiogram data were transferred from the device to our telemedicine centre (Charite´ Campus Mitte, Berlin, Germany); the data were monitored live and stored for a subsequent offline analysis. The primary endpoint was the percentage of successful transfer time of the streaming electrocardiogram compared with Holter electrocardiogram; the secondary endpoint was the percentage of correctly identified arrhythmias in the observed period. Results: It is technically feasible to stream an electrocardiogram during marathon running in the presence of thousands of mobile phone users. In addition, the identification of arrhythmias during a marathon is possible by electrocardiogram streaming. However, during the first race, the data transfer quality was low. After improvement of the software, in the subsequent race there was an extremely good quality in the data transfer via the mobile phone network (89%) and 100% of the rhythm disturbances could be detected in the streamed electrocardiogram. Conclusion: Online electrocardiogram surveillance during marathon running is a promising preventive concept. Intensive further technical development is needed first before further studies with clinical endpoints can start.
Keywords Marathon running, electrocardiogram streaming, telemedicine, arrhythmias, sudden cardiac death Accepted 11 September 2014
Introduction Sudden cardiac death (SCD) in marathon runs is a rare but highly emotional event.1 In Europe, the main cause of SCD in athletes is coronary artery disease (CAD) occurring mainly in middle-aged men.2 Importantly, only a minority of athletes with SCD had symptoms within 2 weeks before the race.3 Marathon running is gaining popularity, even among people over 45 years old. Many of them begin competitive endurance sports in their 40s. A ‘health
1 Medizinische Klinik fu¨r Kardiologie und Angiologie, Campus Mitte, Charite´ – Universita¨tsmedizin Berlin, Berlin, Germany 2 Bundeswehrkrankenhaus Berlin, Abteilung I – Innere Medizin, Berlin, Germany 3 Zentrum fu¨r Kardiovaskula¨re Telemedizin GmbH, Berlin, Germany
Corresponding author: Friedrich Koehler, Center for Cardiovascular Telemedicine, Charite´ – Universita¨tsmedizin Berlin, Charite´platz 1, D-10117 Berlin, Germany. Email: [email protected]
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
Spethmann et al.
33
Table 1. Inclusion and exclusion criteria for participants of the test. Inclusion criteria
Exclusion criteria
To feel healthy in a self-assessment To have run a marathon over the chosen marathon event location at least once during the last 2 years To be a non-professional runner (working in another profession) Male Written informed consent
Hospitalization of any cause over the last 5 years Chronic medication of any cause Participation in any type of study during the 6 months prior to test Any ECG morphology at rest, other than sinus rhythm Any type of known heart disease
ECG: electrocardiogram.
check-up’ is required before accreditation by most marathon organizers. However, there is no consensus on the necessary cardiologic diagnostic tests. The current practice in Germany is to perform a cardiologic check-up consisting of a clinical assessment, an electrocardiogram (ECG) at rest, an exercise test and echocardiography. These methods are valid to detect asymptomatic structural heart disease (e.g. mild to moderate aortic stenosis), but not of sufficient sensitivity to detect exercise-induced myocardial ischemia.4 Coronary angiography, cardiac magnetic resonance imaging (MRI) or cardiac computed tomography (CT) can confirm or exclude CAD with reasonable accuracy.5 However, performing such procedures in every runner over 35 years old before the race is not feasible. A new approach to prevent or help improve survival of SCD could be online ECG monitoring during running, which would allow instantaneous diagnosis of potentially fatal rhythm disorders. In case of life-threatening arrhythmias, the emergency services located along the running track could be alerted to take the runners at risk out of the race and start extended cardiologic diagnostics and treatment. Unfortunately, so far, there is no established system that enables physicians’ continuous real-time ECG monitoring during marathon running. Such a system should be able to transmit ECG signals reliably, even under extreme conditions, such as running, with extensive body movements of the sweating athletes. Moreover, there should be no interruption in ECG data transfer within a mobile phone network, even under the condition of an extreme workload caused by thousands of mobile phone customers (e.g. athletes and spectators) allocated in a very limited geographical area. In the present study, we evaluated the feasibility of a new telemedicine technology for continuous ECG recording, online transference of data by using a public mobile phone network and online diagnosis of heart rhythm in a telemedical centre (TMC), located outside the running track, during two marathon running events.
Figure 1. Holter Tele-ECG CardioMemÕ CM 4000. ß Deutsche Telekom.
Methods This proof-of-concept study was performed during two marathon running events. The second test took place 6 months after the first test.
Study population On each location, five healthy male local athletes were asked to participate through personal contacts. Inclusion and exclusion criteria are listed in Table 1. All 10 participants provided informed consent, and the study was performed according to the principles of the Declaration of Helsinki. The mean age of the volunteers was 41.7 years (range 35–55). All participants were equipped with a Holter TeleECG (Figure 1) and with a smartphone. The data transfer between the Holter Tele-ECG and the smartphone using BluethoothÕ technology was considered ‘near-field communication’. For the first test, each device was worn on one arm using armband cases
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
34
European Journal of Preventive Cardiology 21(2S)
Figure 2. Runner attached with electrodes. ß Deutsche Telekom.
(Figure 2). For the second marathon running event, both devices were worn on one arm.
Definition of the ‘ECG streaming’ via mobile phone network For both events, the published number of runners and spectators were used as indicators for an assumption of the workload within the mobile phone network during the race. The structure of mobile phone network consists in a network of so called ‘mobile phone network cells’. A network cell represents a clearly defined geographical area and an antenna. The antenna takes over all data sent from the mobile phones located in the depending geographical area and transfers this data to a central server via cable. The capacity for the data uptake in each network cell is limited. When there are too many cell phones sending in a network cell at one time, a disconnection of some mobile phones can result. When a huge workload within a mobile phone network is expected, e.g. during marathon events, the network provider increases the number of antennas and decreases the geographical areas of each cell. As result, a running athlete is ‘virtually’ passing through some hundreds of network cells during the race. When a runner is sending an ECG continuously from his mobile phone to a TMC during the race, the sending mobile phone needs a sequential connectivity to all antennas in the mobile phone network cells in which he is running through. This very complex data transfer from a mobile phone via mobile phone network to the TMC is called ‘EGG streaming’ and is influenced by many factors. For medical purposes, the near-field communication between the two devices was defined as a part of the ‘ECG streaming’.
The Holter Tele-ECG represented the benchmark of the ECG-registration during the marathon event and was performed using the standard technique (Figure 2). The Holter Tele-ECG recordings were stored within the device for post-hoc analyses. In addition, the Holter Tele-ECG recorder transmitted all detected ECG signals simultaneously to the mobile phone via near-field communication. The Holter recorder can be assumed as the starting point of the ECG streaming. In contrast, the ECG analysed in the TMC represents the endpoint of the ECG streaming. The difference in the duration (time) between the Holter TeleECG recording and the duration (time) of the ECGregistration received at the TMC, was defined as an indicator for the lost information during ECG streaming. The diagnosis of arrhythmias was used as a surrogate marker for the quality of the streamed ECG-signals. The arrhythmias – post-hoc identified in the Holter recordings – were compared with the results of the TMC online ECG surveillance. This approach addressed the issue of streamed ECG artefacts indirectly. It was defined as an indirect parameter for the amount of readably streamed ECGs.
Definition of the endpoints The primary endpoint was the percentage of ECG streaming duration (time) measured in the TMC compared with the duration of the Holter registration. Secondary endpoints included the number of interrupted ECG streamings during the race and the number of identified premature ventricular beats in the streamed ECG compared with the Holter registration.
ECG recording We used a Holter Tele-ECG (CardioMemÕ CM 400, getemed Medizin- und Informationstechnik AG, Teltow, Germany), which is a digital four-lead portable ECG device with BluetoothÕ 2.0 module for encrypted bidirectional data transfer with a smartphone. The device allows long-term recording for about 48 h and is equipped with a 2.400 touchscreen and integrated micro-SD card for intermediate storage of data. The ECG device was automatically paired with a commercial off-the-shelf smartphone (HTC Touch 2, T3333 with Windows Mobile 6.5 and Bluetooth Stack from Broadcom/Widcomm (BCM100-WCE)) via a BluetoothÕ Radio Frequency Communication Protocol link for continuous data streaming. The smartphone application acting as home broker was implemented by the Hasso-Plattner-Institute at University of Potsdam, Germany. A detailed
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
Spethmann et al.
35
description of technical connection tests and used protocols can be found elsewhere.6 The recorded data were transferred from the smartphone via the mobile network with the transmission control protocol (TCP) directly to a central streaming server located at the TMC, Charite´ – Universita¨tsmedizin Berlin. The ECG presentations on the screens of the TMC were provided by the ECG viewer CardioDayÕ (getemed Medizinund Informationstechnik AG, Teltow, Germany). During the running events, two cardiologists monitored and ran protocol on the incoming ECG data. The incoming data, stored on the TMC server as well as all Holter recordings, which had been stored in the device, were analysed after the event.
Statistics All results are expressed as mean standard deviation (SD). Statistics were calculated using IBMÕ SPSSÕ Statistics Version 21 (IBM Corporation, Armonk, NY).
Results All study participants completed the marathon at the finish without relevant medical problems. The mean running time was 3 h 37 min 21 min.
At the first marathon, there were 7000 runners and almost 150,000 spectators. At the second marathon, there were more than 15,000 runners and nearly 300,000 spectators. The mean time of Holter registration was 4 h 22 min 36 min. Owing to the need of wearing the telemedical devices during the warm-up period before start, the duration of the Holter registration was longer than the running time. In all 10 runners, the quality of the Holter Tele-ECG was sufficient for analysis. In contrast, there were significant differences in the quality of ECG streaming between the two marathon events. Beside some minutes during the warm-up period, there was actually no ECG streaming during the first marathon (see Table 2). During the first marathon, the mean percentage of ECG streaming duration compared with the duration of the Holter registration was less than 3% (see Table 2). As a consequence, no cross-check diagnostics for arrhythmias could be performed after the first marathon. In log file analyses after the marathon, the BluetoothÕ and mobile network connections were identified as error prone. The streaming application software on the smartphone was designed in a way that, in the case of missing mobile network connectivity, the BluetoothÕ connection was aborted as well. Furthermore, a bug in the Bluetooth chip firmware of the ECG device was detected afterwards, which
Table 2. The ratio between of ECG streaming duration via mobile phone network – measured in the TMC – compared with the duration of the simultaneous Holter recording during the first and second marathons.
Participant First marathon 1 2 3 4 5 Mean SD Second marathon 6 7 8 9 10 Mean SD
Duration of Holter recording (h:min)
Interruptions of Holter recording (h:min (%))
Duration of ECG streaming (h:min)
Difference in duration of Holter recording and streaming (h:min (%))
03:54 03:37 03:58 03:41 04:12 03:52 00:12
0 0 0 0 0 0 0
(0) (0) (0) (0) (0) (0) (0)
00:05 00:18 00:01 00:07 00:00 00:06 00:06
03:49 03:19 03:57 03:34 04:12 03:46 00:18
(2.1) (8.3) (0.4) (3.2) (0) (2.8) (3)
04:38 04:59 04:24 05:37 04:46 04:52 00:24
0 0 0 0 0 0 0
(0) (0) (0) (0) (0) (0) (0)
04:10 04:35 04:02 04:24 04:30 04:20 00:12
00:28 00:24 00:22 01:13 00:16 00:32 00:20
(89.9) (92.0) (91.7) (78.3) (94.4) (89.3) (5.7)
ECG: electrocardiogram; SD: standard deviation.
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
36
European Journal of Preventive Cardiology 21(2S)
avoided reconnection after a small number of connection requests. During the first marathon, the Bluetooth connection between all smartphones and their paired ECG devices broke down after 20–30 min. Also, the human body, and in particular the movements of the arms, had influenced the quality of the Bluetooth connectivity, mainly because their frequency band (2.402– 2.480 GHz) is the same as the resonance frequency of water. All these deficiencies did not occur during the inlaboratory tests. Before the second marathon, a complete new software for a better near-field communication between the two devices, as well as for an improved mobile network connectivity, was launched. To improve near-field communication, the runner of the second race had to wear both devices on one arm. As a result, an excellent quality of streaming could be observed in the second marathon event. The mean percentage of ECG streaming duration compared with the duration of the Holter registration was almost 90% (see Table 2). There was only one interruption of the ECG streaming over 3 min. In the Holter ECGs and in the TMC online ECGsurveillance, no episodes of atrial fibrillation were found. However, ventricular arrhythmias could be identified in 2/5 athletes. One athlete had 15 single polymorphic ventricular premature beats (VPB) during running. In a second runner, 117 asymptomatic VPB, including eight couplets and one triplet, were diagnosed. In total, there were no discrepancies in morphology of arrhythmias diagnoses during ECG surveillance compared with the post-hoc analysis of the Holter registration. In addition, the log files of the used mobile phones could be analysed. As a result, the mean transferring time of a QRS complex from beating in the runner’s heart to the screen at the TMC was 72 s 18 s.
Discussion To our knowledge, this proof-of-concept study is the first study demonstrating the feasibility of an online ECG surveillance using a commercial mobile network provider (Deutsche Telekom) during marathon running in a real-life setting. All technical details about the data transfer during the two marathon events concerning the information technology research have been published elsewhere.7 The use of standard hardware (Holter recorder and smartphone) reduced the developmental process down to the design of a specific software only. On the other hand, the chosen hardware reflects the very early stage of the concept. Each participant had to wear two relatively large and heavy devices over almost 42 km. This obstacle was relevant especially during the
second event. For this marathon-running event, the devices with a re-arranged software developed after the unsuccessful ECG surveillance during the first marathon, had to be worn both on one arm. To improve comfort and acceptance, the development of one specific miniaturized ECG device is mandatory, which includes only essential features (e.g. Holter storage, global positioning system module, module for mobile connectivity). Moreover, this approach would overcome the observed near-field connection difficulties. In addition, new surveillance techniques to substitute ECG cables are under development (e.g. functional clothing).8 For clinical purposes, the data transfer from individual mobile phones to the TMC has two relevant parameters: (a) ECG transfer time and (b) the loss of ECG information during periods of disconnectivity between individual mobile phones and the TMC. The observed ECG transfer time of less than 90 s is acceptable. When the runners passed specific geographical points (e.g. at 10 km) during the second marathon event, short periods of disconnectivity of less than 2 s were found in the tracking of all transferred signals of the ECG streaming (e.g. ECG signals plus time stamp). These local instabilities of the mobile network were of no clinical significance because this short period represents a maximum of three QRS complexes. This judgement is based on a post hoc cross-check between the protocol of the online ECG diagnosis done in the TMC and the stored Holter records. Importantly, all premature ventricular beats could be identified. However, the issue of these short disconnectivity periods could be relevant when upscaling the number of users significantly. In the TMC, there was a capacity to follow the incoming ECG data to a maximum of three runners per observing doctor over the whole period of the running event. As a consequence, the development of algorithms for a technical pre-analysis of the incoming ECG data is needed to present only ECGs of runners at risk on the screens in the TMC. That would enable managing a larger number of athletes. In addition to defining the requirements of further technical development resulting from our study, we also performed classical ECG-Holter monitoring during the marathons. In conclusion, online-ECG surveillance during marathon running is a promising preventive concept. Intensive further technical development is needed first before further studies with clinical endpoints can start.
Limitations There was a very small number of healthy participants in this study. As a consequence, the number of identified abnormal ECGs during running was very small.
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
Spethmann et al.
37
Moreover, concerning the vision of an arrhythmia alert system during marathon running, the test setting covered only the afferent arm – from the runner to the TMC. The efferent arm – from the TMC back to the runner, including athlete’s identification and alerting of medical staff – was not a part of this preliminary study and requires a separate test setting.
2.
3.
4.
Acknowledgements The authors thank Michael Scherf and Robert Downes (getemed AG Teltow) for providing the Holter ECG devices and analysis software; the Deutsche Telekom for providing mobile network access and supporting the recruitment of the athletes; and Professor Dr Andreas Polze and Alexander Schacht (Hasso Plattner Institute of University Potsdam) for technical support and programming the home broker software.
5.
6.
Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflicts of interest
7.
None declared.
References 1. Risgaard B, Winkel BG, Jabbari R, et al. Sports-related sudden cardiac death in a competitive and a noncompetitive athlete population aged 12 to 49 years: Data from an unselected nationwide study in Denmark. Heart Rhythm.
8.
Epub ahead of print 23 May 2014. DOI: 10.1016/ j.hrthm.2014.05.026. Harmon KG, Drezner JA, Wilson MG and Sharma S. Incidence of sudden cardiac death in athletes: a state-ofthe-art review. Heart 2014; 100(16): 1227–1234. Kim JH, Malhotra R, Chiampas G, et al. Cardiac arrest during long-distance running races. N Engl J Med 2012; 366(2): 130–140. Grazioli G, Merino B, Montserrat S, et al. Usefulness of echocardiography in preparticipation screening of competitive athletes. Rev Esp Cardiol (Engl) 2014; 67(9): 701–705. Task Force M, Montalescot G, Sechtem U, et al. 2013 ESC guidelines on the management of stable coronary artery disease: the Task Force on the management of stable coronary artery disease of the European Society of Cardiology. Eur Heart J 2013; 34(38): 2949–3003. Schacht AWR, Wolf M, von Lo¨wis M and Polze A. Live streaming of medical data – The Fontane Architecture for remote patient monitoring and its experimental evaluation. In: Proceedings of the 14th IEEE international symposium on object/component/service-oriented real-time distributed computing workshops, Newport Beach, CA, USA, 28–31 March 2011, pp.306–312. Schacht A, von Lo¨wis M and Polze A. Live ECG streaming at the Berlin Marathon 2011. In: Biomedical Engineering/765: Telehealth/766: Assistive Technologies (2012). ACTAPRESS, 2012. ISBN 978-0-88986-909-7. Stoppa M and Chiolerio A. Wearable electronics and smart textiles: a critical review. Sensors (Basel) 2014; 14: 11957–11992.
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
Electrocardiographic monitoring during marathon running: a proof of feasibility for a new telemedical approach Sebastian Spethmann, Sandra Prescher, Henryk Dreger, Herbert Nettlau, Gert Baumann, Fabian Knebel and Friedrich Koehler European Journal of Preventive Cardiology 2014 21: 32 DOI: 10.1177/2047487314553736 The online version of this article can be found at: http://cpr.sagepub.com/content/21/2_suppl/32
Published by: http://www.sagepublications.com
On behalf of: European Society of Cardiology
European Association for Cardiovascular Prevention and Rehabilitation
Additional services and information for European Journal of Preventive Cardiology can be found at: Email Alerts: http://cpr.sagepub.com/cgi/alerts Subscriptions: http://cpr.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Oct 29, 2014 What is This?
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
EURO PEAN SO CIETY O F CARDIOLOGY ®
Original scientific paper
Electrocardiographic monitoring during marathon running: a proof of feasibility for a new telemedical approach
European Journal of Preventive Cardiology 2014, Vol. 21(2S) 32–37 ! The European Society of Cardiology 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/2047487314553736 ejpc.sagepub.com
Sebastian Spethmann1,2, Sandra Prescher1, Henryk Dreger1, Herbert Nettlau3, Gert Baumann1, Fabian Knebel1 and Friedrich Koehler1,3
Abstract Objective: There is a risk for sudden cardiac death and nonfatal arrhythmias for marathon runners. A new telemedical approach to prevent sudden cardiac death could be online electrocardiogram monitoring during endurance sport events, which would allow the emergency services located along the running track to initiate instantaneous therapy. In a first proof-of-concept study we evaluate the feasibility of recording, transferring and analysing an electrocardiogram via a mobile phone (electrocardiogram streaming) and compare the quality to a conventional Holter electrocardiogram during marathon running. Methods: A total of 10 recreational endurance runners are equipped with a Holter Tele-electrocardiogram and a standard smartphone connected via BluetoothÕ to each other in order to continuously record an electrocardiogram during a first marathon event (five runners) and a second marathon event (five runners). All streaming electrocardiogram data were transferred from the device to our telemedicine centre (Charite´ Campus Mitte, Berlin, Germany); the data were monitored live and stored for a subsequent offline analysis. The primary endpoint was the percentage of successful transfer time of the streaming electrocardiogram compared with Holter electrocardiogram; the secondary endpoint was the percentage of correctly identified arrhythmias in the observed period. Results: It is technically feasible to stream an electrocardiogram during marathon running in the presence of thousands of mobile phone users. In addition, the identification of arrhythmias during a marathon is possible by electrocardiogram streaming. However, during the first race, the data transfer quality was low. After improvement of the software, in the subsequent race there was an extremely good quality in the data transfer via the mobile phone network (89%) and 100% of the rhythm disturbances could be detected in the streamed electrocardiogram. Conclusion: Online electrocardiogram surveillance during marathon running is a promising preventive concept. Intensive further technical development is needed first before further studies with clinical endpoints can start.
Keywords Marathon running, electrocardiogram streaming, telemedicine, arrhythmias, sudden cardiac death Accepted 11 September 2014
Introduction Sudden cardiac death (SCD) in marathon runs is a rare but highly emotional event.1 In Europe, the main cause of SCD in athletes is coronary artery disease (CAD) occurring mainly in middle-aged men.2 Importantly, only a minority of athletes with SCD had symptoms within 2 weeks before the race.3 Marathon running is gaining popularity, even among people over 45 years old. Many of them begin competitive endurance sports in their 40s. A ‘health
1 Medizinische Klinik fu¨r Kardiologie und Angiologie, Campus Mitte, Charite´ – Universita¨tsmedizin Berlin, Berlin, Germany 2 Bundeswehrkrankenhaus Berlin, Abteilung I – Innere Medizin, Berlin, Germany 3 Zentrum fu¨r Kardiovaskula¨re Telemedizin GmbH, Berlin, Germany
Corresponding author: Friedrich Koehler, Center for Cardiovascular Telemedicine, Charite´ – Universita¨tsmedizin Berlin, Charite´platz 1, D-10117 Berlin, Germany. Email: [email protected]
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
Spethmann et al.
33
Table 1. Inclusion and exclusion criteria for participants of the test. Inclusion criteria
Exclusion criteria
To feel healthy in a self-assessment To have run a marathon over the chosen marathon event location at least once during the last 2 years To be a non-professional runner (working in another profession) Male Written informed consent
Hospitalization of any cause over the last 5 years Chronic medication of any cause Participation in any type of study during the 6 months prior to test Any ECG morphology at rest, other than sinus rhythm Any type of known heart disease
ECG: electrocardiogram.
check-up’ is required before accreditation by most marathon organizers. However, there is no consensus on the necessary cardiologic diagnostic tests. The current practice in Germany is to perform a cardiologic check-up consisting of a clinical assessment, an electrocardiogram (ECG) at rest, an exercise test and echocardiography. These methods are valid to detect asymptomatic structural heart disease (e.g. mild to moderate aortic stenosis), but not of sufficient sensitivity to detect exercise-induced myocardial ischemia.4 Coronary angiography, cardiac magnetic resonance imaging (MRI) or cardiac computed tomography (CT) can confirm or exclude CAD with reasonable accuracy.5 However, performing such procedures in every runner over 35 years old before the race is not feasible. A new approach to prevent or help improve survival of SCD could be online ECG monitoring during running, which would allow instantaneous diagnosis of potentially fatal rhythm disorders. In case of life-threatening arrhythmias, the emergency services located along the running track could be alerted to take the runners at risk out of the race and start extended cardiologic diagnostics and treatment. Unfortunately, so far, there is no established system that enables physicians’ continuous real-time ECG monitoring during marathon running. Such a system should be able to transmit ECG signals reliably, even under extreme conditions, such as running, with extensive body movements of the sweating athletes. Moreover, there should be no interruption in ECG data transfer within a mobile phone network, even under the condition of an extreme workload caused by thousands of mobile phone customers (e.g. athletes and spectators) allocated in a very limited geographical area. In the present study, we evaluated the feasibility of a new telemedicine technology for continuous ECG recording, online transference of data by using a public mobile phone network and online diagnosis of heart rhythm in a telemedical centre (TMC), located outside the running track, during two marathon running events.
Figure 1. Holter Tele-ECG CardioMemÕ CM 4000. ß Deutsche Telekom.
Methods This proof-of-concept study was performed during two marathon running events. The second test took place 6 months after the first test.
Study population On each location, five healthy male local athletes were asked to participate through personal contacts. Inclusion and exclusion criteria are listed in Table 1. All 10 participants provided informed consent, and the study was performed according to the principles of the Declaration of Helsinki. The mean age of the volunteers was 41.7 years (range 35–55). All participants were equipped with a Holter TeleECG (Figure 1) and with a smartphone. The data transfer between the Holter Tele-ECG and the smartphone using BluethoothÕ technology was considered ‘near-field communication’. For the first test, each device was worn on one arm using armband cases
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
34
European Journal of Preventive Cardiology 21(2S)
Figure 2. Runner attached with electrodes. ß Deutsche Telekom.
(Figure 2). For the second marathon running event, both devices were worn on one arm.
Definition of the ‘ECG streaming’ via mobile phone network For both events, the published number of runners and spectators were used as indicators for an assumption of the workload within the mobile phone network during the race. The structure of mobile phone network consists in a network of so called ‘mobile phone network cells’. A network cell represents a clearly defined geographical area and an antenna. The antenna takes over all data sent from the mobile phones located in the depending geographical area and transfers this data to a central server via cable. The capacity for the data uptake in each network cell is limited. When there are too many cell phones sending in a network cell at one time, a disconnection of some mobile phones can result. When a huge workload within a mobile phone network is expected, e.g. during marathon events, the network provider increases the number of antennas and decreases the geographical areas of each cell. As result, a running athlete is ‘virtually’ passing through some hundreds of network cells during the race. When a runner is sending an ECG continuously from his mobile phone to a TMC during the race, the sending mobile phone needs a sequential connectivity to all antennas in the mobile phone network cells in which he is running through. This very complex data transfer from a mobile phone via mobile phone network to the TMC is called ‘EGG streaming’ and is influenced by many factors. For medical purposes, the near-field communication between the two devices was defined as a part of the ‘ECG streaming’.
The Holter Tele-ECG represented the benchmark of the ECG-registration during the marathon event and was performed using the standard technique (Figure 2). The Holter Tele-ECG recordings were stored within the device for post-hoc analyses. In addition, the Holter Tele-ECG recorder transmitted all detected ECG signals simultaneously to the mobile phone via near-field communication. The Holter recorder can be assumed as the starting point of the ECG streaming. In contrast, the ECG analysed in the TMC represents the endpoint of the ECG streaming. The difference in the duration (time) between the Holter TeleECG recording and the duration (time) of the ECGregistration received at the TMC, was defined as an indicator for the lost information during ECG streaming. The diagnosis of arrhythmias was used as a surrogate marker for the quality of the streamed ECG-signals. The arrhythmias – post-hoc identified in the Holter recordings – were compared with the results of the TMC online ECG surveillance. This approach addressed the issue of streamed ECG artefacts indirectly. It was defined as an indirect parameter for the amount of readably streamed ECGs.
Definition of the endpoints The primary endpoint was the percentage of ECG streaming duration (time) measured in the TMC compared with the duration of the Holter registration. Secondary endpoints included the number of interrupted ECG streamings during the race and the number of identified premature ventricular beats in the streamed ECG compared with the Holter registration.
ECG recording We used a Holter Tele-ECG (CardioMemÕ CM 400, getemed Medizin- und Informationstechnik AG, Teltow, Germany), which is a digital four-lead portable ECG device with BluetoothÕ 2.0 module for encrypted bidirectional data transfer with a smartphone. The device allows long-term recording for about 48 h and is equipped with a 2.400 touchscreen and integrated micro-SD card for intermediate storage of data. The ECG device was automatically paired with a commercial off-the-shelf smartphone (HTC Touch 2, T3333 with Windows Mobile 6.5 and Bluetooth Stack from Broadcom/Widcomm (BCM100-WCE)) via a BluetoothÕ Radio Frequency Communication Protocol link for continuous data streaming. The smartphone application acting as home broker was implemented by the Hasso-Plattner-Institute at University of Potsdam, Germany. A detailed
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
Spethmann et al.
35
description of technical connection tests and used protocols can be found elsewhere.6 The recorded data were transferred from the smartphone via the mobile network with the transmission control protocol (TCP) directly to a central streaming server located at the TMC, Charite´ – Universita¨tsmedizin Berlin. The ECG presentations on the screens of the TMC were provided by the ECG viewer CardioDayÕ (getemed Medizinund Informationstechnik AG, Teltow, Germany). During the running events, two cardiologists monitored and ran protocol on the incoming ECG data. The incoming data, stored on the TMC server as well as all Holter recordings, which had been stored in the device, were analysed after the event.
Statistics All results are expressed as mean standard deviation (SD). Statistics were calculated using IBMÕ SPSSÕ Statistics Version 21 (IBM Corporation, Armonk, NY).
Results All study participants completed the marathon at the finish without relevant medical problems. The mean running time was 3 h 37 min 21 min.
At the first marathon, there were 7000 runners and almost 150,000 spectators. At the second marathon, there were more than 15,000 runners and nearly 300,000 spectators. The mean time of Holter registration was 4 h 22 min 36 min. Owing to the need of wearing the telemedical devices during the warm-up period before start, the duration of the Holter registration was longer than the running time. In all 10 runners, the quality of the Holter Tele-ECG was sufficient for analysis. In contrast, there were significant differences in the quality of ECG streaming between the two marathon events. Beside some minutes during the warm-up period, there was actually no ECG streaming during the first marathon (see Table 2). During the first marathon, the mean percentage of ECG streaming duration compared with the duration of the Holter registration was less than 3% (see Table 2). As a consequence, no cross-check diagnostics for arrhythmias could be performed after the first marathon. In log file analyses after the marathon, the BluetoothÕ and mobile network connections were identified as error prone. The streaming application software on the smartphone was designed in a way that, in the case of missing mobile network connectivity, the BluetoothÕ connection was aborted as well. Furthermore, a bug in the Bluetooth chip firmware of the ECG device was detected afterwards, which
Table 2. The ratio between of ECG streaming duration via mobile phone network – measured in the TMC – compared with the duration of the simultaneous Holter recording during the first and second marathons.
Participant First marathon 1 2 3 4 5 Mean SD Second marathon 6 7 8 9 10 Mean SD
Duration of Holter recording (h:min)
Interruptions of Holter recording (h:min (%))
Duration of ECG streaming (h:min)
Difference in duration of Holter recording and streaming (h:min (%))
03:54 03:37 03:58 03:41 04:12 03:52 00:12
0 0 0 0 0 0 0
(0) (0) (0) (0) (0) (0) (0)
00:05 00:18 00:01 00:07 00:00 00:06 00:06
03:49 03:19 03:57 03:34 04:12 03:46 00:18
(2.1) (8.3) (0.4) (3.2) (0) (2.8) (3)
04:38 04:59 04:24 05:37 04:46 04:52 00:24
0 0 0 0 0 0 0
(0) (0) (0) (0) (0) (0) (0)
04:10 04:35 04:02 04:24 04:30 04:20 00:12
00:28 00:24 00:22 01:13 00:16 00:32 00:20
(89.9) (92.0) (91.7) (78.3) (94.4) (89.3) (5.7)
ECG: electrocardiogram; SD: standard deviation.
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
36
European Journal of Preventive Cardiology 21(2S)
avoided reconnection after a small number of connection requests. During the first marathon, the Bluetooth connection between all smartphones and their paired ECG devices broke down after 20–30 min. Also, the human body, and in particular the movements of the arms, had influenced the quality of the Bluetooth connectivity, mainly because their frequency band (2.402– 2.480 GHz) is the same as the resonance frequency of water. All these deficiencies did not occur during the inlaboratory tests. Before the second marathon, a complete new software for a better near-field communication between the two devices, as well as for an improved mobile network connectivity, was launched. To improve near-field communication, the runner of the second race had to wear both devices on one arm. As a result, an excellent quality of streaming could be observed in the second marathon event. The mean percentage of ECG streaming duration compared with the duration of the Holter registration was almost 90% (see Table 2). There was only one interruption of the ECG streaming over 3 min. In the Holter ECGs and in the TMC online ECGsurveillance, no episodes of atrial fibrillation were found. However, ventricular arrhythmias could be identified in 2/5 athletes. One athlete had 15 single polymorphic ventricular premature beats (VPB) during running. In a second runner, 117 asymptomatic VPB, including eight couplets and one triplet, were diagnosed. In total, there were no discrepancies in morphology of arrhythmias diagnoses during ECG surveillance compared with the post-hoc analysis of the Holter registration. In addition, the log files of the used mobile phones could be analysed. As a result, the mean transferring time of a QRS complex from beating in the runner’s heart to the screen at the TMC was 72 s 18 s.
Discussion To our knowledge, this proof-of-concept study is the first study demonstrating the feasibility of an online ECG surveillance using a commercial mobile network provider (Deutsche Telekom) during marathon running in a real-life setting. All technical details about the data transfer during the two marathon events concerning the information technology research have been published elsewhere.7 The use of standard hardware (Holter recorder and smartphone) reduced the developmental process down to the design of a specific software only. On the other hand, the chosen hardware reflects the very early stage of the concept. Each participant had to wear two relatively large and heavy devices over almost 42 km. This obstacle was relevant especially during the
second event. For this marathon-running event, the devices with a re-arranged software developed after the unsuccessful ECG surveillance during the first marathon, had to be worn both on one arm. To improve comfort and acceptance, the development of one specific miniaturized ECG device is mandatory, which includes only essential features (e.g. Holter storage, global positioning system module, module for mobile connectivity). Moreover, this approach would overcome the observed near-field connection difficulties. In addition, new surveillance techniques to substitute ECG cables are under development (e.g. functional clothing).8 For clinical purposes, the data transfer from individual mobile phones to the TMC has two relevant parameters: (a) ECG transfer time and (b) the loss of ECG information during periods of disconnectivity between individual mobile phones and the TMC. The observed ECG transfer time of less than 90 s is acceptable. When the runners passed specific geographical points (e.g. at 10 km) during the second marathon event, short periods of disconnectivity of less than 2 s were found in the tracking of all transferred signals of the ECG streaming (e.g. ECG signals plus time stamp). These local instabilities of the mobile network were of no clinical significance because this short period represents a maximum of three QRS complexes. This judgement is based on a post hoc cross-check between the protocol of the online ECG diagnosis done in the TMC and the stored Holter records. Importantly, all premature ventricular beats could be identified. However, the issue of these short disconnectivity periods could be relevant when upscaling the number of users significantly. In the TMC, there was a capacity to follow the incoming ECG data to a maximum of three runners per observing doctor over the whole period of the running event. As a consequence, the development of algorithms for a technical pre-analysis of the incoming ECG data is needed to present only ECGs of runners at risk on the screens in the TMC. That would enable managing a larger number of athletes. In addition to defining the requirements of further technical development resulting from our study, we also performed classical ECG-Holter monitoring during the marathons. In conclusion, online-ECG surveillance during marathon running is a promising preventive concept. Intensive further technical development is needed first before further studies with clinical endpoints can start.
Limitations There was a very small number of healthy participants in this study. As a consequence, the number of identified abnormal ECGs during running was very small.
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
Spethmann et al.
37
Moreover, concerning the vision of an arrhythmia alert system during marathon running, the test setting covered only the afferent arm – from the runner to the TMC. The efferent arm – from the TMC back to the runner, including athlete’s identification and alerting of medical staff – was not a part of this preliminary study and requires a separate test setting.
2.
3.
4.
Acknowledgements The authors thank Michael Scherf and Robert Downes (getemed AG Teltow) for providing the Holter ECG devices and analysis software; the Deutsche Telekom for providing mobile network access and supporting the recruitment of the athletes; and Professor Dr Andreas Polze and Alexander Schacht (Hasso Plattner Institute of University Potsdam) for technical support and programming the home broker software.
5.
6.
Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflicts of interest
7.
None declared.
References 1. Risgaard B, Winkel BG, Jabbari R, et al. Sports-related sudden cardiac death in a competitive and a noncompetitive athlete population aged 12 to 49 years: Data from an unselected nationwide study in Denmark. Heart Rhythm.
8.
Epub ahead of print 23 May 2014. DOI: 10.1016/ j.hrthm.2014.05.026. Harmon KG, Drezner JA, Wilson MG and Sharma S. Incidence of sudden cardiac death in athletes: a state-ofthe-art review. Heart 2014; 100(16): 1227–1234. Kim JH, Malhotra R, Chiampas G, et al. Cardiac arrest during long-distance running races. N Engl J Med 2012; 366(2): 130–140. Grazioli G, Merino B, Montserrat S, et al. Usefulness of echocardiography in preparticipation screening of competitive athletes. Rev Esp Cardiol (Engl) 2014; 67(9): 701–705. Task Force M, Montalescot G, Sechtem U, et al. 2013 ESC guidelines on the management of stable coronary artery disease: the Task Force on the management of stable coronary artery disease of the European Society of Cardiology. Eur Heart J 2013; 34(38): 2949–3003. Schacht AWR, Wolf M, von Lo¨wis M and Polze A. Live streaming of medical data – The Fontane Architecture for remote patient monitoring and its experimental evaluation. In: Proceedings of the 14th IEEE international symposium on object/component/service-oriented real-time distributed computing workshops, Newport Beach, CA, USA, 28–31 March 2011, pp.306–312. Schacht A, von Lo¨wis M and Polze A. Live ECG streaming at the Berlin Marathon 2011. In: Biomedical Engineering/765: Telehealth/766: Assistive Technologies (2012). ACTAPRESS, 2012. ISBN 978-0-88986-909-7. Stoppa M and Chiolerio A. Wearable electronics and smart textiles: a critical review. Sensors (Basel) 2014; 14: 11957–11992.
Downloaded from cpr.sagepub.com at UNIV OF GUELPH on November 7, 2014
