Jpn Circ J 1998; 62: 844–848

Decrease in the High Frequency QRS Components Depending on the Local Conduction Delay

Tetsu Watanabe, MD; Michiyasu Yamaki, MD; Hidetada Tachibana, MD; Isao Kubota, MD; Hitonobu Tomoike, MD

The high frequency components contained in the QRS complex (HF-QRS) are a powerful indicator for the risk of sudden cardiac death. However, it is controversial whether conduction delay increases or decreases the HF- QRS. In 21 anesthetized, open-chest dogs, the right atrium was constantly paced. A cannula was inserted into the left anterior descending artery and flecainide, lidocaine or disopyramide was infused to slow the local conduc- tion. Sixty unipolar electrograms were recorded from the entire ventricular surface and were signal-averaged. Data were filtered (30–250Hz) by using fast-Fourier transform. The HF-QRS was calculated by integrating the filtered QRS signal. Activation time (AT; dV/dt minimum) was delayed and the HF-QRS was reduced in the area perfused by flecainide, lidocaine or disopyramide. The percent increase in AT closely correlated the percentage decrease in the HF-QRS; the correlation coefficients were 0.75, 0.83 and 0.76 for flecainide, lido- caine and disopyramide infusion, respectively, (p<0.001). Decrease in the HF-QRS linearly correlated with the local conduction delay. This study proved that conduction delay decreases the HF-QRS, and that the HF-QRS is a potent indicator of disturbed local conduction. (Jpn Circ J 1998; 62: 844–848) Key Words: Disopyramide; Fast-Fourier transform; Flecainide; Lidocaine; Signal averaging

ate potentials, or high-frequency low-amplitude anesthetized with sodium pentobarbital (30mg/kg, iv) and signals in the terminal portion of the QRS received supplemental doses as needed. Dogs were venti- L complex, are considered to be a powerful indicator lated by a respirator with room air supplemented with for the risk of ventricular and sudden cardiac oxygen (3–5 L/min). The thorax was opened at the fifth death in patients with ischemic heart disease. Studies on intercostal space, then the was opened and a experimental chronic led researchers to believe pericardial cradle was made to support the heart in the that this originated in the slowed conduction.1–3 Recently, appropriate position. The sinus node was crushed, and the frequency-domain analysis has been used as a reliable and right atrium was paced at a cycle length of 400msec using quantitative tool for detecting high-frequency low-ampli- a model SEN-7203 stimulator (Nihon Koden, Tokyo, tude signals on ECG. Several reports have shown that Japan). After an intravenous bolus administration of acute4–7 or chronic myocardial ischemia8–10 does not heparin (10,000 IU), a 24-gauge plastic cannula was increase, but there is a decrease in the high frequency inserted into the left anterior descending artery (LAD) at components in the QRS complex (HF-QRS). the distal site of the second diagonal branch. The cannula However, when examining the causality of the HF-QRS, was kept open by continuous infusion of saline at 1ml/min. myocardial ischemia may be not an ideal tool, because A globe-shaped electrode array was placed on a ventricular myocardial ischemia includes complicated electrical surface for simultaneous monitoring of electrograms from changes such as the altered potential for potassium ion, and 60 epicardial sites. Each unipolar electrode consisted of anaerobic metabolism, which leads to intracellular acidifi- fine silver wire (0.2-mm diameter) sutured to the sock. The cation and an increase in cellular osmolarity.11–13 In the electrode array was 6 rows (1–6) and 10 columns (A–J). present study, we evoked a conduction delay by using All recording electrodes were referenced to the Wilson’s sodium channel blockers and examined the direct relation- central terminal, and multichannel electrograms were digi- ship between the HF-QRS and the conduction delay. The tized every millisecond using a multiplexed data process- purpose of this study was to define whether the conduction ing system (CD-G015, Chunichi Denshi, Nagoya, Japan) as delay increases or decreases the HF-QRS. described in a previous study.14 The thoracic cavity was covered with plastic wrap to prevent cooling and dehumid- ifying. Body temperature was maintained at 37–38°C. An Methods arterial line was inserted into the right femoral artery to Instrumentation continuously monitor the mean arterial pressure. The elec- Twenty-one adult mongrel dogs (weight, 12–33kg) were trocardiogram lead II and blood pressure were recorded simultaneously throughout the study on a model 2G66 (Received May 11, 1998; revised manuscript received July 21, 1998; recorder (NEC San-ei, Tokyo, Japan). accepted August 5, 1998) First Department of Internal Medicine, Yamagata University School Experimental Protocol of Medicine, Yamagata, Japan Flecainide (low dose: 10μg/kg per min; high dose: Mailing address: Tetsu Watanabe, MD, First Department of Internal Medicine, Yamagata University School of Medicine, 2-2-2 Iida-Nishi, 100μg/kg per min, n=7), lidocaine (low dose: 0.12mg/kg Yamagata 990-9585, Japan. E-mail: [email protected] per min; high dose: 0.6mg/kg per min, n=7) or disopyra- u.ac.jp mide (low dose: 20μg/kg per min; high dose: 200μg/kg per

Japanese Circulation Journal Vol.62, November 1998 Conduction Delay and High Frequency QRS Components 845

Fig1. Representative maps of activation times at control, and during low-dose and high-dose infusion of flecainide. (*)The site of ‘the earliest activation’. Arrows indicate the epicardial electrograms at the center of perfused region: lead ‘D2’ and non-perfused region: lead ‘H2’.

Fig2. The HF-QRS area maps at control, and during low-dose and high-dose infusion of flecainide in the same each experiment as indicated in Fig1. Arrows indicate filtered QRS complexes at the center of the perfused region: lead ‘D2’ and non-perfused region: lead ‘H2’.

Japanese Circulation Journal Vol.62, November 1998 846 WATANABE T et al.

Fig3. Changes in ATs (A) and the HF-QRS (B) for each sodium channel blocker at the center of the perfused region. ATs were signifi- cantly delayed by a high-dose infu- sion of each sodium channel blocker. Low-dose disopyramide infusion significantly delayed ATs. The HF-QRS was reduced by infu- sion of each sodium channel blocker. Low-dose lidocaine and each high-dose sodium channel blocker significantly reduced the HF-QRS. AT, activation time; HF- QRS, high frequency QRS compo- nent; *p<0.05 vs control, †p<0.05 vs low dose. min, n=7) was intracoronarily infused by an infusion pump components was calculated by integrating the electrogram (model SP-100, JMS, Hiroshima, Japan). The lower dose during the QRS interval,16 which we called ‘HF-QRS’. of flecainide, disopyramide or lidocaine was 1% of the Epicardial activation of each electrogram was defined as intravenous dose in experimental study and the higher dose the minimum derivative of the QRS signal.18 The earliest was 5–10%.15 After baseline measurements, the low-dose activation measured on the cardiac surface was assigned to protocol was performed during the first 10 min, and the time zero and the activation time (AT) was determined high one was continued for the next 10 min. Epicardial from time zero to each activation, and an isochronal map electrograms were recorded 10 min after beginning the was constructed.14 Changes in AT and the HF-QRS for each low-dose infusion. When wide QRS complexes, which sodium channel blocker were measured at the center of the were produced by each sodium channel blockade, or spon- perfused region. taneous ventricular tachyarrhythmias were recognized, electrograms were immediately recorded. Statistical Analysis Quantitative data are reported as mean±SE. Statistical Analysis of Multichannel Epicardial Electrograms analysis was performed with the Wilcoxon signed-ranks Electrograms were averaged to cancel random noise by test. A confidence level of 95% was considered statistically means of a SUN 4/2 microcomputer (SUN Microsystems, significant. Mountain View, CA, USA) as previously reported.14 The cross-correlation function was used to reject ectopic beats and artifacts (cross-correlation coefficients below a thresh- Results old value: 0.95).16 The number of beats averaged was from Effects of Sodium Channel Blockers on AT and the HF- 50 to 56 (mean, 54). The signal-averaged ECG data were QRS Area stored on a magneto-optical disk for later analysis. A root The 21 dogs were divided into 3 groups: 7 were given mean square (RMS) voltage based on all 60 leads was flecainide, 7 lidocaine, 7 disopyramide. Fig1 illustrates the plotted to determine both the onset and offset of QRS. The cardiac surface distribution of the ATs before (control) and flat portion of the PR segment was defined as the zero after flecainide infusion (low and high dose of flecainide) level. The RMS values were computed from the averaged in a representative experiment with an asterisk indicating (but not filtered) ECG wave forms. the site of ‘the earliest activation’. The delay in ATs was For analysis with the fast-Fourier transform algorithm not apparent on the maps at control or during low-dose (FFT), ECG data were padded with zeros for a total 1024 infusion of flecainide. However, a marked delay in ATs points. A time series f(t) representing the ECG complex was seen in the perfused area during high-dose infusion. was subjected to a 1,024-point FFT analysis to produce a The configuration of the electrogram showed a widened frequency series F(w). A product of F(w) with a transfer QRS complex and an increased R wave in the perfused function, H(w), was computed to produce a filtered wave area. G(w) in a frequency domain. Fig2 shows the HF-QRS maps before and after low and G(w)=F(w)H(w), where H(w) is given by the following high-dose infusion of flecainide. With high-dose flecainide, function, previously reported by Abboud et al17 a decrease in the HF-QRS was clearly observed in the perfused area. This region corresponded to the area H(w) = [1 1+(WL/w)n]{1–[1/ 1+(WH/w)m]} showing conduction delays in the AT map (Fig1). In Fig2, it can also be seen that the amplitude of the high frequency with WL and WH being the low and high cut-off frequency components of the QRS complex markedly decreased with of the filter. The frequency range we used was 30–250Hz, high-dose flecainide infusion. designated as the ‘high frequency component’. Parameters Fig 3A shows changes in the ATs for each sodium m and n determine the decay slope of the filter. In this channel blocker; flecainide, lidocaine or disopyramide, study, parameter m–n was 4–100, respectively. A filtered measured at the center of the perfused region. ATs were wave form g(t) was produced by transformation (inverse significantly delayed by the high-dose infusion of each FFT) of G(w) to the time domain. The high frequency QRS sodium channel blocker (p<0.05 vs control). In case of

Japanese Circulation Journal Vol.62, November 1998 Conduction Delay and High Frequency QRS Components 847

tion in themyocardium has been required. Thus, the present study used sodium channel blockers in order to produce local conduction disturbances without ischemia. The results strongly suggest that conduction delay decreases the HF-QRS. The mechanism of the attenuation of the high frequency components due to slowed ventricular conduction is still unclear. Abboud et al simulated the phenomenon on a 3 dimensional heart model; slowing of the conduction veloc- ity produced a decreased amplitude of the high frequency component in the QRS complex in the model heart.20 They suggested that slowed conduction may decrease the number of terminal branches in the activation process. This causes synchronized activation in a large amount of myocardium, which shifts high-frequency activation to a lower frequency. In the present study, we used 3 different types of sodium channel blockers to minimize their different effects on the Fig 4. Correlation between AT and the HF-QRS at control, and myocardium, because heterogeneity in sodium channel during low-dose and high-dose infusion of flecainide, lidocaine or blockers has been noticed recently (eg, different binding disopyramide. Correlation coefficients: flecainide, r=0.75, p<0.001; kinetics and affinities) depending on the channel state lidocaine, r=0.83, p<0.001; disopyramide, r=0.76, p<0.001. (open or closed). The results showed similar effects among the sodium channel blockers in the relation between the HF-QRS and the conduction delay, which indicated that disopyramide, a low dose significantly delayed ATs differences in the subclass of sodium channel blockers did (p<0.05 vs control). High-dose sodium channel blockers not interfere with the relation. significantly reduced the HF-QRS in the perfused region (Fig3B). Infusions of low-dose lidocaine also reduced the Clinical Implication: Relation to Clinical Observation HF-QRS (p<0.05 vs control). Prolongation of QRS duration has been recognized in exercise-induced myocardial ischemia.21–23 It reflects Correlation Between AT and the HF-QRS delayed conduction, and is thought to be an indicator of The relation between the degree of increased AT and multiple diseased vessels and exercise-induced wall motion decreased HF-QRS is shown in Fig 4. The percentage abnormalities in patients with .22 increase in AT (%AT) closely correlated with the percent- However, the sensitivity of this parameter for detecting age decrease in the HF-QRS (%HF-QRS); correlation coef- delayed conduction is not high. We previously reported a ficients for flecainide, lidocaine and disopyramide infusion method for detecting local activation using a multiple were 0.75 (p<0.001), 0.83 (p<0.001) and 0.76 (p<0.001), unipolar lead system,24–27 which allowed us to evaluate the respectively. degree of conduction delay quantitatively, from body24,26 or cardiac surface electrodes.28 The constructed activation sequence clearly detected the abnormality of local activa- Discussion tion in myocardial ischemia,24 infarction25,27 or ventricular The conduction delay was locally evoked by the intra- hypertrophy,26 which indicated that the local conduction coronary administration of a sodium channel blocker. The delay could predict the severity of exercise-induced results showed that decreased levels of the HF-QRS are myocardial ischemia,24 the site of tiny myocardial infarc- well correlated with the degree of delayed AT. tion,25 or the severity of hypertrophy.26 However, to obtain this information, a complicated system is required. The Relation of the HF-QRS to Local Delay in Ventricular present study may provide an alternative method for detect- Conduction ing local activation by use of a limited-numbered lead Local conduction delay has been presumed as the origin system. of the high frequency component during QRS complex The reduction of the HF-QRS is possibly correlated with from studies on acute or chronic myocardial ischemia.4–6,8 arrhythmogenecity. We previously reported a relationship However, whether the conduction delay increases1–3,19 or between arrhythmogenecity and a reduction of the high decreases8,10 the HF-QRS has been controversial and can frequency QRS components in patients with chronic not be solved only by studies on myocardial ischemia, .16 Because the conduction delay is because myocardial ischemia produces various electro- one of the most important factors for the re-entrant circuit, physiological changes, accompanying a decrease in the HF-QRS may be a potent indicator of the risk for conduction velocities. Cellular electrophysiology suggests cardiovascular events. However, further examination will the following probable changes in ischemic myocardium: be needed to confirm the clinical significance of the HF- (1) a depolarization of the resting membrane potential; (2) QRS. diminishing the upstroke velocity, amplitude, and duration of action potential; (3) an accumulation of K+ in the extra- Acknowledgments cellular space, which also inhibits the Na+/K+ pump; and + This study was supported in part by Grant-in-Aid for Scientific (4) acidification (increase H ) in both intracellular and Research from the Ministry of Education, Science, Culture and Sports, 4 extracellular compartments. To avoid these effects, a Japan (08670756, 08457200) and a grant from Suzuken Memorial condition, other than ischemia, that creates slow conduc- Foundation.

Japanese Circulation Journal Vol.62, November 1998 848 WATANABE T et al.

References J Am Coll Cardiol 1997; 29: 1639–1644 16. Hosoya Y, Kubota I, Shibata T, Yamaki M, Ikeda K, Tomoike H: 1. Anderson GJ, Blieden MF: The high frequency electrocardiogram in Spectral analysis of 87-lead body surface signal-averaged ECGs in coronary artery disease. Am Heart J 1975; 89: 349–358 patients with previous anterior myocardial infarction as a marker of 2. Flowers NC, Horan LG, Tolleson WJ, Thomas JR: Localization of ventricular . Circulation 1992; 85: 2060–2064 the site of myocardial scarring in man by high-frequency compo- 17. Abboud S, Belhassen B, Miller HI, Sadeh D, Laniado S: High nents. Circulation 1969; 40: 927–934 frequency using an advanced method of signal 3. Flowers NC, Horan LG, Thomas JR, Tolleson WJ: The anatomic averaging for non-invasive detection of coronary artery disease in basis for high-frequency components in the electrocardiogram. patients with normal conventional electrocardiogram. J Electrocardiol Circulation 1969; 39: 531–539 1986; 19: 371–380 4. Abboud S: High-frequency electrocardiogram analysis of the entire 18. Spach MS, Miller W3, Dolber PC, Kootsey JM, Sommer JR, Mosher QRS in the diagnosis and assessment of coronary artery disease. CJ: The functional role of structural complexities in the propagation Prog Cardiovasc Dis 1993; 35: 311–328 of depolarization in the atrium of the dog: Cardiac conduction distur- 5. Abboud S, Cohen RJ, Selwyn A, Ganz P, Sadeh D, Friedman PL: bances due to discontinuities of effective axial resistivity. Circ Res Detection of transient myocardial ischemia by computer analysis of 1982; 50: 175–191 standard and signal-averaged high-frequency electrocardiograms in 19. Langner PJ, Geselowitz DB, Briller SA: Wide band recording of the patients undergoing percutaneous transluminal coronary angioplasty. electrocardiogram and coronary heart disease. Am Heart J 1973; 86: Circulation 1987; 76: 585–596 308–317 6. Mor AV, Shargorodsky B, Abboud S, Laniado S, Akselrod S: Effects 20. Abboud S, Berenfeld O, Sadeh D: Simulation of high-resolution of coronary occlusion on high-frequency content of the epicardial QRS complex using a ventricular model with a fractal conduction electrogram and body surface electrocardiogram. Circulation 1987; system: Effects of ischemia on high-frequency QRS potentials. Circ 76: 237–243 Res 1991; 68: 1751–1760 7. Berkalp B, Oral D, Caglar N, Omurlu K, Pamir G, Alpman A, et al: 21. Michaelides AP, Boudoulas H, Vyssoulis GP, Skouros C, Toutouzas Effects of percutaneous transluminal coronary angioplasty on late PK: Exercise-induced S-wave prolongation in left anterior descend- potentials and high frequency mid-QRS potentials. Cardiology 1994; ing coronary artery stenosis. Am J Cardiol 1992; 18: 1407–1411 85: 216–221 22. Michaelides A, Ryan JM, VanFossen D, Pozderac R, Boudoulas H: 8. Goldberger AL, Bhargava V, Froelicher V, Covell J: Effect of Exercise-induced QRS prolongation in patients with coronary artery myocardial infarction on high-frequency QRS potentials. Circulation disease: a marker of myocardial ischemia. Am Heart J 1993; 6: 1981; 64: 34–42 1320–1325 9. Goldberger AL, Bhargava V, Froelicher V, Covell J, Mortara D: 23. Lenihan DJ, Coyne E, Feldman B, Black R, Collins G: Frequency of Effect of myocardial infarction on the peak amplitude of high late potentials on signal-averaged electrocardiograms during thal- frequency QRS potentials. J Electrocardiol 1980; 13: 367–371 lium stress testing in coronary artery disease. Am J Cardiol 1992; 70: 10. Bhargava V, Goldberger A: Myocardial infarction diminishes both 432–435 low and high frequency QRS potentials: power spectrum analysis of 24. Igarashi H, Yamaki M, Kubota I, Ikeda K, Matsui M, Tsuiki K, et al: lead II. J Electrocardiol 1981; 14: 57–60 Relation between localization of coronary artery disease and local 11. Kléber AG, Fleischhauer J, Cascio WE: Ischemia-induced propaga- abnormalities in ventricular activation during exercise tests. tion failure in the heart. In: Zipes DP, Jalife J, editors. Cardiac elec- Circulation 1990; 81: 461–469 trophysiology: from cell to bedside. Philadelphia: WB Saunders, 25. Ikeda K, Kubota I, Igarashi A, Yamaki M, Tsuiki K, Yasui S: 1995: 174–181 Detection of local abnormalities in ventricular activation sequence 12. Kléber AG: Extracellular potassium accumulation in acute myocar- by body surface isochrone mapping in patients with previous dial ischemia. J Mol Cell Cardiol 1984; 16: 389–394 myocardial infarction. Circulation 1985; 72: 801–809 13. Shaw RM, Rudy Y: Electrophysiologic effects of acute myocardial 26. Yamaki M, Ikeda K, Kubota I, Nakamura K, Hanashima K, Tsuiki ischemia: A mechanistic investigation of action potential conduction K, et al: Improved diagnostic performance on the severity of left and conduction failure. Circ Res 1997; 80: 124–138 with body surface mapping. Circulation 14. Kubota I, Yamaki M, Shibata T, Ikeno E, Hosoya Y, Tomoike H: 1989; 79: 312–323 Role of ATP-sensitive K+ channel on ECG ST segment elevation 27. Konta T, Ikeda K, Kubota I, Tsuiki K, Yasui S: Relationship during a bout of myocardial ischemia: A study on epicardial mapping between late potentials and left ventricular function in patients with in dogs. Circulation 1993; 88: 1845–1851 coronary artery disease. Jpn Circ J 1988; 52: 105–110 15. Kondo T, Yamaki M, Kubota I, Tachibana H, Tomoike H: 28. Haugland FN, Johnson SB, Packer DL: Direct quantification of Electrophysiological effects of sodium channel blockades on apparent binding indices from quinidine-induced in vivo conduction anisotropic conduction and conduction block in canine myocardium. delay in canine myocardium. J Clin Invest 1994; 93: 1798–1811

Japanese Circulation Journal Vol.62, November 1998