- Email: [email protected]
Electrocardiographic abnormalities in patients with subarachnoid hemorrhage and normal adults: A comparison study
Journal of Electrocardiology Vol. 37 Supplement 2004
Poster Session I
Augmentation of the Amplitude of QRS Complexes in Patients With Heart Failure Treated With Diuresis for 48 Hours
patients with AN should await abatement of A. The above can be extrapolated to the patients with end-stage renal failure undergoing hemodialysis, and congestive heart failure, who also reveal attenuation of the ECG QRS potentials at the stage of poor compensation of their edematous state.
John E. Madias, MD,* Jessica Song, Pharm D,† C. Michael White, Pharm D,† James S. Kalus, Pharm D,† and Jeffrey Kluger, MD,† Mount Sinai School of Medicine,* and University of Connecticut,† Elmhurst, NY, and Hartford CT USA
Genesis of the Normal T-U Wave: A Slice Model of the Left Ventricular Myocardium
Augmentation in the amplitude of ECG QRS complexes (AUQRS) has been described in patients treated for heart failure (HF), but the underlying mechanism has remained enigmatic. We assessed the effect of diuresis-based fluid loss in patients treated for HF on the AU-QRS. Twenty-one patients aged 70.5⫾12.7 years, 13 with ischemic and 8 with non-ischemic cardiomyopathy, received diuresis in the hospital for exacerbated HF, and had standard ECGs recorded prior to the initiation of therapy, and at 24, and 48 hours. Percent change (%⌬) over the course of observation in the sums of the amplitude of QRS complexes from 12 leads (⌺QRS12), 6 limb leads (⌺QRS6), and leads 1⫹2 (⌺QRS2) in mm were correlated with net fluid loss, corrected for admission weight in mL/Kg. Fluid loss amounted to 3,204.9⫾1,399.5 mL, and ⌺QRS12 was 160.9⫾42.3 mm before and 170.0⫾50.7 mm after diuresis (p ⫽ 0. 024). %⌬ in ⌺QRS12, ⌺QRS6, and ⌺QRS2 correlated well with the net fluid loss (r ⫽ ⫺0.70, ⫺0.82, ⫺0.61, and p ⫽ 0.002, 0.0005, 0.001), correspondingly. AU-QRS correlates well with net fluid loss from shortterm diuresis in patients with HF, and can be used as an easily obtainable and universally available bedside index of the net fluid loss, experienced by bed-ridden patients with exacerbated HF. undergoing therapy.
Henk J. Ritsema Van Eck, Jan A. Kors, and Gerard Van Herpen, Department of Medical Informatics, Erasmus MC, Rotterdam, The Netherlands The U wave has always lived in the shadow of the preceding, bigger T wave. Its clinical relevance has been recognized but a satisfactory explanation of its origin is still outstanding. Earlier we presented a simple digital model, consisting of a single string of 12 myocardial cells. We have extended the model to a more realistic one in the form of a slice of Left Ventricle (LV), 12 cell layers thick and containing 1328 cells. To each layer an Action Potential (AP) is assigned with shape and duration (APD90) according to published data. The mid-myocardial M-cell layers have appropriately prolonged AP’s. The AP’s have slowly tapering tails, as evidenced from literature data. The depolarization sequence follows from the endocardial Purkinje activation pattern. The potential differences between the AP’s produce timevarying electrical sources between each of the 1328 myocardial cells and its neighbors. The resultant dipole vector is assigned to the cell center. Each cell dipole contributes to the potential in an arbitrary point P in or on the body. The specific lead vector linking P to that particular cell determines the strength of this contribution. The course of the potentials at P is then calculated as the sum of the contributions from all 1328 cells, as they vary with time. This is done for various distances of P to the epicardium and for various angles with respect to the LV axis. The repolarization waves constructed in this way reproduce the natural aspects of a T wave followed by a U. The size and morphology of the T-U complex depend on the position (distance and angle) of P with respect to the myocardium and on the transmural dispersion of the repolarization. Clear-cut U waves are produced with transmural dispersions of as little as 36 ms (APD90). We conclude that T and U are the resultant of one and the same repolarization process and that they form a single, continuous T-U complex. The creation of normal U waves is conditional on the presence of small voltage differences between the tail ends of the AP’s. Further investigation is necessary to elucidate the genesis of pathological T-U morphologies.
Anasarca Peripheral Edema of Varying Etiology Conceals the ECG Diagnosis of Left Ventricular Hypertrophy John E. Madias, MD, Mount Sinai School of Medicine, New York, NY USA The diagnosis of left ventricular hypertrophy (LVH) is in large part based on measurements of R-, and S-waves from various ECG leads; since anasarca peripheral edema (AN) leads to attenuation of the amplitude of QRS complexes, it could interfere with the diagnosis of LVH. ECG measurements and qualitative assessment (“Cornell”, “Sokolow-Lyon”, and “Romhilt-Estes”) were carried out in 14 patients with LVH (6 patients with AN and 8 patients who did not gain weight during their hospitalization, and served as “controls”). Patients with AN were evaluated on admission, and on the days of half of peak weight (HF-W) gain, and peak weight (P-W) gain. The “controls” were evaluated on admission and at discharge. While LVH characterization remained unaltered for the “controls”, the patients with AN revealed statistically significant drop in all ECG measurements on the days of HF-W gain (p ⫽ 0.016 – 0.0005), and P-W gain (p ⫽ 0.02 – 0.0005), on which the diagnosis of LVH could not be anymore made. The ECG-based diagnosis of LVH is interfered with by AN, and absence of LVH in
Electrocardiographic Abnormalities in Patients with Subarachnoid Hemorrhage and Normal Adults: A Comparison Study Claire E. Sommargren, Robert Warner, Jonathan G. Zaroff, Nader M. Banki, Alexander Kopelnik, Avinash A. Kothavale,
42
Poster Session I
43
Poyee P. Tung, and Barbara J. Drew, University of California, San Francisco, CA
before and after applying the adaptive filter. A percent improvement was calculated for each data set. These results are shown in the table.
A wide variety of electrocardiographic (ECG) abnormalities have been reported in patients with subarachnoid hemorrhage (SAH). The purpose of this study was to compare the ECG in patients with SAH to those in normal persons in order to identify differences in ECG characteristics between the two groups, and to determine which waveform and interval abnormalities occur more frequently in SAH. Method: A standard ECG was recorded as soon as possible after admission. Computer-assisted measurements of ST segment deviation and QTc intervals were made with the Mortara ST Review Station (Milwaukee, WI). Results: ECGs of 200 patients with SAH and 30 patients in a normal control group were analyzed. Of 237 quantitative ECG measurements, 91 were significantly different between the two groups. Mean PR interval duration was shorter in the SAH group (p⫽.000). Mean R wave amplitude was higher in lead I (p⫽.001) and lead V2 (p⫽.016) in the normal group, and higher in V6 (p⫽.003) in the SAH group. S wave amplitudes were greater in SAH patients across all six precordial leads (p⫽.003–.027). T wave axis differed significantly between the two groups (p⫽.001), and. T wave inversion was found to occur in 44 (22%) of SAH patients but only 1 (3%) of the normal control group (p⫽.013). In addition, 48 (24%) patients in the SAH group met ECG criteria for left ventricular hypertrophy, while none did so in the control group (p⫽.011). Conclusions: There are significant ECG differences between normal patients and those with SAH. It is further concluded that T wave inversion and ECG evidence of left ventricular hypertrophy occur significantly more frequently in patients with SAH than in normal patients.
Percent Improvement After Filtering
Using Motion Sensors to Reduce Motion Artifact in the ECG David A. Tong, Keith A. Bartels, and Kevin S. Honeyager, Southwest Research Institute, San Antonio, TX, USA Motion artifact is an unsolved problem in electrocardiography. Motion artifact may produce large amplitude signals in the ECG that may be misinterpreted by clinicians and automated systems resulting in misdiagnosis, prolonged procedure duration, and delayed or inappropriate treatment decisions. The hypothesis being investigated in this project is that motion artifact can be reduced by using electrode motion as a noise reference in an adaptive filter. The initial results of a novel approach to reducing motion artifact in the ECG using motion sensors attached to the ECG electrodes and adaptive filtering are presented. To measure electrode motion, two custom-designed motion sensors were developed and attached to the ECG electrode. One sensor utilized a two axis anisotrophic magnetoresistive (AMR) sensor and the other sensor utilized a three axis accelerometer (ACC). In an iterative fashion, both sensors were attached to the right arm electrode where motion was induced by (1) pushing directly on the ECG electrode, (2) pushing on the skin around the electrode, and (3) pulling on the ECG lead wire. The motion sensor signals and ECG leads I and II were recorded from eight subjects at 500 Hz. An LMS adaptive filter was implemented in Matlab and was used to post-process the electrode motion and ECG data. System performance was assessed by calculating two measures of the noise in the ECG (the L2 Norm and the MaxMin statistic)
AMR Sensor Noise Type All Data Sets Push Electrode Pull Lead Push Skin
ACC Sensor
L2 Norm
Max Min
L2 Norm
Max Min
67.6% 70.6% 45.2% 77.3%
37.7% 48.7% 24.5% 51.0%
83.6% 78.6% 91.3% 89.6%
58.5% 50.7% 49.2% 67.5%
As shown in the table, the data support the hypothesis that electrode motion can be used in an adaptive filter to reduce motion artifact in the ECG. The ACC sensor out performed the AMR sensor because the ACC sensor provided an additional axis of measurement and the relationship between the ACC signal and the motion artifact noise is better modeled by a linear time-varying system.
What Are the Implications for Using Modified (Mason-Likar) Exercise Lead System in Research? Shu-Fen Wung, PhD, RN, Amelia Sieger, BS, Monique Leon, BS, Jessica Redondo, BS, Vincent Sorrell, MD, and Steven Goldman, MD, University of Arizona, Southern Arizona VA Health Care System Tucson, Arizona, USA Researchers often use the Mason-Likar exercise lead system to maintain stability of the waveforms for continuous ECG monitoring research. The purpose of this large prospective study was conducted to determine the agreement in axes and the diagnosis of myocardial infarction (MI) between the 12-lead ECGs recorded via standard and Mason-Likar systems in 133 patients who presented to Emergency Departments for evaluation of an acute MI. Method: 12-lead ECGs were first recorded by the standard immediately followed by the Mason-Likar method. Computerized measurements of waveforms and intervals were obtained. Percent of agreement was used to determine the similarities and differences in the interpretations. Results: By moving the limb electrode closer to the torso, there was a mean 27° right-ward shift in the QRS axis. A high agreement (94%) was found in the QRS axis interpretation between the 2 lead systems. Agreements in the presence or absence of abnormal Q wave duration between the 2 lead systems were higher in leads I (99%), II (96%), V4 (98%), V5 (99%), and V6 (96%) and were lower in leads aVL (90%) and aVF (86%). All subjects had 100% agreement in the presence and absence of anterior MI. The majority (87%) of the subjects also had high agreement in the presence or absence of inferior wall MI. The agreements in the presence and absence of posterior MI by widen R waves in leads V1 and V2 were also high, 99% and 95%, respectively. Conclusions: The Mason-Likar electrode sites produce a 27° rightward frontal plane QRS axis shift; however, in the majority of subjects, the QRS axis interpretation remained the same. This MasonLikar system also had a high agreement in the presence or absence of anterior, posterior, and inferior MI in this population. Moving electrodes to the body torso can produce false-positive and false-negative abnormal Q waves in lead I in 10% of subject and lead aVF in 14% of subjects. Funding source: NlH/NINR: RO1 NR008092
Poster Session I
Augmentation of the Amplitude of QRS Complexes in Patients With Heart Failure Treated With Diuresis for 48 Hours
patients with AN should await abatement of A. The above can be extrapolated to the patients with end-stage renal failure undergoing hemodialysis, and congestive heart failure, who also reveal attenuation of the ECG QRS potentials at the stage of poor compensation of their edematous state.
John E. Madias, MD,* Jessica Song, Pharm D,† C. Michael White, Pharm D,† James S. Kalus, Pharm D,† and Jeffrey Kluger, MD,† Mount Sinai School of Medicine,* and University of Connecticut,† Elmhurst, NY, and Hartford CT USA
Genesis of the Normal T-U Wave: A Slice Model of the Left Ventricular Myocardium
Augmentation in the amplitude of ECG QRS complexes (AUQRS) has been described in patients treated for heart failure (HF), but the underlying mechanism has remained enigmatic. We assessed the effect of diuresis-based fluid loss in patients treated for HF on the AU-QRS. Twenty-one patients aged 70.5⫾12.7 years, 13 with ischemic and 8 with non-ischemic cardiomyopathy, received diuresis in the hospital for exacerbated HF, and had standard ECGs recorded prior to the initiation of therapy, and at 24, and 48 hours. Percent change (%⌬) over the course of observation in the sums of the amplitude of QRS complexes from 12 leads (⌺QRS12), 6 limb leads (⌺QRS6), and leads 1⫹2 (⌺QRS2) in mm were correlated with net fluid loss, corrected for admission weight in mL/Kg. Fluid loss amounted to 3,204.9⫾1,399.5 mL, and ⌺QRS12 was 160.9⫾42.3 mm before and 170.0⫾50.7 mm after diuresis (p ⫽ 0. 024). %⌬ in ⌺QRS12, ⌺QRS6, and ⌺QRS2 correlated well with the net fluid loss (r ⫽ ⫺0.70, ⫺0.82, ⫺0.61, and p ⫽ 0.002, 0.0005, 0.001), correspondingly. AU-QRS correlates well with net fluid loss from shortterm diuresis in patients with HF, and can be used as an easily obtainable and universally available bedside index of the net fluid loss, experienced by bed-ridden patients with exacerbated HF. undergoing therapy.
Henk J. Ritsema Van Eck, Jan A. Kors, and Gerard Van Herpen, Department of Medical Informatics, Erasmus MC, Rotterdam, The Netherlands The U wave has always lived in the shadow of the preceding, bigger T wave. Its clinical relevance has been recognized but a satisfactory explanation of its origin is still outstanding. Earlier we presented a simple digital model, consisting of a single string of 12 myocardial cells. We have extended the model to a more realistic one in the form of a slice of Left Ventricle (LV), 12 cell layers thick and containing 1328 cells. To each layer an Action Potential (AP) is assigned with shape and duration (APD90) according to published data. The mid-myocardial M-cell layers have appropriately prolonged AP’s. The AP’s have slowly tapering tails, as evidenced from literature data. The depolarization sequence follows from the endocardial Purkinje activation pattern. The potential differences between the AP’s produce timevarying electrical sources between each of the 1328 myocardial cells and its neighbors. The resultant dipole vector is assigned to the cell center. Each cell dipole contributes to the potential in an arbitrary point P in or on the body. The specific lead vector linking P to that particular cell determines the strength of this contribution. The course of the potentials at P is then calculated as the sum of the contributions from all 1328 cells, as they vary with time. This is done for various distances of P to the epicardium and for various angles with respect to the LV axis. The repolarization waves constructed in this way reproduce the natural aspects of a T wave followed by a U. The size and morphology of the T-U complex depend on the position (distance and angle) of P with respect to the myocardium and on the transmural dispersion of the repolarization. Clear-cut U waves are produced with transmural dispersions of as little as 36 ms (APD90). We conclude that T and U are the resultant of one and the same repolarization process and that they form a single, continuous T-U complex. The creation of normal U waves is conditional on the presence of small voltage differences between the tail ends of the AP’s. Further investigation is necessary to elucidate the genesis of pathological T-U morphologies.
Anasarca Peripheral Edema of Varying Etiology Conceals the ECG Diagnosis of Left Ventricular Hypertrophy John E. Madias, MD, Mount Sinai School of Medicine, New York, NY USA The diagnosis of left ventricular hypertrophy (LVH) is in large part based on measurements of R-, and S-waves from various ECG leads; since anasarca peripheral edema (AN) leads to attenuation of the amplitude of QRS complexes, it could interfere with the diagnosis of LVH. ECG measurements and qualitative assessment (“Cornell”, “Sokolow-Lyon”, and “Romhilt-Estes”) were carried out in 14 patients with LVH (6 patients with AN and 8 patients who did not gain weight during their hospitalization, and served as “controls”). Patients with AN were evaluated on admission, and on the days of half of peak weight (HF-W) gain, and peak weight (P-W) gain. The “controls” were evaluated on admission and at discharge. While LVH characterization remained unaltered for the “controls”, the patients with AN revealed statistically significant drop in all ECG measurements on the days of HF-W gain (p ⫽ 0.016 – 0.0005), and P-W gain (p ⫽ 0.02 – 0.0005), on which the diagnosis of LVH could not be anymore made. The ECG-based diagnosis of LVH is interfered with by AN, and absence of LVH in
Electrocardiographic Abnormalities in Patients with Subarachnoid Hemorrhage and Normal Adults: A Comparison Study Claire E. Sommargren, Robert Warner, Jonathan G. Zaroff, Nader M. Banki, Alexander Kopelnik, Avinash A. Kothavale,
42
Poster Session I
43
Poyee P. Tung, and Barbara J. Drew, University of California, San Francisco, CA
before and after applying the adaptive filter. A percent improvement was calculated for each data set. These results are shown in the table.
A wide variety of electrocardiographic (ECG) abnormalities have been reported in patients with subarachnoid hemorrhage (SAH). The purpose of this study was to compare the ECG in patients with SAH to those in normal persons in order to identify differences in ECG characteristics between the two groups, and to determine which waveform and interval abnormalities occur more frequently in SAH. Method: A standard ECG was recorded as soon as possible after admission. Computer-assisted measurements of ST segment deviation and QTc intervals were made with the Mortara ST Review Station (Milwaukee, WI). Results: ECGs of 200 patients with SAH and 30 patients in a normal control group were analyzed. Of 237 quantitative ECG measurements, 91 were significantly different between the two groups. Mean PR interval duration was shorter in the SAH group (p⫽.000). Mean R wave amplitude was higher in lead I (p⫽.001) and lead V2 (p⫽.016) in the normal group, and higher in V6 (p⫽.003) in the SAH group. S wave amplitudes were greater in SAH patients across all six precordial leads (p⫽.003–.027). T wave axis differed significantly between the two groups (p⫽.001), and. T wave inversion was found to occur in 44 (22%) of SAH patients but only 1 (3%) of the normal control group (p⫽.013). In addition, 48 (24%) patients in the SAH group met ECG criteria for left ventricular hypertrophy, while none did so in the control group (p⫽.011). Conclusions: There are significant ECG differences between normal patients and those with SAH. It is further concluded that T wave inversion and ECG evidence of left ventricular hypertrophy occur significantly more frequently in patients with SAH than in normal patients.
Percent Improvement After Filtering
Using Motion Sensors to Reduce Motion Artifact in the ECG David A. Tong, Keith A. Bartels, and Kevin S. Honeyager, Southwest Research Institute, San Antonio, TX, USA Motion artifact is an unsolved problem in electrocardiography. Motion artifact may produce large amplitude signals in the ECG that may be misinterpreted by clinicians and automated systems resulting in misdiagnosis, prolonged procedure duration, and delayed or inappropriate treatment decisions. The hypothesis being investigated in this project is that motion artifact can be reduced by using electrode motion as a noise reference in an adaptive filter. The initial results of a novel approach to reducing motion artifact in the ECG using motion sensors attached to the ECG electrodes and adaptive filtering are presented. To measure electrode motion, two custom-designed motion sensors were developed and attached to the ECG electrode. One sensor utilized a two axis anisotrophic magnetoresistive (AMR) sensor and the other sensor utilized a three axis accelerometer (ACC). In an iterative fashion, both sensors were attached to the right arm electrode where motion was induced by (1) pushing directly on the ECG electrode, (2) pushing on the skin around the electrode, and (3) pulling on the ECG lead wire. The motion sensor signals and ECG leads I and II were recorded from eight subjects at 500 Hz. An LMS adaptive filter was implemented in Matlab and was used to post-process the electrode motion and ECG data. System performance was assessed by calculating two measures of the noise in the ECG (the L2 Norm and the MaxMin statistic)
AMR Sensor Noise Type All Data Sets Push Electrode Pull Lead Push Skin
ACC Sensor
L2 Norm
Max Min
L2 Norm
Max Min
67.6% 70.6% 45.2% 77.3%
37.7% 48.7% 24.5% 51.0%
83.6% 78.6% 91.3% 89.6%
58.5% 50.7% 49.2% 67.5%
As shown in the table, the data support the hypothesis that electrode motion can be used in an adaptive filter to reduce motion artifact in the ECG. The ACC sensor out performed the AMR sensor because the ACC sensor provided an additional axis of measurement and the relationship between the ACC signal and the motion artifact noise is better modeled by a linear time-varying system.
What Are the Implications for Using Modified (Mason-Likar) Exercise Lead System in Research? Shu-Fen Wung, PhD, RN, Amelia Sieger, BS, Monique Leon, BS, Jessica Redondo, BS, Vincent Sorrell, MD, and Steven Goldman, MD, University of Arizona, Southern Arizona VA Health Care System Tucson, Arizona, USA Researchers often use the Mason-Likar exercise lead system to maintain stability of the waveforms for continuous ECG monitoring research. The purpose of this large prospective study was conducted to determine the agreement in axes and the diagnosis of myocardial infarction (MI) between the 12-lead ECGs recorded via standard and Mason-Likar systems in 133 patients who presented to Emergency Departments for evaluation of an acute MI. Method: 12-lead ECGs were first recorded by the standard immediately followed by the Mason-Likar method. Computerized measurements of waveforms and intervals were obtained. Percent of agreement was used to determine the similarities and differences in the interpretations. Results: By moving the limb electrode closer to the torso, there was a mean 27° right-ward shift in the QRS axis. A high agreement (94%) was found in the QRS axis interpretation between the 2 lead systems. Agreements in the presence or absence of abnormal Q wave duration between the 2 lead systems were higher in leads I (99%), II (96%), V4 (98%), V5 (99%), and V6 (96%) and were lower in leads aVL (90%) and aVF (86%). All subjects had 100% agreement in the presence and absence of anterior MI. The majority (87%) of the subjects also had high agreement in the presence or absence of inferior wall MI. The agreements in the presence and absence of posterior MI by widen R waves in leads V1 and V2 were also high, 99% and 95%, respectively. Conclusions: The Mason-Likar electrode sites produce a 27° rightward frontal plane QRS axis shift; however, in the majority of subjects, the QRS axis interpretation remained the same. This MasonLikar system also had a high agreement in the presence or absence of anterior, posterior, and inferior MI in this population. Moving electrodes to the body torso can produce false-positive and false-negative abnormal Q waves in lead I in 10% of subject and lead aVF in 14% of subjects. Funding source: NlH/NINR: RO1 NR008092