Available online at www.sciencedirect.com
ScienceDirect Journal of Electrocardiology xx (2014) xxx – xxx www.jecgonline.com
Precordial electrode placement for optimal ECG monitoring: Implications for ambulatory monitor devices and event recorders☆ Sotirios Nedios, MD, a,⁎ Iñaki Romero, MSc, PhD, b Jin-Hong Gerds-Li, MD, c Eckard Fleck, MD, c Charalampos Kriatselis, MD c a
c
Heart Center, University of Leipzig, Department of Electrophysiology, Strümpelstr. 39, Leipzig, Germany b Holst Centre/IMEC, High Tech Campus 31, AE, Eindhoven, Netherlands German Heart Institute Berlin, Department of Internal Medicine and Cardiology, Augustenburgerplatz 1, Berlin, Germany
Abstract
Introduction: Detection of QRS complexes, P-waves and atrial fibrillation f-waves in electrocardiographic (ECG) signals is critical for the correct diagnosis of arrhythmias. We aimed to find the best bipolar lead (BL) with the highest signal amplitude and shortest inter-electrode spacing. Methods: ECG signals (120 seconds) were recorded in 36 patients with 16 precordial electrodes placed in a standardized pattern. An average signal was analysed for each of 120 possible BLs obtained by calculating the difference between pairs of unipolar leads. Peak-to-peak amplitudes of QRS waves (50 ms around R-peak) and P waves (270–70 ms before R-peak) were calculated. For patients with atrial fibrillation, power of the fibrillatory (f) wave was used instead. Maximum values at each distance were considered and differentiation analysis was performed based on incremental changes (amplitude to distance). Results: There was a significant correlation between distance and QRS-amplitude (r = 0.78, p b 0.001), P-wave amplitude (r = 0.60, p b 0.01) and f-wave power (r = 0.79, p b 0.001). The range of values was: QRS-amplitude 0.7–2.33 mV, P-wave amplitude 0.07–0.18 mV, and f-wave power 0.55–2.12 mV 2/s. The maximum value for the shortest distance was on a heart-aligned axis over the left ventricle for the QRS complex (1.9 mV at 8.7 cm) and over the atria for the P-wave (0.98 mV) and f-waves (1.45 mV 2/s at 8 cm, respectively). Conclusion: There is a strong positive correlation between electrode distance and ECG signalamplitude. Distance of 8 cm on a heart-aligned axis and over the relevant heart-chamber provides the highest signal amplitude for the shortest distance. These findings are essential for the design and use of ambulatory monitoring devices. © 2014 Elsevier Inc. All rights reserved.
Keywords:
Ambulatory; Patient monitoring; Electrocardiography (ECG); Atrial fibrillation
Introduction Palpitations and arrhythmias are among the most common symptoms which prompt patients to consult healthcare services [1]. The diagnosis and treatment of these symptoms often proves to be strenuous, frustrating and elusive, despite numerous time-consuming and costly investigations. Cardiac arrhythmias and especially atrial fibrillation (AF) are the most common cause of palpitations [2]. Electrocardiographic (ECG) documentation during symptomatic episodes is
☆
Disclosure: None of the authors have any conflict of interest, financial or otherwise. ⁎ Corresponding author at: Department of Electrophysiology, Heart Center, University of Leipzig, Strümpelstrasse 39, 04289 Leipzig, Germany. E-mail address: [email protected] http://dx.doi.org/10.1016/j.jelectrocard.2014.04.003 0022-0736/© 2014 Elsevier Inc. All rights reserved.
necessary to establish causality, in order to guide the therapeutic strategy [3,4]. This is of paramount importance when it comes to detection of AF, which partially remains asymptomatic, especially after an ablation procedure [5–7]. Ambulatory monitoring with Holter-ECG, telemetry, implantable cardiac monitors, loop or event recorders have improved the detection and differentiation of arrhythmias. However, existing detection algorithms still fail to achieve high levels of sensitivity and specificity, due to motion artefacts and other ECG-noise caused by muscle myopotential activity [8–10]. While rhythm detection has traditionally focused on the QRS complex, P- and f-wave detection has been lately implemented in some diagnostic algorithms [11–15]. The amplitude of these waves highly depends on the positioning of the electrodes, which in term affects patient convenience
2
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
[16]. In principle, a shorter distance between electrodes, possibly in a single device would yield a better patient compliance. Low-power electronics have allowed the development of numerous small devices that can be used for several days or weeks, especially for patients with AF [7,17,18]. However, hitherto little is known about the optimal size, electrode spacing and placement for such devices. In this study we analysed the ECG signals from different precordial electrode positions of both AF patients and subjects without AF in order to estimate an optimal electrode location that provides the optimal signal quality with the smallest inter-electrode spacing possible.
Methods We included 36 patients (61 ± 16 years old) that presented in the outpatient clinic of our cardiology clinic for evaluation or follow-up of known heart disease, as a representative sample of typical real-world patients, occasionally requiring ambulatory monitoring. Patients were examined at rest in a supine position with 16 electrodes covering the precordial area, in order to minimize muscle myopotentials. Electrodes were placed between the 2nd and 5th inter-costal space (ICS) in a standardized pattern, aligned in columns over the sternum, the bilateral parasternal lines and the left mid-clavicular line. Left papilla mammaria was excluded due to patient compliance and the respective electrode was placed on the 5th ICS at the anterior auxiliary line. The distance between adjacent electrodes was fixed at 4 cm left to right, 3.5 cm superior to inferior and 5.5 cm diagonal. Fig. 1 illustrates the placement, numbering and distances between electrodes. The 16 unipolar ECG signals were recorded against a right foot reference, over a period of 120 seconds and were
digitalized with a sampling frequency of 4800 Hz. An averaged electrical signal was then calculated for all possible bipolar leads (BL) which were obtained by taking the difference between each pair of unipolar leads. From a total of 256 combinations generated, we excluded duplicates and zero values to finally examine 120 different BLs for each patient. Mean values were then obtained for each BL for all patients in the study. We then analysed BL amplitudes for the 19 different distances and selected the BL with the maximum value for each specific distance. BLs with a value more than 91% of the maximum value were defined as clinically equivalent (Tables 2–4). In order to specify the electrodes with the highest amplitude for the shortest distance, we performed differentiation analysis based on incremental change. Examples of signal analysis for the QRS, P and f-wave are presented in (Fig. 2). In order to evaluate the validity of our findings we performed an additional differentiation analysis for every individual patient. We calculated the optimal BLs and their equivalents for every patient and then compared these with the optimal BLs as previously calculated (Table 5). QRS amplitude An averaged beat was obtained in order to remove noise, by applying a Butterworth filter with low-pass frequency of 100 Hz cut-off and notch Filter of 50 Hz. No high-pass filter was applied. The peak-to-peak QRS amplitude was defined as the difference between maximum and minimum amplitude (mV) in a window of 50 ms around the R-peak, which was sufficient to capture the largest QRS amplitude in all patients. P-wave amplitude For patients in sinus rhythm, the P-wave amplitude was calculated from an averaged ECG beat. The peak-to-peak
Fig. 1. Electrode position and distance between adjacent electrodes. Electrodes were covering the precordial area and named as depicted. The arrows illustrate bipolar electrode leads for optimal ECG monitoring based on the registration of ventricular (black), atrial (white) or both activities (grey). ICS = inter-costal space.
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx Table 1 Patient characteristics.
Number of patients, n Age, years Sex, male, n (%) Body weight, kg Height, cm Coronary artery disease, n (%) Diabetes mellitus, n (%) Hypertension, n (%) Hyperlipidemia, n (%) QRS duration, ms Left anterior hemiblock, n (%) Left BBB, n (%) QRS axis deviation Normal:−30° to +90°, n (%) Left:−30° to − 90°, n (%) Right: + 90° to +180°, n (%) LV-EF, % LV–EDD, mm IVSD, mm
Total
Sinus rhythm
Atrial fibrillation
p
36 56 ± 16 29 (81%) 83 ± 18 175 ± 10 26 (72%) 26 (72%) 26 (72%) 23 (64%) 90 ± 15 2 (6%) 1 (3%)
21 (58%) 56 ± 17 15 (72%) 78 ± 17 174 ± 11 15 (71%) 16 (76%) 11 (52%) 10 (47%) 95 ± 21 1 (5%) -
15 (42%) 73 ± 9 b 0.01 14 (93%) 0.20 90 ± 17 0.29 177 ± 8 0.06 11 (73%) 0.65 10 (66%) 0.71 15 (100%) b 0.01 13 (87%) 0.03 87 ± 9 0.16 1 (7%) 0.67 1 (7%) 0.58
25 (70%) 2 (5%) 9 (25%) 54 ± 16 55 ± 11 11 ± 2
14 (66%) 1 (5%) 6 (29%) 60 ± 12 52 ± 8 11 ± 2
11 (73%) 1 (7%) 3 (20%) 49 ± 19 58 ± 13 11 ± 2
0.11
0.06 0.14 0.70
BBB-bundle branch block, EF = ejection fraction, LV = left ventricular, EDD = end-diastolic diameter, IVSD = interventricular septum diameter.
amplitude was defined as the difference between maximum and minimum (mV) in a window that started 270 ms and ended 70 ms before the R peak. f-wave power In AF patients, the P-wave is absent and replaced by a fibrillatory wave (f-wave) which is inconsistent and whose amplitude does not give accurate information. Therefore, the energy of the wave defined as the area under the wave was
3
Table 3 Distance vs. P-wave amplitude in bipolar-leads (BL) from patients with sinus rhythm. Distance BLs (cm) (n)
BL with Equivalent BL max. value (≥ 91%max)
Max value Increment (mV) (mV/cm)
3.50 4.00 5.32 7.00 8.00 8.06 8.73 10.50 10.63 11.24 12.00 12.50 13.20 13.89 15.95 16.00 16.38 17.46 19.14
6–10 9–10 5–10 2–10 9–11 1–10 1–7 2–13 1–11 1–13 5–8 1–8 1–14 5–15 1–15 12–16 9–16 5–16 1–16
0.07 0.07 0.10 0.12 0.09 0.15 0.12 0.15 0.15 0.17 0.10 0.13 0.18 0.13 0.17 0.11 0.12 0.14 0.17
10 12 16 7 8 12 11 4 8 7 4 5 5 4 3 1 1 1 1
5–11 11–12 2–14 12–15, 13–16 4–13 6–16
− 0.02 0.02 0.01 − 0.04 0.98 − 0.05 0.02 0.05 0.02 − 0.09 0.06 0.07 − 0.06 0.02 − 1.31 0.04 0.02 0.01
The BL with the highest positive increment for the shortest distance are marked bold.
calculated instead. In order to isolate atrial activity, ventricular activity (QRS-T complex) was cancelled with a previously described method [10]. Electrical AF activity on the ECG is characterized by higher energy concentration in the band of 4–10 Hz as compared to normal ECGs. Therefore, the remaining signals were filtered using a band pass filter with low and high cut-off frequencies of 3 and 15Hz respectively [19] and the f-wave power was calculated as the averaged energy per time unit (mV 2/s).
Table 2 Distance vs. QRS amplitude in bipolar-leads (BL) from all study patients.
Table 4 Distance vs. f-wave power in bipolar-leads (BL) from patients with atrial fibrillation.
Distance BLs BL with Equivalent BL (cm) (n) max. value (≥91%max)
Distance BLs BL with Equivalent BL (cm) (n) max. value (≥91%max)
Max value Increment (mV2/s) (mV2/s*cm)
3.50 4.00 5.32 7.00 8.00 8.06 8.73 10.50 10.63 11.24 12.00 12.50 13.20 13.89 15.95 16.00 16.38 17.46 19.14
0.55 0.98 0.84 1.25 1.15 1.45 1.39 1.74 1.74 1.94 1.78 1.90 2.05 1.97 2.12 1.69 2.01 1.76 1.56
3.50 4.00 5.32 7.00 8.00 8.06 8.73 10.50 10.63 11.24 12.00 12.50 13.20 13.89 15.95 16.00 16.38 17.46 19.14
10 12 16 7 8 12 11 4 8 7 4 5 5 4 3 1 1 1 1
7–11 15–16 11–15 3–11 14–16 2–11 11–16 3–14 7–16 2–14 13–16 10–16 2–15 6–16 2–16 12–16 9–16 5–16 1–16
3–7, 11–14 14–15 6–11, 10–14 7–14
Max value Increment (mV) (mV/cm)
0.70 0.99 1.15 1.30 1.82 6–14, 7–15, 8–16 1.48 1.88 4–15 1.70 5–14, 6–15 1.90 3–15, 4–16 1.86 2.08 2.17 1–14, 3–16 2.04 5–15 2.18 1–15 2.26 2.02 2.13 2.26 2.33
0.59 0.12 0.09 0.52 − 5.72 0.61 − 0.11 1.59 − 0.08 0.29 0.18 − 0.19 0.20 0.04 − 4.94 0.31 0.12 0.04
The BLs with the highest positive increment for the shortest distance are marked bold.
10 12 16 7 8 12 11 4 8 7 4 5 5 4 3 1 1 1 1
10–13 14–15 10–14 1–9 12–14 2–9 9–14 1–12 3–9 2–12 12–15 9–15 3–12 4–9 4–12 12–16 9–16 5–16 1–16
9–12 13–14 9–13 2–12 13–15 7–9, 10–15 2–13 1–13
0.88 − 0.11 0.24 − 0.10 4.97 − 0.10 0.20 − 0.01 0.31 − 0.20 0.23 0.22 − 0.12 0.07 − 8.55 0.85 − 0.23 − 0.12
The BLs with the highest positive increment for the shortest distance are marked bold.
4
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
in 47% of them and other comorbidities as summarized in Table 1. At the time of the recording, 15 patients were in atrial fibrillation, while the remaining 21 were in sinus rhythm. Hypertension, hyperlipidaemia and older age were more prevalent in AF patients than those in sinus rhythm. There was one patient with left bundle brunch block (QRS 140 ms) and AF, as well as two with left anterior fascicular hemiblock (QRS 95 ± 10 ms); one in each group. Most patients (70%) had a normal QRS-axis and all had a normal P-wave axis, when present. QRS amplitude The QRS amplitude ranged from 0.70 mV in BL 7–11 at a distance of 3.5 cm to 2.33 mV in BL 1–16 at a distance of 19.1 cm. There was a significant correlation between QRSamplitude and distance (r = 0.78, p b 0.001). Differentiation analysis with calculation of the ratio amplitude to distance change revealed that BL 11–16 (at 8.73 cm) had the highest positive increment for the shortest distance. The second best BL was between 14–16 (at 8 cm) with similar position and comparable amplitude to BL 11–16 (1.82 vs. 1.88 mV). There was no equivalent BL for these pairs, since the second highest values were 75% (1.37 mV, BL 12–14) and 83% (1.57 mV, BL 10–15) of the maximum value at 8 and 8.73 cm respectively. Results are shown in Table 2 and Fig. 3A. P-wave amplitude The P-wave amplitude ranged from 0.07 mV in BL 9–10 at a distance of 4.0 cm to 0.18 mV in BL 1–14 at a distance of 13.2 cm. There was a significant correlation between P-wave amplitude and inter-electrode distance (r = 0.60, p b 0.01). Differentiation analysis based on the ratio amplitude to distance change, revealed that BL 1–10 (at 8.06 cm) had the highest positive increment for the shortest distance. There was no equivalent BL for this pairs, since the second highest value was 88% (0.12 mV, BL 2–11) of the maximum value at 8.06 cm. Results are shown in Table 2 and Fig. 3B. f-wave power
Fig. 2. Examples of averaged bipolar-leads (BL) from 3 different patients illustrating the spectrum of four BLs on a plot diagram of amplitude (mV) over time for A. the QRS-wave, B. the zoomed-in P-wave and C. the filtered fibrillatory (f) wave.
Results Patients (n = 36, 61 ± 16 years old) had a preserved ejection fraction (EF 54 ± 16%) with coronary artery disease
The power of the f-wave ranged from 0.55 mV 2/s in BL 10–13 at a distance of 3.5 cm to 2.12 mV 2/s in BL 4–12 at a distance of 16.0 cm. There was a significant correlation between f-wave power and distance (r = 0.79, p b 0.001). Differentiation analysis based on the ratio amplitude to distance change, revealed that BL 2–9 (at 8.06 cm) had the highest positive increment for the shortest distance. There was no equivalent BL for this pairs, since the second highest value was 85% (1.23 mV, BL 1–10) of the maximum value at 8.06 cm. Results are shown in Table 2 and Fig. 3C. Prevalence of optimal distance and optimal BLs The optimal distance, optimal BL or the equivalent optimal BLs for every individual patient are summarized in Table 5. For QRS detection, 67% of the patients had an optimal BL at 8.73 cm and 33% at 8.06 cm. Maximum BL values at 8.73 cm were equivalent or higher to those of shorter distances in 86% (n = 31) of the cases.
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
5
Table 5 Optimal bipolar-leads (BL) for QRS, P and f (fibrillatory) wave detection after differentiation analysis for every individual patient. QRS wave
P wave
Pat. Nr.
Heart Axis
Dist. (cm)
BL of max.
Equivalents (≥ 91%max)
Dist. (cm)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
N N A A N N N N A A N A N A N A A A N N A N N A N N N A N A N A N A A N
8.73 8.73 8.06 8.73 8.73 8.73 8.73 8.73 8.06 8.06 8.73 8.06 8.73 8.06 8.73 8.06 8.06 8.73 8.73 8.73 8.06 8.73 8.73 8.06 8.73 8.73 8.73 8.73 8.73 8.73 8.73 8.06 8.73 8.06 8.06 8.73
11–16 11–16 2–11 9–14 11–16 11–16 10–15 11–16 7–15 2–11 11–16 6–14 11–16 7–13 11–16 2–11 2–11 7–15 11–16 11–16 6–14 11–16 11–16 2–11 10–15 11–16 11–16 9–14 11–16 9–14 11–16 8–14 11–16 6–14 6–14 11–16
10–15
8.00 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 5.32 7.00 8.06 7.00 7.00 7.00 8.06 8.00 8.00 8.06 8.06 8.06 8.06 8.06 7.00 4.00 8.06 7.00 8.00 8.06
5–11 11–16
3–10, 4–11
10–15 8–14
4–11
11–16
10–15
7–15 10–15, 8–10
BL of max. 5–7
Equivalents (≥91%max)
f wave BL of max.
Equivalents (≥ 91%max)
1–3, 9–11 2–9 2–9
1–10 1–10
5–13 2–11
1–10
2–11
2–9
3–10
5–13 1–10 2–11
2–11
5–13
6–12, 7–13
1–10 1–10 5–13
2–11, 5–13 2–9
1–10 7–10 2–10 1–10 6–13 6–13 6–13
2–11, 5–13 5–10 3–11
5–13
2–10
3–10 5–7 9–11 1–10
6–8
12–14 2–11 7–15 6–14
2–11
1–10, 5–13 3–10
6–13 14–15 1–10 2–10
2–11, 5–13
1–10
5–13, 2–11
1–3
Results (distance or equivalent BLs with the highest positive increment for the shortest distance) verifying the initial analysis are marked bold. N = normal heart axis, A = abnormal heart axis.
BL 11–16 was the optimal or equivalent lead in 20 (56%) patients, BL 9–14 in 9 (25%), BL 10–15 in 6 (17%), BL 2–11 in 5 (14%), BL 7–15 in 3 (7%), BL 8–14 in 2 (6%), BL 5–11 and 8–10 in 1 (3%). In 5 patients with optimal BL at 8.06 cm, BL 11–16 could provide equivalent or higher values and therefore could serve as optimal BL in 69% (n = 20 + 5) of the patients. Patients with optimal BL other than 11–16 had significantly higher prevalence of an abnormal heart axis (69% vs. 30%, p = 0.038). For atrial wave detection, 67% of the patients had an optimal BL at 8.06 cm, 17% at 7 cm, 11% at 8 cm and 3% at shorter distance. Maximum BL values at 8.06 cm were always equivalent or higher to those of shorter distances. For P wave detection in patients with sinus rhythm, BL 1–10 was the optimal lead in 11 (52%) patients, BL 5–13 in 7 (33%), BL 2–11 in 8 (38%), BL 6–13 in 4 (19%), BL 2–10 in 3 (14%), BL 9–11 in 2 (10%) and BLs 1–3, 2–9, 3–11, 5–7, 5–10, 7–10 in individuals (5%). In comparison to other BL values, BL 1–10 could always provide equivalent or higher values for 100% of the patients.
For f wave detection in AF patients, BL 2–9, 3–10 and 5–13 were the optimal lead in 3 (20%) patients respectively, BL 2–11 in 2 (13%) and BLs 1–10, 5–7, 6–8, 6–12, 6–14 7–13, 7–15, 14–15 in individuals (5%). In comparison to other BL values, BL 2–9 could always provide equivalent or higher values for all of the patients.
Discussion This study examined the ventricular and atrial rate detection value of all different combinations of bipolar leads covering the precordial area and revealed that there is a strong positive non-linear correlation between signal amplitude and electrode distance. To the best of our knowledge, this is the first study that evaluated all precordial leads in order to identify position of electrodes with optimal detection of QRS complex, P or f-wave at the shortest distance. These results are very useful when placing or designing a device for ECG monitoring in ambulatory patients.
6
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
provide a signal with at least 27% lower amplitude and maximum of 1.3 mV. A better signal, with at least 10% more amplitude than at 8.7 cm (BL 11–16), would only be available with a distance of ≥ 12 cm (BL 13–16), which would mean a 50% longer length. A distance of 8.7 cm could serve the majority (86%) of the patients, most of them (64%) sharing BL 11–16 as the best bipolar lead. A different optimal lead was associated with an abnormal heart axis and could be achieved with a heart-aligned registration. Therefore, a distance of about 8 cm on a heart-aligned axis would provide the optimal length-effective depiction of QRS. To examine atrial activity in patients during sinus rhythm, electrode positioning over the atrial area provides the best P-wave amplitudes. Here, electrodes at 8 cm (BL 1–10) can achieve amplitudes that are comparable to the maximal precordial values (0.15 mV vs. 0.18 mV, respectively) recorded at a 60% longer distance (13.2 cm). Again, a distance of 5 or 7 cm would provide a signal with at least 20% lower amplitude (max. 1.2 mV) than those at 8 cm. Similarly, f-wave power in patients with AF was best recorded over the atrial area. Electrodes at 8 cm (BL 2–9) achieved a relatively good power (1.45 mV 2/s), whereas at 5 or 7 cm distance had almost 10% lower signal. A better signal, with at least 10% more amplitude would be available at 10.5 cm (BL 1–12) or 11.2 cm (BL 2–12) for 31% or 41% longer length respectively. Although there were patients with optimal BL at shorter distances, a distance of 8.06 cm could provide adequately high values for all patients. Despite the variation of optimal leads among individuals, BL 1–10 and BL 2–9 had equivalent or higher values, so that they could serve all patients. In conclusion, a bipolar lead of approximately 8 cm over the atria would be the optimal detection size for atrial activity. Clinical perspective
Fig. 3. Plot diagram with max. values of averaged bipolar-leads (BL) in relation to inter-electrode distance (cm) for A. the QRS-wave, B. the P-wave and C. the fibrillatory (f) wave. QRS- and P-wave are measured as amplitude (mV) and the f-wave is calculated as averaged energy per time unit (mV2/s).
Precordial signal detection Signal detection was dependent on both the electrode positioning as well as the inter-electrode spacing. Electrodes in proximity to the left ventricle (LV) and towards the LV apex provide a good QRS signal with smaller inter-electrode distances. These positions achieve QRS amplitudes of ≥ 1.8 mV with distances of 8–8.7 cm (BL 11–16 or 14–16). In comparison, a distance of 5 or 7 cm, in the size range of many currently available monitor devices, would
Most algorithms for rhythm monitoring are based on the presence, timing and morphology of QRS registered by Holter-ECG or implantable devices. Additionally, some algorithms include the presence or absence of a P-wave, whereas fibrillatory waves are usually absent from rhythm differentiation efforts [11–14]. Conventional Holter-ECG electrodes are placed far apart from one another and tend to register myopotentials from proximal muscle groups. As a result these positions sometimes fail to achieve optimal ECG recording and hinder rhythm recognition [16]. Considering the above, an optimal electrode placement should have the maximal signals for the shortest distance in order to combine convenience and efficacy, leading to improved patient’s compliance and better ECG detection. Previous studies have reported substantially reduced compliance with lead-based extended monitoring because of skin irritation, difficulty of use and lifestyle disruption [20–23]. Newer smaller devices on the contrary, have higher patient compliance and longer uninterrupted wear time [18,20,24]. Patients consider smaller devices more comfortable because they interfere less with everyday life [24]. Therefore, short inter-electrode distance would improve patient compliance.
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
In this study we found that the optimal inter-electrode distance for QRS detection is about 8 cm. For optimal detection of both P and QRS-waves a longer distance is necessary. Data analysis, using a common electrode (#1) and the incremental change analysis showed that BL 1–14 at 13.2 results at max. values for both P (0.18 mV) and QRSwave (1.96 mV). If a multiple-lead configuration were possible, then BL 11–16 (or BL 14–16) and BL 1–10 (or BL 1–11) would be a good combination for the simultaneous optimal registration of ventricular and atrial activities respectively. Alternatively, a modified BL between BL 1–10 with optimal P-wave detection and BL 2–11 for the same distance of 8.06 cm could be considered, but this would result in approximately 20% lower QRS amplitude. However, all these combinations are above the maximum QRS value (1.3 mV) that a common monitor device with maximum size of 7 cm is able to detect. Fig. 1 illustrates the BL configurations for optimal ECG monitoring. These findings supplement the results of previously published studies [16]. Waktare et al. found that inferior and leftward oriented bipolar leads provide the best registration for sinus P-waves, whereas a bipole from the left clavicle to a position below V1 offers the optimal recording of fibrillatory waves. Previous studies have also shown that unipolar V1 lead is of great value for assessing atrial activity [12,25]. Therefore, it is not surprising that a bipolar modification of V1 should be recommended for recording AF activity. This is in accordance with the modified BL proposed in our study (Fig. 1), which carries the benefit of minimal interelectrode distance. Currently available monitor devices are usually 5–7 cm and are placed in a vertical or oblique orientation in the high parasternal region [18,24,26,27]. The results of our study, however, support that future ambulatory ECG systems should be placed on a heart-aligned axis, over the relevant heart-chamber and have a size of ≥ 8 cm. The current study adds to our understanding not only by identifying the axes of optimal signal detection but also by identifying the shortest distance for the maximum signal amplitudes. Limitations The study was based on 36 randomly selected patients of an outpatient clinic presenting for evaluation or follow-up of known heart disease. Patients with AF had some different clinical characteristics, but altogether they represent a realistic sample of real-world patients that need ambulatory ECG monitor to diagnose or monitor arrhythmias. Even though we included some patients with conduction defects and abnormal heart axis, further studies are needed to evaluate the present findings in such patients. Analysis was restricted to supine position at rest only, which is mostly applied in sleep studies. Therefore further evaluation in a regular ambulatory setting is required. The study was not designed to analyse inter- or intrapatient variability. Although we detected an important effect of heart axis on QRS detection, the possible bias of cardiac rotation, respiratory variation and breast size was not included in this study. Therefore, clinical and physical characteristics of each patient should be appropriately considered when inter-
7
preting the findings of this study in order to choose the optimal design or electrode positioning of an ECG monitor device.
Conclusion There is a close correlation between electrode spacing, position and ECG signal-amplitude. The optimal precordial ECG detection with the best amplitude for the shortest distance is achieved at a distance of about 8 cm on a heartaligned axis and over the relevant chamber of interest. These findings are important for design considerations and optimal placement of ambulatory monitoring devices. Acknowledgments The authors would like to thank Dr. Hedi Razavi (St. Jude Medical, USA) for editorial assistance.
References [1] Kroenke K, Arrington ME, Mangelsdorff AD. The prevalence of symptoms in medical outpatients and the adequacy of therapy. Arch Intern Med 1990;150:1685–9. [2] Weber BE, Kapoor WN. Evaluation and outcomes of patients with palpitations. Am J Med 1996;100:138–48. [3] Bollmann A, Mende M, Neugebauer A, Pfeiffer D. Atrial fibrillatory frequency predicts atrial defibrillation threshold and early arrhythmia recurrence in patients undergoing internal cardioversion of persistent atrial fibrillation. Pacing Clin Electrophysiol 2002;25:1179–84. [4] Bollmann A, Husser D, Stridh M, et al. Frequency measures obtained from the surface electrocardiogram in atrial fibrillation research and clinical decision-making. J Cardiovasc Electrophysiol 2003;14(10 Suppl):S154–61. [5] Hindricks G, Piorkowski C, Tanner H, et al. Perception of atrial fibrillation before and after radiofrequency catheter ablation: relevance of asymptomatic arrhythmia recurrence. Circulation 2005;112:307–13. [6] Hindricks G, Pokushalov E, Urban L, et al. Performance of a new leadless implantable cardiac monitor in detecting and quantifying atrial fibrillation: results of the XPECT trial. Circ Arrhythm Electrophysiol 2010;3:141–7. [7] Mittal S, Movsowitz C, Steinberg JS. Ambulatory external electrocardiographic monitoring: focus on atrial fibrillation. J Am Coll Cardiol 2011;58:1741–9. [8] Babaeizadeh S, Gregg RE, Helfenbein ED, Lindauer JM, Zhou SH. Improvements in atrial fibrillation detection for real-time monitoring. J Electrocardiol 2009;42:522–6. [9] Romero I, Grundlehner B, Penders J. Robust beat detector for ambulatory cardiac monitoring. Conf Proc IEEE Eng Med Biol Soc 2009:950–3. [10] Larburu N, Lopetegi T, Romero I. Comparative study of algorithms for atrial fibrillation detection. Comput Cardiol 2011;38:265–8. [11] Slocum J, Sahakian A, Swiryn S. Diagnosis of atrial fibrillation from surface electrocardiograms based on computer-detected atrial activity. J Electrocardiol 1992;25:1–8. [12] Holm M, Pehrson S, Ingemansson M, et al. Non-invasive assessment of the atrial cycle length during atrial fibrillation in man: introducing, validating and illustrating a new ECG method. Cardiovasc Res 1998;38:69–81. [13] Logan B, Healey J. Robust detection of atrial fibrillation for a long term telemonitoring system. IEEE Comput Cardiol 2005;32:619–22. [14] Jiang K, Huang C, Ye SM, Chen H. High accuracy in automatic detection of atrial fibrillation for Holter monitoring. J Zhejiang Univ Sci B 2012;13:751–6.
8
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
[15] Bollmann A, Husser D, Mainardi L, et al. Analysis of surface electrocardiograms in atrial fibrillation: techniques, research, and clinical applications. Europace 2006;8:911–26. [16] Waktare JE, Gallagher MM, Murtagh A, Camm AJ, Malik M. Optimum lead positioning for recording bipolar atrial electrocardiograms during sinus rhythm and atrial fibrillation. Clin Cardiol 1998;21:825–30. [17] Yazicioglu RF, Torfs T, Penders J, et al. Ultra-low-power wearable biopotential sensor nodes. Conf Proc IEEE Eng Med Biol Soc 2009;2009:3205–8. [18] Turakhia MP, Hoang DD, Zimetbaum P, et al. Diagnostic utility of a novel leadless arrhythmia monitoring device. Am J Cardiol 2013;112:520–4. [19] Romero I, Wubbeler G, C E. Common spatial pattern: an improved method for atrial fibrillation wave extraction based on the principal component analysis. Comput Cardiol 2007;34:501–4. [20] Ackermans PA, Solosko TA, Spencer EC, et al. A user-friendly integrated monitor-adhesive patch for long-term ambulatory electrocardiogram monitoring. J Electrocardiol 2012;45:148–53. [21] Zimetbaum P, Goldman A. Ambulatory arrhythmia monitoring: choosing the right device. Circulation 2010;122:1629–36.
[22] Rothman SA, Laughlin JC, Seltzer J, et al. The diagnosis of cardiac arrhythmias: a prospective multi-center randomized study comparing mobile cardiac outpatient telemetry versus standard loop event monitoring. J Cardiovasc Electrophysiol 2007;18:241–7. [23] Vasamreddy CR, Dalal D, Dong J, et al. Symptomatic and asymptomatic atrial fibrillation in patients undergoing radiofrequency catheter ablation. J Cardiovasc Electrophysiol 2006;17:134–9. [24] Rosenberg MA, Samuel M, Thosani A, Zimetbaum PJ. Use of a noninvasive continuous monitoring device in the management of atrial fibrillation: a pilot study. Pacing Clin Electrophysiol 2013;36:328–33. [25] Herzog LR, Marcus FI, Scott WA, Faitelson LH, Ott P, Hahn E. Evaluation of electrocardiographic leads for detection of atrial activity (P wave) in ambulatory ECG monitoring: a pilot study. Pacing Clin Electrophysiol 1992;15:131–4. [26] Chrysostomakis SI, Klapsinos NC, Simantirakis EN, Marketou ME, Kambouraki DC, Vardas PE. Sensing issues related to the clinical use of implantable loop recorders. Europace 2003;5:143–8. [27] Himmrich E, Zellerhoff C, Nebeling D, et al. Where should the implantable ECG event recorder be placed? Z Kardiol 2000;89:289–94.
ScienceDirect Journal of Electrocardiology xx (2014) xxx – xxx www.jecgonline.com
Precordial electrode placement for optimal ECG monitoring: Implications for ambulatory monitor devices and event recorders☆ Sotirios Nedios, MD, a,⁎ Iñaki Romero, MSc, PhD, b Jin-Hong Gerds-Li, MD, c Eckard Fleck, MD, c Charalampos Kriatselis, MD c a
c
Heart Center, University of Leipzig, Department of Electrophysiology, Strümpelstr. 39, Leipzig, Germany b Holst Centre/IMEC, High Tech Campus 31, AE, Eindhoven, Netherlands German Heart Institute Berlin, Department of Internal Medicine and Cardiology, Augustenburgerplatz 1, Berlin, Germany
Abstract
Introduction: Detection of QRS complexes, P-waves and atrial fibrillation f-waves in electrocardiographic (ECG) signals is critical for the correct diagnosis of arrhythmias. We aimed to find the best bipolar lead (BL) with the highest signal amplitude and shortest inter-electrode spacing. Methods: ECG signals (120 seconds) were recorded in 36 patients with 16 precordial electrodes placed in a standardized pattern. An average signal was analysed for each of 120 possible BLs obtained by calculating the difference between pairs of unipolar leads. Peak-to-peak amplitudes of QRS waves (50 ms around R-peak) and P waves (270–70 ms before R-peak) were calculated. For patients with atrial fibrillation, power of the fibrillatory (f) wave was used instead. Maximum values at each distance were considered and differentiation analysis was performed based on incremental changes (amplitude to distance). Results: There was a significant correlation between distance and QRS-amplitude (r = 0.78, p b 0.001), P-wave amplitude (r = 0.60, p b 0.01) and f-wave power (r = 0.79, p b 0.001). The range of values was: QRS-amplitude 0.7–2.33 mV, P-wave amplitude 0.07–0.18 mV, and f-wave power 0.55–2.12 mV 2/s. The maximum value for the shortest distance was on a heart-aligned axis over the left ventricle for the QRS complex (1.9 mV at 8.7 cm) and over the atria for the P-wave (0.98 mV) and f-waves (1.45 mV 2/s at 8 cm, respectively). Conclusion: There is a strong positive correlation between electrode distance and ECG signalamplitude. Distance of 8 cm on a heart-aligned axis and over the relevant heart-chamber provides the highest signal amplitude for the shortest distance. These findings are essential for the design and use of ambulatory monitoring devices. © 2014 Elsevier Inc. All rights reserved.
Keywords:
Ambulatory; Patient monitoring; Electrocardiography (ECG); Atrial fibrillation
Introduction Palpitations and arrhythmias are among the most common symptoms which prompt patients to consult healthcare services [1]. The diagnosis and treatment of these symptoms often proves to be strenuous, frustrating and elusive, despite numerous time-consuming and costly investigations. Cardiac arrhythmias and especially atrial fibrillation (AF) are the most common cause of palpitations [2]. Electrocardiographic (ECG) documentation during symptomatic episodes is
☆
Disclosure: None of the authors have any conflict of interest, financial or otherwise. ⁎ Corresponding author at: Department of Electrophysiology, Heart Center, University of Leipzig, Strümpelstrasse 39, 04289 Leipzig, Germany. E-mail address: [email protected] http://dx.doi.org/10.1016/j.jelectrocard.2014.04.003 0022-0736/© 2014 Elsevier Inc. All rights reserved.
necessary to establish causality, in order to guide the therapeutic strategy [3,4]. This is of paramount importance when it comes to detection of AF, which partially remains asymptomatic, especially after an ablation procedure [5–7]. Ambulatory monitoring with Holter-ECG, telemetry, implantable cardiac monitors, loop or event recorders have improved the detection and differentiation of arrhythmias. However, existing detection algorithms still fail to achieve high levels of sensitivity and specificity, due to motion artefacts and other ECG-noise caused by muscle myopotential activity [8–10]. While rhythm detection has traditionally focused on the QRS complex, P- and f-wave detection has been lately implemented in some diagnostic algorithms [11–15]. The amplitude of these waves highly depends on the positioning of the electrodes, which in term affects patient convenience
2
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
[16]. In principle, a shorter distance between electrodes, possibly in a single device would yield a better patient compliance. Low-power electronics have allowed the development of numerous small devices that can be used for several days or weeks, especially for patients with AF [7,17,18]. However, hitherto little is known about the optimal size, electrode spacing and placement for such devices. In this study we analysed the ECG signals from different precordial electrode positions of both AF patients and subjects without AF in order to estimate an optimal electrode location that provides the optimal signal quality with the smallest inter-electrode spacing possible.
Methods We included 36 patients (61 ± 16 years old) that presented in the outpatient clinic of our cardiology clinic for evaluation or follow-up of known heart disease, as a representative sample of typical real-world patients, occasionally requiring ambulatory monitoring. Patients were examined at rest in a supine position with 16 electrodes covering the precordial area, in order to minimize muscle myopotentials. Electrodes were placed between the 2nd and 5th inter-costal space (ICS) in a standardized pattern, aligned in columns over the sternum, the bilateral parasternal lines and the left mid-clavicular line. Left papilla mammaria was excluded due to patient compliance and the respective electrode was placed on the 5th ICS at the anterior auxiliary line. The distance between adjacent electrodes was fixed at 4 cm left to right, 3.5 cm superior to inferior and 5.5 cm diagonal. Fig. 1 illustrates the placement, numbering and distances between electrodes. The 16 unipolar ECG signals were recorded against a right foot reference, over a period of 120 seconds and were
digitalized with a sampling frequency of 4800 Hz. An averaged electrical signal was then calculated for all possible bipolar leads (BL) which were obtained by taking the difference between each pair of unipolar leads. From a total of 256 combinations generated, we excluded duplicates and zero values to finally examine 120 different BLs for each patient. Mean values were then obtained for each BL for all patients in the study. We then analysed BL amplitudes for the 19 different distances and selected the BL with the maximum value for each specific distance. BLs with a value more than 91% of the maximum value were defined as clinically equivalent (Tables 2–4). In order to specify the electrodes with the highest amplitude for the shortest distance, we performed differentiation analysis based on incremental change. Examples of signal analysis for the QRS, P and f-wave are presented in (Fig. 2). In order to evaluate the validity of our findings we performed an additional differentiation analysis for every individual patient. We calculated the optimal BLs and their equivalents for every patient and then compared these with the optimal BLs as previously calculated (Table 5). QRS amplitude An averaged beat was obtained in order to remove noise, by applying a Butterworth filter with low-pass frequency of 100 Hz cut-off and notch Filter of 50 Hz. No high-pass filter was applied. The peak-to-peak QRS amplitude was defined as the difference between maximum and minimum amplitude (mV) in a window of 50 ms around the R-peak, which was sufficient to capture the largest QRS amplitude in all patients. P-wave amplitude For patients in sinus rhythm, the P-wave amplitude was calculated from an averaged ECG beat. The peak-to-peak
Fig. 1. Electrode position and distance between adjacent electrodes. Electrodes were covering the precordial area and named as depicted. The arrows illustrate bipolar electrode leads for optimal ECG monitoring based on the registration of ventricular (black), atrial (white) or both activities (grey). ICS = inter-costal space.
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx Table 1 Patient characteristics.
Number of patients, n Age, years Sex, male, n (%) Body weight, kg Height, cm Coronary artery disease, n (%) Diabetes mellitus, n (%) Hypertension, n (%) Hyperlipidemia, n (%) QRS duration, ms Left anterior hemiblock, n (%) Left BBB, n (%) QRS axis deviation Normal:−30° to +90°, n (%) Left:−30° to − 90°, n (%) Right: + 90° to +180°, n (%) LV-EF, % LV–EDD, mm IVSD, mm
Total
Sinus rhythm
Atrial fibrillation
p
36 56 ± 16 29 (81%) 83 ± 18 175 ± 10 26 (72%) 26 (72%) 26 (72%) 23 (64%) 90 ± 15 2 (6%) 1 (3%)
21 (58%) 56 ± 17 15 (72%) 78 ± 17 174 ± 11 15 (71%) 16 (76%) 11 (52%) 10 (47%) 95 ± 21 1 (5%) -
15 (42%) 73 ± 9 b 0.01 14 (93%) 0.20 90 ± 17 0.29 177 ± 8 0.06 11 (73%) 0.65 10 (66%) 0.71 15 (100%) b 0.01 13 (87%) 0.03 87 ± 9 0.16 1 (7%) 0.67 1 (7%) 0.58
25 (70%) 2 (5%) 9 (25%) 54 ± 16 55 ± 11 11 ± 2
14 (66%) 1 (5%) 6 (29%) 60 ± 12 52 ± 8 11 ± 2
11 (73%) 1 (7%) 3 (20%) 49 ± 19 58 ± 13 11 ± 2
0.11
0.06 0.14 0.70
BBB-bundle branch block, EF = ejection fraction, LV = left ventricular, EDD = end-diastolic diameter, IVSD = interventricular septum diameter.
amplitude was defined as the difference between maximum and minimum (mV) in a window that started 270 ms and ended 70 ms before the R peak. f-wave power In AF patients, the P-wave is absent and replaced by a fibrillatory wave (f-wave) which is inconsistent and whose amplitude does not give accurate information. Therefore, the energy of the wave defined as the area under the wave was
3
Table 3 Distance vs. P-wave amplitude in bipolar-leads (BL) from patients with sinus rhythm. Distance BLs (cm) (n)
BL with Equivalent BL max. value (≥ 91%max)
Max value Increment (mV) (mV/cm)
3.50 4.00 5.32 7.00 8.00 8.06 8.73 10.50 10.63 11.24 12.00 12.50 13.20 13.89 15.95 16.00 16.38 17.46 19.14
6–10 9–10 5–10 2–10 9–11 1–10 1–7 2–13 1–11 1–13 5–8 1–8 1–14 5–15 1–15 12–16 9–16 5–16 1–16
0.07 0.07 0.10 0.12 0.09 0.15 0.12 0.15 0.15 0.17 0.10 0.13 0.18 0.13 0.17 0.11 0.12 0.14 0.17
10 12 16 7 8 12 11 4 8 7 4 5 5 4 3 1 1 1 1
5–11 11–12 2–14 12–15, 13–16 4–13 6–16
− 0.02 0.02 0.01 − 0.04 0.98 − 0.05 0.02 0.05 0.02 − 0.09 0.06 0.07 − 0.06 0.02 − 1.31 0.04 0.02 0.01
The BL with the highest positive increment for the shortest distance are marked bold.
calculated instead. In order to isolate atrial activity, ventricular activity (QRS-T complex) was cancelled with a previously described method [10]. Electrical AF activity on the ECG is characterized by higher energy concentration in the band of 4–10 Hz as compared to normal ECGs. Therefore, the remaining signals were filtered using a band pass filter with low and high cut-off frequencies of 3 and 15Hz respectively [19] and the f-wave power was calculated as the averaged energy per time unit (mV 2/s).
Table 2 Distance vs. QRS amplitude in bipolar-leads (BL) from all study patients.
Table 4 Distance vs. f-wave power in bipolar-leads (BL) from patients with atrial fibrillation.
Distance BLs BL with Equivalent BL (cm) (n) max. value (≥91%max)
Distance BLs BL with Equivalent BL (cm) (n) max. value (≥91%max)
Max value Increment (mV2/s) (mV2/s*cm)
3.50 4.00 5.32 7.00 8.00 8.06 8.73 10.50 10.63 11.24 12.00 12.50 13.20 13.89 15.95 16.00 16.38 17.46 19.14
0.55 0.98 0.84 1.25 1.15 1.45 1.39 1.74 1.74 1.94 1.78 1.90 2.05 1.97 2.12 1.69 2.01 1.76 1.56
3.50 4.00 5.32 7.00 8.00 8.06 8.73 10.50 10.63 11.24 12.00 12.50 13.20 13.89 15.95 16.00 16.38 17.46 19.14
10 12 16 7 8 12 11 4 8 7 4 5 5 4 3 1 1 1 1
7–11 15–16 11–15 3–11 14–16 2–11 11–16 3–14 7–16 2–14 13–16 10–16 2–15 6–16 2–16 12–16 9–16 5–16 1–16
3–7, 11–14 14–15 6–11, 10–14 7–14
Max value Increment (mV) (mV/cm)
0.70 0.99 1.15 1.30 1.82 6–14, 7–15, 8–16 1.48 1.88 4–15 1.70 5–14, 6–15 1.90 3–15, 4–16 1.86 2.08 2.17 1–14, 3–16 2.04 5–15 2.18 1–15 2.26 2.02 2.13 2.26 2.33
0.59 0.12 0.09 0.52 − 5.72 0.61 − 0.11 1.59 − 0.08 0.29 0.18 − 0.19 0.20 0.04 − 4.94 0.31 0.12 0.04
The BLs with the highest positive increment for the shortest distance are marked bold.
10 12 16 7 8 12 11 4 8 7 4 5 5 4 3 1 1 1 1
10–13 14–15 10–14 1–9 12–14 2–9 9–14 1–12 3–9 2–12 12–15 9–15 3–12 4–9 4–12 12–16 9–16 5–16 1–16
9–12 13–14 9–13 2–12 13–15 7–9, 10–15 2–13 1–13
0.88 − 0.11 0.24 − 0.10 4.97 − 0.10 0.20 − 0.01 0.31 − 0.20 0.23 0.22 − 0.12 0.07 − 8.55 0.85 − 0.23 − 0.12
The BLs with the highest positive increment for the shortest distance are marked bold.
4
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
in 47% of them and other comorbidities as summarized in Table 1. At the time of the recording, 15 patients were in atrial fibrillation, while the remaining 21 were in sinus rhythm. Hypertension, hyperlipidaemia and older age were more prevalent in AF patients than those in sinus rhythm. There was one patient with left bundle brunch block (QRS 140 ms) and AF, as well as two with left anterior fascicular hemiblock (QRS 95 ± 10 ms); one in each group. Most patients (70%) had a normal QRS-axis and all had a normal P-wave axis, when present. QRS amplitude The QRS amplitude ranged from 0.70 mV in BL 7–11 at a distance of 3.5 cm to 2.33 mV in BL 1–16 at a distance of 19.1 cm. There was a significant correlation between QRSamplitude and distance (r = 0.78, p b 0.001). Differentiation analysis with calculation of the ratio amplitude to distance change revealed that BL 11–16 (at 8.73 cm) had the highest positive increment for the shortest distance. The second best BL was between 14–16 (at 8 cm) with similar position and comparable amplitude to BL 11–16 (1.82 vs. 1.88 mV). There was no equivalent BL for these pairs, since the second highest values were 75% (1.37 mV, BL 12–14) and 83% (1.57 mV, BL 10–15) of the maximum value at 8 and 8.73 cm respectively. Results are shown in Table 2 and Fig. 3A. P-wave amplitude The P-wave amplitude ranged from 0.07 mV in BL 9–10 at a distance of 4.0 cm to 0.18 mV in BL 1–14 at a distance of 13.2 cm. There was a significant correlation between P-wave amplitude and inter-electrode distance (r = 0.60, p b 0.01). Differentiation analysis based on the ratio amplitude to distance change, revealed that BL 1–10 (at 8.06 cm) had the highest positive increment for the shortest distance. There was no equivalent BL for this pairs, since the second highest value was 88% (0.12 mV, BL 2–11) of the maximum value at 8.06 cm. Results are shown in Table 2 and Fig. 3B. f-wave power
Fig. 2. Examples of averaged bipolar-leads (BL) from 3 different patients illustrating the spectrum of four BLs on a plot diagram of amplitude (mV) over time for A. the QRS-wave, B. the zoomed-in P-wave and C. the filtered fibrillatory (f) wave.
Results Patients (n = 36, 61 ± 16 years old) had a preserved ejection fraction (EF 54 ± 16%) with coronary artery disease
The power of the f-wave ranged from 0.55 mV 2/s in BL 10–13 at a distance of 3.5 cm to 2.12 mV 2/s in BL 4–12 at a distance of 16.0 cm. There was a significant correlation between f-wave power and distance (r = 0.79, p b 0.001). Differentiation analysis based on the ratio amplitude to distance change, revealed that BL 2–9 (at 8.06 cm) had the highest positive increment for the shortest distance. There was no equivalent BL for this pairs, since the second highest value was 85% (1.23 mV, BL 1–10) of the maximum value at 8.06 cm. Results are shown in Table 2 and Fig. 3C. Prevalence of optimal distance and optimal BLs The optimal distance, optimal BL or the equivalent optimal BLs for every individual patient are summarized in Table 5. For QRS detection, 67% of the patients had an optimal BL at 8.73 cm and 33% at 8.06 cm. Maximum BL values at 8.73 cm were equivalent or higher to those of shorter distances in 86% (n = 31) of the cases.
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
5
Table 5 Optimal bipolar-leads (BL) for QRS, P and f (fibrillatory) wave detection after differentiation analysis for every individual patient. QRS wave
P wave
Pat. Nr.
Heart Axis
Dist. (cm)
BL of max.
Equivalents (≥ 91%max)
Dist. (cm)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
N N A A N N N N A A N A N A N A A A N N A N N A N N N A N A N A N A A N
8.73 8.73 8.06 8.73 8.73 8.73 8.73 8.73 8.06 8.06 8.73 8.06 8.73 8.06 8.73 8.06 8.06 8.73 8.73 8.73 8.06 8.73 8.73 8.06 8.73 8.73 8.73 8.73 8.73 8.73 8.73 8.06 8.73 8.06 8.06 8.73
11–16 11–16 2–11 9–14 11–16 11–16 10–15 11–16 7–15 2–11 11–16 6–14 11–16 7–13 11–16 2–11 2–11 7–15 11–16 11–16 6–14 11–16 11–16 2–11 10–15 11–16 11–16 9–14 11–16 9–14 11–16 8–14 11–16 6–14 6–14 11–16
10–15
8.00 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 8.06 5.32 7.00 8.06 7.00 7.00 7.00 8.06 8.00 8.00 8.06 8.06 8.06 8.06 8.06 7.00 4.00 8.06 7.00 8.00 8.06
5–11 11–16
3–10, 4–11
10–15 8–14
4–11
11–16
10–15
7–15 10–15, 8–10
BL of max. 5–7
Equivalents (≥91%max)
f wave BL of max.
Equivalents (≥ 91%max)
1–3, 9–11 2–9 2–9
1–10 1–10
5–13 2–11
1–10
2–11
2–9
3–10
5–13 1–10 2–11
2–11
5–13
6–12, 7–13
1–10 1–10 5–13
2–11, 5–13 2–9
1–10 7–10 2–10 1–10 6–13 6–13 6–13
2–11, 5–13 5–10 3–11
5–13
2–10
3–10 5–7 9–11 1–10
6–8
12–14 2–11 7–15 6–14
2–11
1–10, 5–13 3–10
6–13 14–15 1–10 2–10
2–11, 5–13
1–10
5–13, 2–11
1–3
Results (distance or equivalent BLs with the highest positive increment for the shortest distance) verifying the initial analysis are marked bold. N = normal heart axis, A = abnormal heart axis.
BL 11–16 was the optimal or equivalent lead in 20 (56%) patients, BL 9–14 in 9 (25%), BL 10–15 in 6 (17%), BL 2–11 in 5 (14%), BL 7–15 in 3 (7%), BL 8–14 in 2 (6%), BL 5–11 and 8–10 in 1 (3%). In 5 patients with optimal BL at 8.06 cm, BL 11–16 could provide equivalent or higher values and therefore could serve as optimal BL in 69% (n = 20 + 5) of the patients. Patients with optimal BL other than 11–16 had significantly higher prevalence of an abnormal heart axis (69% vs. 30%, p = 0.038). For atrial wave detection, 67% of the patients had an optimal BL at 8.06 cm, 17% at 7 cm, 11% at 8 cm and 3% at shorter distance. Maximum BL values at 8.06 cm were always equivalent or higher to those of shorter distances. For P wave detection in patients with sinus rhythm, BL 1–10 was the optimal lead in 11 (52%) patients, BL 5–13 in 7 (33%), BL 2–11 in 8 (38%), BL 6–13 in 4 (19%), BL 2–10 in 3 (14%), BL 9–11 in 2 (10%) and BLs 1–3, 2–9, 3–11, 5–7, 5–10, 7–10 in individuals (5%). In comparison to other BL values, BL 1–10 could always provide equivalent or higher values for 100% of the patients.
For f wave detection in AF patients, BL 2–9, 3–10 and 5–13 were the optimal lead in 3 (20%) patients respectively, BL 2–11 in 2 (13%) and BLs 1–10, 5–7, 6–8, 6–12, 6–14 7–13, 7–15, 14–15 in individuals (5%). In comparison to other BL values, BL 2–9 could always provide equivalent or higher values for all of the patients.
Discussion This study examined the ventricular and atrial rate detection value of all different combinations of bipolar leads covering the precordial area and revealed that there is a strong positive non-linear correlation between signal amplitude and electrode distance. To the best of our knowledge, this is the first study that evaluated all precordial leads in order to identify position of electrodes with optimal detection of QRS complex, P or f-wave at the shortest distance. These results are very useful when placing or designing a device for ECG monitoring in ambulatory patients.
6
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
provide a signal with at least 27% lower amplitude and maximum of 1.3 mV. A better signal, with at least 10% more amplitude than at 8.7 cm (BL 11–16), would only be available with a distance of ≥ 12 cm (BL 13–16), which would mean a 50% longer length. A distance of 8.7 cm could serve the majority (86%) of the patients, most of them (64%) sharing BL 11–16 as the best bipolar lead. A different optimal lead was associated with an abnormal heart axis and could be achieved with a heart-aligned registration. Therefore, a distance of about 8 cm on a heart-aligned axis would provide the optimal length-effective depiction of QRS. To examine atrial activity in patients during sinus rhythm, electrode positioning over the atrial area provides the best P-wave amplitudes. Here, electrodes at 8 cm (BL 1–10) can achieve amplitudes that are comparable to the maximal precordial values (0.15 mV vs. 0.18 mV, respectively) recorded at a 60% longer distance (13.2 cm). Again, a distance of 5 or 7 cm would provide a signal with at least 20% lower amplitude (max. 1.2 mV) than those at 8 cm. Similarly, f-wave power in patients with AF was best recorded over the atrial area. Electrodes at 8 cm (BL 2–9) achieved a relatively good power (1.45 mV 2/s), whereas at 5 or 7 cm distance had almost 10% lower signal. A better signal, with at least 10% more amplitude would be available at 10.5 cm (BL 1–12) or 11.2 cm (BL 2–12) for 31% or 41% longer length respectively. Although there were patients with optimal BL at shorter distances, a distance of 8.06 cm could provide adequately high values for all patients. Despite the variation of optimal leads among individuals, BL 1–10 and BL 2–9 had equivalent or higher values, so that they could serve all patients. In conclusion, a bipolar lead of approximately 8 cm over the atria would be the optimal detection size for atrial activity. Clinical perspective
Fig. 3. Plot diagram with max. values of averaged bipolar-leads (BL) in relation to inter-electrode distance (cm) for A. the QRS-wave, B. the P-wave and C. the fibrillatory (f) wave. QRS- and P-wave are measured as amplitude (mV) and the f-wave is calculated as averaged energy per time unit (mV2/s).
Precordial signal detection Signal detection was dependent on both the electrode positioning as well as the inter-electrode spacing. Electrodes in proximity to the left ventricle (LV) and towards the LV apex provide a good QRS signal with smaller inter-electrode distances. These positions achieve QRS amplitudes of ≥ 1.8 mV with distances of 8–8.7 cm (BL 11–16 or 14–16). In comparison, a distance of 5 or 7 cm, in the size range of many currently available monitor devices, would
Most algorithms for rhythm monitoring are based on the presence, timing and morphology of QRS registered by Holter-ECG or implantable devices. Additionally, some algorithms include the presence or absence of a P-wave, whereas fibrillatory waves are usually absent from rhythm differentiation efforts [11–14]. Conventional Holter-ECG electrodes are placed far apart from one another and tend to register myopotentials from proximal muscle groups. As a result these positions sometimes fail to achieve optimal ECG recording and hinder rhythm recognition [16]. Considering the above, an optimal electrode placement should have the maximal signals for the shortest distance in order to combine convenience and efficacy, leading to improved patient’s compliance and better ECG detection. Previous studies have reported substantially reduced compliance with lead-based extended monitoring because of skin irritation, difficulty of use and lifestyle disruption [20–23]. Newer smaller devices on the contrary, have higher patient compliance and longer uninterrupted wear time [18,20,24]. Patients consider smaller devices more comfortable because they interfere less with everyday life [24]. Therefore, short inter-electrode distance would improve patient compliance.
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
In this study we found that the optimal inter-electrode distance for QRS detection is about 8 cm. For optimal detection of both P and QRS-waves a longer distance is necessary. Data analysis, using a common electrode (#1) and the incremental change analysis showed that BL 1–14 at 13.2 results at max. values for both P (0.18 mV) and QRSwave (1.96 mV). If a multiple-lead configuration were possible, then BL 11–16 (or BL 14–16) and BL 1–10 (or BL 1–11) would be a good combination for the simultaneous optimal registration of ventricular and atrial activities respectively. Alternatively, a modified BL between BL 1–10 with optimal P-wave detection and BL 2–11 for the same distance of 8.06 cm could be considered, but this would result in approximately 20% lower QRS amplitude. However, all these combinations are above the maximum QRS value (1.3 mV) that a common monitor device with maximum size of 7 cm is able to detect. Fig. 1 illustrates the BL configurations for optimal ECG monitoring. These findings supplement the results of previously published studies [16]. Waktare et al. found that inferior and leftward oriented bipolar leads provide the best registration for sinus P-waves, whereas a bipole from the left clavicle to a position below V1 offers the optimal recording of fibrillatory waves. Previous studies have also shown that unipolar V1 lead is of great value for assessing atrial activity [12,25]. Therefore, it is not surprising that a bipolar modification of V1 should be recommended for recording AF activity. This is in accordance with the modified BL proposed in our study (Fig. 1), which carries the benefit of minimal interelectrode distance. Currently available monitor devices are usually 5–7 cm and are placed in a vertical or oblique orientation in the high parasternal region [18,24,26,27]. The results of our study, however, support that future ambulatory ECG systems should be placed on a heart-aligned axis, over the relevant heart-chamber and have a size of ≥ 8 cm. The current study adds to our understanding not only by identifying the axes of optimal signal detection but also by identifying the shortest distance for the maximum signal amplitudes. Limitations The study was based on 36 randomly selected patients of an outpatient clinic presenting for evaluation or follow-up of known heart disease. Patients with AF had some different clinical characteristics, but altogether they represent a realistic sample of real-world patients that need ambulatory ECG monitor to diagnose or monitor arrhythmias. Even though we included some patients with conduction defects and abnormal heart axis, further studies are needed to evaluate the present findings in such patients. Analysis was restricted to supine position at rest only, which is mostly applied in sleep studies. Therefore further evaluation in a regular ambulatory setting is required. The study was not designed to analyse inter- or intrapatient variability. Although we detected an important effect of heart axis on QRS detection, the possible bias of cardiac rotation, respiratory variation and breast size was not included in this study. Therefore, clinical and physical characteristics of each patient should be appropriately considered when inter-
7
preting the findings of this study in order to choose the optimal design or electrode positioning of an ECG monitor device.
Conclusion There is a close correlation between electrode spacing, position and ECG signal-amplitude. The optimal precordial ECG detection with the best amplitude for the shortest distance is achieved at a distance of about 8 cm on a heartaligned axis and over the relevant chamber of interest. These findings are important for design considerations and optimal placement of ambulatory monitoring devices. Acknowledgments The authors would like to thank Dr. Hedi Razavi (St. Jude Medical, USA) for editorial assistance.
References [1] Kroenke K, Arrington ME, Mangelsdorff AD. The prevalence of symptoms in medical outpatients and the adequacy of therapy. Arch Intern Med 1990;150:1685–9. [2] Weber BE, Kapoor WN. Evaluation and outcomes of patients with palpitations. Am J Med 1996;100:138–48. [3] Bollmann A, Mende M, Neugebauer A, Pfeiffer D. Atrial fibrillatory frequency predicts atrial defibrillation threshold and early arrhythmia recurrence in patients undergoing internal cardioversion of persistent atrial fibrillation. Pacing Clin Electrophysiol 2002;25:1179–84. [4] Bollmann A, Husser D, Stridh M, et al. Frequency measures obtained from the surface electrocardiogram in atrial fibrillation research and clinical decision-making. J Cardiovasc Electrophysiol 2003;14(10 Suppl):S154–61. [5] Hindricks G, Piorkowski C, Tanner H, et al. Perception of atrial fibrillation before and after radiofrequency catheter ablation: relevance of asymptomatic arrhythmia recurrence. Circulation 2005;112:307–13. [6] Hindricks G, Pokushalov E, Urban L, et al. Performance of a new leadless implantable cardiac monitor in detecting and quantifying atrial fibrillation: results of the XPECT trial. Circ Arrhythm Electrophysiol 2010;3:141–7. [7] Mittal S, Movsowitz C, Steinberg JS. Ambulatory external electrocardiographic monitoring: focus on atrial fibrillation. J Am Coll Cardiol 2011;58:1741–9. [8] Babaeizadeh S, Gregg RE, Helfenbein ED, Lindauer JM, Zhou SH. Improvements in atrial fibrillation detection for real-time monitoring. J Electrocardiol 2009;42:522–6. [9] Romero I, Grundlehner B, Penders J. Robust beat detector for ambulatory cardiac monitoring. Conf Proc IEEE Eng Med Biol Soc 2009:950–3. [10] Larburu N, Lopetegi T, Romero I. Comparative study of algorithms for atrial fibrillation detection. Comput Cardiol 2011;38:265–8. [11] Slocum J, Sahakian A, Swiryn S. Diagnosis of atrial fibrillation from surface electrocardiograms based on computer-detected atrial activity. J Electrocardiol 1992;25:1–8. [12] Holm M, Pehrson S, Ingemansson M, et al. Non-invasive assessment of the atrial cycle length during atrial fibrillation in man: introducing, validating and illustrating a new ECG method. Cardiovasc Res 1998;38:69–81. [13] Logan B, Healey J. Robust detection of atrial fibrillation for a long term telemonitoring system. IEEE Comput Cardiol 2005;32:619–22. [14] Jiang K, Huang C, Ye SM, Chen H. High accuracy in automatic detection of atrial fibrillation for Holter monitoring. J Zhejiang Univ Sci B 2012;13:751–6.
8
S. Nedios et al. / Journal of Electrocardiology xx (2014) xxx–xxx
[15] Bollmann A, Husser D, Mainardi L, et al. Analysis of surface electrocardiograms in atrial fibrillation: techniques, research, and clinical applications. Europace 2006;8:911–26. [16] Waktare JE, Gallagher MM, Murtagh A, Camm AJ, Malik M. Optimum lead positioning for recording bipolar atrial electrocardiograms during sinus rhythm and atrial fibrillation. Clin Cardiol 1998;21:825–30. [17] Yazicioglu RF, Torfs T, Penders J, et al. Ultra-low-power wearable biopotential sensor nodes. Conf Proc IEEE Eng Med Biol Soc 2009;2009:3205–8. [18] Turakhia MP, Hoang DD, Zimetbaum P, et al. Diagnostic utility of a novel leadless arrhythmia monitoring device. Am J Cardiol 2013;112:520–4. [19] Romero I, Wubbeler G, C E. Common spatial pattern: an improved method for atrial fibrillation wave extraction based on the principal component analysis. Comput Cardiol 2007;34:501–4. [20] Ackermans PA, Solosko TA, Spencer EC, et al. A user-friendly integrated monitor-adhesive patch for long-term ambulatory electrocardiogram monitoring. J Electrocardiol 2012;45:148–53. [21] Zimetbaum P, Goldman A. Ambulatory arrhythmia monitoring: choosing the right device. Circulation 2010;122:1629–36.
[22] Rothman SA, Laughlin JC, Seltzer J, et al. The diagnosis of cardiac arrhythmias: a prospective multi-center randomized study comparing mobile cardiac outpatient telemetry versus standard loop event monitoring. J Cardiovasc Electrophysiol 2007;18:241–7. [23] Vasamreddy CR, Dalal D, Dong J, et al. Symptomatic and asymptomatic atrial fibrillation in patients undergoing radiofrequency catheter ablation. J Cardiovasc Electrophysiol 2006;17:134–9. [24] Rosenberg MA, Samuel M, Thosani A, Zimetbaum PJ. Use of a noninvasive continuous monitoring device in the management of atrial fibrillation: a pilot study. Pacing Clin Electrophysiol 2013;36:328–33. [25] Herzog LR, Marcus FI, Scott WA, Faitelson LH, Ott P, Hahn E. Evaluation of electrocardiographic leads for detection of atrial activity (P wave) in ambulatory ECG monitoring: a pilot study. Pacing Clin Electrophysiol 1992;15:131–4. [26] Chrysostomakis SI, Klapsinos NC, Simantirakis EN, Marketou ME, Kambouraki DC, Vardas PE. Sensing issues related to the clinical use of implantable loop recorders. Europace 2003;5:143–8. [27] Himmrich E, Zellerhoff C, Nebeling D, et al. Where should the implantable ECG event recorder be placed? Z Kardiol 2000;89:289–94.
