Wireless ECG Body Sensor During Field Testing

Today's lifestyle presents high risk for cardiovascular diseases, which is a leading cause of death worldwide. Previous studies have targeted various risk factors for development of cardiovascular diseases, such as high blood pressure, smoking, type 2 diabetes, obesity, psychological stress and physical inactivity.

Increased level of physical activity has numerous beneficial long-term effects on cardiovascular system, and might decrease relative risk of sudden cardiac death. However, the risk of unexpected cardiovascular events is increased during, and immediately after intense exercise. Two studies conducted in France, reported a daily incidence of three sudden deaths and four myocardial infarctions during regular physical activity in the general population. Twelve-year longitudinal study showed that average age of young recreational athletes in Switzerland who died during physical activity was 27.6 years, with the incidence of 0.52/100.000 per year. Some studies reported dramatic increase of unexpected cardiovascular events in a middle age group of recreational population, which was at the level of 6.5/100.000 per year. Furthermore, several studies reported higher incidence of sports related cardiovascular events between general population in comparison with professional athletes. Males are more affected, and the most affected age group is middle-aged between 35 and 59 years.

Previous studies noticed that the key challenge is identification of structural heart disease and inheritable conditions that increase incidence for lethal arrhythmias during exercise, because sudden cardiac events often occurs without any warning or prediction. Therefore, early detection of persons with higher risk for sudden cardiac death is of crucial importance. In this regard, numerous European countries have introduced a mandatory medical examination for all persons registered in sports clubs or associations. Exercise stress test (EST) which includes electrocardiographic (ECG) monitoring is an important part of examination. However, the problem arises from the fact that EST is especially sensitive for testing individuals with previous known symptoms of cardiac disease, while the positive predictive value of an EST in asymptomatic subjects is relatively low. On the other hand, it was proven that long term ECG monitoring during regular daily activities can be very useful in detection of heart disease in people who had normal 12-lead ECG, and negative Master two-step tests. Also, it was proven that it is possible to detect atrial fibrillation in asymptomatic people with long-term one-channel ECG monitoring.

Recent studies showed that the wireless body ECG sensor is a feasible solution for reliable and accurate long-term heart rhythm monitoring, but the sensor placement and fixation is an important factor which can influence the signal quality and needs to be taken into consideration. In accordance with previously said, our main goal was to determine whether it is possible to measure quality electrocardiographic report using ECG body-sensor during standardized outdoor tests. Additional aim was to determine whether different types of sensors' fixation affect quality of the ECG signal during submaximal and maximal running settings.

Twenty-three were included, 13 males and 10 females, age between 19 and 23 years (Age = 20.56±1.19 years; Body height = 177.28±9.76 cm; Body mass = 70.11±7.68 kg). All participants were physically active non-athletes, with daily involvement (3-5 times per week, 30 to 60 minutes of moderate to high intensity) in various physical activities (team sports, cyclic activities, resistance training). To meet the criteria for participation in the study, participants had to confirm they were healthy, and without known previous cardiac problems. In accordance with the Declaration of Helsinki participants gave the signed consent for participation in the study. Prior to the tests, the purpose and protocols of the study were explained in details to each participant. Ethics approval for the study was obtained from the Institutional Ethical committee (Approval no. 02-1359/18-2).

ECG Signal Monitoring Wireless ECG body-sensor Savvy (Saving d.o.o., Ljubljana, Slovenia), which is a certified medical device, was used for collecting ECG data. The body-sensor allows ECG measurements during long-term exercise, with a sampling rate between 125 Hz and 1000 Hz. An Android application, MobECG, which runs on a smartphone, connected to the sensor via Bluetooth, captures and displays the measured data and saves it in the smartphone memory for further processing.

The optimal placement for the electrodes was close to the heart in order to obtain the appropriate amplitude of ECG signal. Considering that the signals from the electrical muscular activity (EMG) could disturb the ECG signal, especially during physical activity, sensor position was placed to avoid large muscle groups. In this study, we tested position marked as left inferior (LI). At the LI position electrodes were placed at the positions of standard anterior precordial leads V1 and V2, and then sensor was translated downward by approximately 10 cm (below the xiphoid), where the influence of muscular disturbances is expected to be minimal.

In the first test standard self-adhesive Skintact ECG electrodes type PREMIER T-60 were used (Leonhard Lang GmbH, Innsbruck, Austria). The electrodes were in the original packaging and used according to the manufacturer's instructions. The ECG electrodes were positioned 5 cm apart. Before the positioning, the skin of the subjects was cleaned with diluted ethanol. Then, two electrodes were stuck at the LI position, and the ECG body-sensor was connected. Sensor was additionally fixed with self-adhesive Omniplast 2.5 cm tape (Paul Hartmann AG, Heidenheim, Germany), specially designed to fix Holter electrodes (see Figure 1). Both parts of the sensor were fixed together with one, approximately 40-cm-long strip of tape. In the second measurement Polar soft strap belt (Polar Electro Oy, Finland) was used to connect the sensor.

The measured ECG data were continuously stored in the smartphone memory and transferred after the completion of the measurements to a personal computer. All the ECG measurements were analysed retrospectively. The subsequent ECG analyses and signal quality were made with medically certified Holter interpretation software QuickReader®AFT-1000 (Holter Supplies, Paris, France). Moreover, all the measurements and the results were visually examined by a medical doctor. Only signals where QRS waves were clear and recognized without doubt where marked as useful. Variables used in study were: Detected QRS, QRS detected as negative (%), False negative detection (%), Correct software detection (%), Load / speed with still assessable HR (%), Maximal HR [bpm], Average running time (s) (for Cooper 2400 and 100 m tests) and Average achieved intensity (for Shutlle run).

In this study we used three different standardized running tests, one sub-maximal test - Cooper 2400 m (C), and two maximal (all-out) tests - 100 m sprint (S) and shuttle run (SR). In C test participants had to run 2400 m, while in S test they had to run 100 m at the highest possible speed. In SR test standard test protocol was used. Participants had to run 20 meters and step across the line before audio sign (beep). Test started with speed of 8.5 km/h which was increased 0.5 km/h every minute, until exhaustion. Test was finished when subject couldn't reach the line two times in a row, before audio sign.

Every participant made six tests, two C, S and SR tests. All tests were made in the morning, and participant performed one test per day and retest until next day. Break between two tests was at least 24 hours. Before performing the tests, the electrodes and sensor were positioned at the LI position, and the participants sat down for 5 minutes, while the ECG was recording. When the test was finished the participant sat down and rest for next 5 minutes without interrupting the ECG recording.

Descriptive statistics were calculated for all measured variables, and test of normality was made for all the data. For the data which was not distributed normally by Kolmogorov-Smirnov and Shapiro-Wilk test results, based on the results of Skewness, transformation with function Log10 or Log10 reflexion was made. For all original and transformed variables that showed normal distribution, for determination of differences student t-test for dependent variables was used, while non-parametric Wilcoxon test was used for those which were not distributed normally. Intra-class correlation was counted for repeated measurements. The significance level was set to p < 0.05. The statistical analysis was conducted using IBM SPSS Statistics 20 software.