bs_bs_banner
1040
THOUGHTS AND PROGRESS
Detection of Premature Ventricular Contractions on a Ventricular Electrocardiogram for Patients With Left Ventricular Assist Devices †Sung Min Park, *Jin Hong Lee, and *Seong Wook Choi *Department of Mechanical and Biomedical Engineering, College of Engineering, Kangwon National University; and †School of Medicine, Kangwon National University, Chuncheon, Korea Abstract: The ventricular electrocardiogram (v-ECG) was developed for long-term monitoring of heartbeats in patients with a left ventricular assist device (LVAD) and does not normally have the functionality necessary to detect additional heart irregularities that can progress to critical arrhythmias. Although the v-ECG has the benefits of physiological optimization and counterpulsation control, when abnormal heartbeats occur, the v-ECG does not show the distinct abnormal waveform that enables easy detection of an abnormal heartbeat among normal heartbeats on the conventional ECG. In this study, the v-ECGs of normal and abnormal heartbeats are compared with each other with respect to peak-to-peak voltage, area, and maximal slopes, and a new method to detect abnormal heartbeats is suggested. In a series of animal experiments with three porcine models (Yorkshire pigs weighing 30–40 kg), a v-ECG and conventional ECG were taken simultaneously during LVAD perfusion. Clinical experts found 104 abnormal heartbeats from the saved conventional ECG data and confirmed that the other 3159 heartbeats were normal. Almost all of the abnormal heartbeats were premature ventricular contractions (PVCs), and there was short-term tachycardia for 3 s. A personal computer was used to automatically detect abnormal heartbeats with the v-ECG according to the new method, and its results were compared with the clinicians’ results. The new method found abnormal heartbeats with 90% accuracy, and less than 15% of the total PVCs were missed. Key Words: Left ventricular assist device—Ventricular electrocardiogram—Arrhythmia—Algorithm.
The left ventricular assist device (LVAD) is considered the most effective treatment to improve a patient’s chance for survival in end-stage heart failure (1–7). However, there are some reports that LVADs can induce additional heart irregularities, such as arrhythmias, by causing changes in the load on the heart (8–12). In fact, such heart irregularities may be the cause of 5% of all LVAD patient deaths doi:10.1111/aor.12306 Received October 2013; revised February 2014. Address correspondence and reprint requests to Prof. Seong Wook Choi, Department of Mechanical and Biomedical Engineering, College of Engineering, Kangwon National University, 192-1 Hyja-dong, Chuncheon-si 200-701, Korea. E-mail: swchoe@ kangwon.ac.kr Artif Organs, Vol. 38, No. 12, 2014
(6). In some cases, an implantable cardioverter– defibrillator (ICD) had been implanted in the LVAD patient, and in other cases, a pacemaker had been implanted after weaning from the LVAD (13,14). For patients who develop arrhythmias within a week after LVAD implantation, patient survival rates are lower than among those who do not develop arrhythmias or do so only after a week (6). Therefore, heart function should be continuously monitored to detect abnormal heartbeat that can develop into a critical heart irregularity. Although accurate detection of abnormal heartbeat is usually accomplished with a conventional electrocardiogram (ECG), the conventional ECG requires many electrodes and long lead wires that are susceptible to noise from the LVAD (15–19). These additional devices and lead wires severely burden LVAD patients, who are already required to carry an external controller and batteries (17). To enable patients to move freely and to eliminate the various external noise and motion artifacts, ECG electrodes and wires may be substituted with implantable electrodes and wires set on ventricular muscle (20–22). The ventricular electrocardiogram (v-ECG) is obtained through one or several electrodes directly located on the ventricular muscle and has been applied in patients with pacemakers, ICDs, and LVADs with high reliability. However, the v-ECG has a different waveform from the conventional ECG and needs to be studied carefully to accurately analyze the heart’s condition (22). In this study, a new method is suggested for automatic detection of arrhythmias, especially premature ventricular contractions (PVCs), by v-ECG; until now, the v-ECG has not been used to detect PVCs (23–25). A series of animal experiments with three porcine models was conducted to evaluate the effectiveness of the new method. MATERIALS AND METHODS PVC detection algorithm On a conventional ECG, PVCs have distinct characteristics, such as a QRS complex with different widths and amplitudes compared with normal contractions (Fig. 1a); the changes in the QRS complex are caused by an abnormal path of electrical stimuli in heart muscle, and PVCs can be easily differentiated by the waveform of the QRS complex. In contrast, in the v-ECG, the distinct characteristics of the PVC R waves are not readily observable without deliberate comparison, as shown in Fig. 1b. To determine the differences between a normal contraction and a PVC on v-ECG, three parameters
THOUGHTS AND PROGRESS of the R wave were selected: (i) the peak-to-peak value, which represents the amplitude of the rapidly changing v-ECG during heart contraction; (ii) the area, which represents the area within the R wave and baseline; and (iii) maximal slope, which represents the highest value of the measured rate of change of the R wave. The peak-to-peak value of the R wave was obtained using Eq. 1: t2 ′
Peak-to-peak value = ∫ Sn dt t 1′
Sn = 0 when
Sn =
dVv-ECG ( t ) dt
(1)
dVv-ECG ( t ) ≥0 dt
when
dVv-ECG ( t )
1040
THOUGHTS AND PROGRESS
Detection of Premature Ventricular Contractions on a Ventricular Electrocardiogram for Patients With Left Ventricular Assist Devices †Sung Min Park, *Jin Hong Lee, and *Seong Wook Choi *Department of Mechanical and Biomedical Engineering, College of Engineering, Kangwon National University; and †School of Medicine, Kangwon National University, Chuncheon, Korea Abstract: The ventricular electrocardiogram (v-ECG) was developed for long-term monitoring of heartbeats in patients with a left ventricular assist device (LVAD) and does not normally have the functionality necessary to detect additional heart irregularities that can progress to critical arrhythmias. Although the v-ECG has the benefits of physiological optimization and counterpulsation control, when abnormal heartbeats occur, the v-ECG does not show the distinct abnormal waveform that enables easy detection of an abnormal heartbeat among normal heartbeats on the conventional ECG. In this study, the v-ECGs of normal and abnormal heartbeats are compared with each other with respect to peak-to-peak voltage, area, and maximal slopes, and a new method to detect abnormal heartbeats is suggested. In a series of animal experiments with three porcine models (Yorkshire pigs weighing 30–40 kg), a v-ECG and conventional ECG were taken simultaneously during LVAD perfusion. Clinical experts found 104 abnormal heartbeats from the saved conventional ECG data and confirmed that the other 3159 heartbeats were normal. Almost all of the abnormal heartbeats were premature ventricular contractions (PVCs), and there was short-term tachycardia for 3 s. A personal computer was used to automatically detect abnormal heartbeats with the v-ECG according to the new method, and its results were compared with the clinicians’ results. The new method found abnormal heartbeats with 90% accuracy, and less than 15% of the total PVCs were missed. Key Words: Left ventricular assist device—Ventricular electrocardiogram—Arrhythmia—Algorithm.
The left ventricular assist device (LVAD) is considered the most effective treatment to improve a patient’s chance for survival in end-stage heart failure (1–7). However, there are some reports that LVADs can induce additional heart irregularities, such as arrhythmias, by causing changes in the load on the heart (8–12). In fact, such heart irregularities may be the cause of 5% of all LVAD patient deaths doi:10.1111/aor.12306 Received October 2013; revised February 2014. Address correspondence and reprint requests to Prof. Seong Wook Choi, Department of Mechanical and Biomedical Engineering, College of Engineering, Kangwon National University, 192-1 Hyja-dong, Chuncheon-si 200-701, Korea. E-mail: swchoe@ kangwon.ac.kr Artif Organs, Vol. 38, No. 12, 2014
(6). In some cases, an implantable cardioverter– defibrillator (ICD) had been implanted in the LVAD patient, and in other cases, a pacemaker had been implanted after weaning from the LVAD (13,14). For patients who develop arrhythmias within a week after LVAD implantation, patient survival rates are lower than among those who do not develop arrhythmias or do so only after a week (6). Therefore, heart function should be continuously monitored to detect abnormal heartbeat that can develop into a critical heart irregularity. Although accurate detection of abnormal heartbeat is usually accomplished with a conventional electrocardiogram (ECG), the conventional ECG requires many electrodes and long lead wires that are susceptible to noise from the LVAD (15–19). These additional devices and lead wires severely burden LVAD patients, who are already required to carry an external controller and batteries (17). To enable patients to move freely and to eliminate the various external noise and motion artifacts, ECG electrodes and wires may be substituted with implantable electrodes and wires set on ventricular muscle (20–22). The ventricular electrocardiogram (v-ECG) is obtained through one or several electrodes directly located on the ventricular muscle and has been applied in patients with pacemakers, ICDs, and LVADs with high reliability. However, the v-ECG has a different waveform from the conventional ECG and needs to be studied carefully to accurately analyze the heart’s condition (22). In this study, a new method is suggested for automatic detection of arrhythmias, especially premature ventricular contractions (PVCs), by v-ECG; until now, the v-ECG has not been used to detect PVCs (23–25). A series of animal experiments with three porcine models was conducted to evaluate the effectiveness of the new method. MATERIALS AND METHODS PVC detection algorithm On a conventional ECG, PVCs have distinct characteristics, such as a QRS complex with different widths and amplitudes compared with normal contractions (Fig. 1a); the changes in the QRS complex are caused by an abnormal path of electrical stimuli in heart muscle, and PVCs can be easily differentiated by the waveform of the QRS complex. In contrast, in the v-ECG, the distinct characteristics of the PVC R waves are not readily observable without deliberate comparison, as shown in Fig. 1b. To determine the differences between a normal contraction and a PVC on v-ECG, three parameters
THOUGHTS AND PROGRESS of the R wave were selected: (i) the peak-to-peak value, which represents the amplitude of the rapidly changing v-ECG during heart contraction; (ii) the area, which represents the area within the R wave and baseline; and (iii) maximal slope, which represents the highest value of the measured rate of change of the R wave. The peak-to-peak value of the R wave was obtained using Eq. 1: t2 ′
Peak-to-peak value = ∫ Sn dt t 1′
Sn = 0 when
Sn =
dVv-ECG ( t ) dt
(1)
dVv-ECG ( t ) ≥0 dt
when
dVv-ECG ( t )
