electrocardiographic indexes of dispersion of ventricular repolarization: an isolated heart...

746 JACC Vol. 25, No. 3 March 1, 1995:746-52

EXPERIMENTAL STUDIES

Electrocardiographic Indexes of Dispersion of Ventricular Repolarization: An Isolated Heart Validation Study

M A R K U S Z A B E L , MD, S T U A R T P O R T N O Y , MD, M I C H A E L R. F R A N Z , MD, PI-ID, F A C C

Washington, D. C.

Objectives. This study tested the correlation of QT and JT dispersion and other potentially useful electrocardiographic (ECG) indexes with dispersion of repolarization and recovery time.

Background. Dispersion of ventricular repolarization is cur- rently being assessed noninvasively from the surface ECG by means of QT and JT dispersion, although their correlation with dispersion of repolarization as measured directly from the myocardium is not well established.

Methods. Multiple monophasic action potentials were recorded simultaneously with a 12-lead ECG from isolated Langendorff- perfused rabbit hearts. The QT and JT dispersion was compared with the dispersion of monophasic action potential duration at 90% repolarization (APD9o) and with dispersion of recovery time. As new ECG indexes, total T wave area, T wave area after the peak (late T wave area) and T peak to T end interval were also tested.

Results. The QT and JT dispersion showed a significant correlation with the dispersion of APDgo and the dispersion of recovery time (r values between 0.58 and 0.64, respectively, p < 0.001). However, total T wave area showed better correlation, respectively, with dispersion of APDgo and recovery time (r = 0.79 and r = 0.82, p < 0.0001), as did late T wave area (r = 0.81 and r = 0.81, p < 0.0001) and T peak to T end interval (r = 0.81 and r = 0.82, p < 0.0001).

Conclusions. The JT and QT dispersion correlate significantly with dispersion of APDgo and recovery time. The ECG assessment of dispersion of repolarization can be improved by three new ECG dispersion indexes: T peak to T end interval, total T wave area and late T wave area. These new indexes should be tested clinically.

(J Am Coli Cardiol 1995;25:746-52)

The importance of dispersion of repolarization in the genesis of ventricular arrhythmias has been shown in both experimental and clinical electrophysiologic studies (1-5). Clinically, dispersion of repolarization can be measured invasively using endocardial or epicardial catheter mapping (6,7) and noninvasively using extensive multilead body surface potential mapping (8-11). However, neither of these two methods is available or practical for widespread use.

The QT dispersion, calculated as the range of QT intervals (QTmax - QTmin), has been proposed as a simple noninva- sive measurement of dispersion of repolarization available from the 12-lead surface electrocardiogram (ECG) (12). Sev- eral clinical studies (12-18) have since demonstrated the potential usefulness of this variable as a predictor of arrhythmias in patients with the congenital long QT syndrome (12,18) and for monitoring the effects of antiarrhythmic drugs with class III action, including their risk of proarrhythmia (13,15-

From the Divisions of Clinical Pharmacology and Cardiology, Georgetown University and Department of Veterans Affairs Medical Center, Washington, D.C. This study was supported by a Merck International Fellowship in Clinical Pharmacology (Dr. Zabel) and by a Department of Veterans Affairs Merit Review Grant (Dr. Franz). It was presented at the 1994 North American Society. of Pacing and Electrophysiology Young Investigator Award competition, Nash- ville, Tennessee, May 1994.

Manuscript received May 12, 1994; revised manuscript received October 4, 1994, accepted October 12, 1994.

Address for correspondence: Dr. Michael R. Franz, Division of Cardiology, Department of Veterans Affairs Medical Center, 50 Irving Street, NW, Wash- ington, D.C. 20422.

17). Before widespread clinical use of this readily available ECG variable is warranted, it should be demonstrated that a sutficiently accurate assessment of dispersion of repolarization or recovery time is possible from the 12-lead ECG. Only one study (19) in abstract form is available that correlates QT dispersion with dispersion of recovery time as calculated from monophasic action potentials recorded during cardiac surgery; however, the number of patients and recordings was low in this study. Moreover, monophasic action potentials were acquired in sequence rather than simultaneously. It is also unclear whether QT and JT dispersion (JTmax - JTmin) are the optimal ECG variables for assessing dispersion of repolarization.

It is postulated from the theory of T wave genesis that the T wave vector is generated by locally different levels of repolarization in the heart (20-23), so that the T wave emerges from "inhomogeneous" recovery throughout the heart. There- fore, T wave width, which represents the time interval during which differences in repolarization occur in the heart, and T wave area, which represents a summation of vectors during this time interval, should exhibit high correlation with dispersion of repolarization. The objective of this study was to determine the correlation of various potentially useful ECG variables with the dispersion of action potential duration and recovery time recorded directly from an intact rabbit heart as a reference standard. To accomplish this, we developed a novel isolated heart preparation that was immersed in a saline-filled tank and from which we could simultaneously record seven monophasic

©1995 by the American College of Cardiology 0735-1097/95/$9.50 0735-1097(94)00446-W

JACC Vol. 25, No. 3 ZABEL ET AL. 747 March 1, 1995:746-52 ECG VARIABLES FOR DISPERSION OF REPOLARIZATION

action potential signals (estimating the value of the true reference standard that would be the difference in action potential duration at 90% repolarization (APD9o) from all cells of a single heart) and a 12-lead volume-conducted ECG.

Methods

Isolated rabbit heart preparation and experimental setup. The experimental techniques used to prepare isolated Langendorff-perfused rabbit hearts and the experimental setup have recently been described in detail (24). Experiments and animal care conformed to the "Position of the American Heart Association on Research Animal Use" adopted by the Asso- ciation in November 1984. In brief, 10 New Zealand white rabbits (3.0 to 4.0 kg) were anesthetized with intravenous pentobarbital after injection of 500 IU of intravenous heparin. After removal of the heart, the aorta was cannulated and perfused with oxygenized Tyrode's solution heated to 3°C. The heart was then mounted on a modified Langendorff apparatus and immersed in a tissue bath filled with Tyrode's solution. To slow down the intrinsic heart rate, the atrioventricular node was destroyed by applying radiofrequency energy through a pair of customized tweezers.

Monophasic action potential recordings. A plastic ring with six evenly spaced cantilever arms holding spring-loaded silver-silver chloride contact monophasic action potential electrodes (EP Technologies) was mounted around the central stub of the Langendorff setup (24). Two electrodes were placed on the right ventricular epicardium, and another four were evenly spread over the left ventricle. The heart was driven from the pacing electrode of a monophasic action potential combi- nation catheter (25) recording a seventh monophasic action potential and positioned at the right ventricular endocardium on or near the intraventricular septum throughout the exper- iment.

Twelve-lead volume-conducted ECG recordings. Four silver-silver chloride electrode pellets (In Vivo Metric) were positioned in a simulated "Einthoven" configuration with the reference and "foot" electrodes fixed beneath the heart and the "arm" electrodes fixed to the walls of the tissue bath, which had a diameter similar to that of a rabbit thorax (Fig. 1). This setup permits the recording of a volume-conducted true ECG and has recently been validated in our laboratory (24,26) for the unipolar Einthoven leads I to aVF. For this study, six additional electrodes were mounted on the anterior wall of the bath in a 90 ° circular pattern (Fig. 1) to record unipolar Wilson leads V 1 to V 6 from their simulated electrode positions. The ECGs closely resemble a body surface ECG and reflect global changes in activation and repolarization of the isolated heart (26). The signals were amplified by a custom-built ECG amplifier (Stellartech) that allowed the simultaneous recording of the signals from all 12 leads (I, II, III, aVL, aVR, aVF, V 1 t o V6) ,

Protocols. To shorten repolarization, the basic cycle length was decreased from 1,200 to 400 ms in 100-ms steps after achieving steady state at each cycle length (100 beats undis-

Tissue Bath

L 12 cm 5 [- -I

Figure 1. Schematic of electrode positions for electrocardiographic recordings, with top and side views showing all electrode positions, F = foot; LA (RA) = left (right) arm; REF = reference.

turbed by premature ventricular contractions for at least 25 beats). To modify the dispersion of repolarization, d-sotalol (Bristol Myers, Inc.) was added to the buffer solution in four experiments at increasing concentrations, and the protocol was repeated 45 rain after a new concentration was started.

Data acquisition. Recordings were acquired on a strip chart recorder (TA4000, Gould Electronics) with a paper speed of 100 mm/s. The data were simultaneously digitized at 1 kHz for storage on the hard disk of a Macintosh IIfx computer. Both strip chart recorder and digital acquisition board (National Instruments) were configured for 16 channels. Therefore, all seven monophasic action potential and nine ECG signals were acquired simultaneously. During a steady state recording, leads I to aVF, V1, V3 and V 5 were first acquired. Then, by means of a switchboard, leads V2, V 4 and V 6 were acquired instead of leads V a, V 3 and V 5.

Measurements. Two beats from each set of steady state recordings were chosen for analysis, one directly before and one directly after the switch from leads V 1, V 3 and V 5 to leads V2, V4 and V 6. The respective signals were processed channel by channel and displayed on a 20-in. (50-cm) computer screen by means of customized, interactive software programs written in LabVIEW 2.2 language (National Instruments). All analyses were performed by a single observer (M.Z.). The ECG signals were filtered selectively for 60-Hz noise, whereas monophasic

748 ZABEL ET AL. JACC Vol. 25, No. 3 ECG VARIABLES FOR DISPERSION OF REPOLARIZAT1ON March 1, 1995:746-52

_~__~, MAP 1 (RV)

: MAP 2 (RV)

MAP3(LV)

MAP 4 (LV)

~ I . MAP 5 (LV)

1 m V •

100 mseC

MAP 6 (LV)

III

a V ~ ~

aVF

V1

V2

i i ! "

V3

V4

V6 ~ 100 msec

Figure 2. Example of a recording of 6 monophasic action potential (MAP) (left) and 12 electrocardiographic (ECG) signals (right) during baseline steady state pacing at a cycle length of 1,000 ms. The six monophasic action potentials and ECG leads I to aVF, V1, V 3 and V 5 are simultaneous from a single beat; ECG leads V2, V4 and V6 were recorded 3 s later after switching. The measurement spikes are also depicted (see Methods for explanation). ECG lead II was discarded. Measurements (all in ms) from this example were QT 277, QTpeak 243, JT 219, JTpeak 186, QT dispersion 24, JT dispersion 22, T peak to T end 33, total T wave area 33.1, late T wave area 7.7, action potential duration at 90% repolarization (APD9o) 218, recovery time 248, dispersion of APDgo 22 and dispersion of recovery time 36.

action potential signals were processed unfiltered. The computer program used customized algorithms to detect important features of the monophasic action potential and ECG wave- forms (pacing spike, monophasic action potential upstroke and 90% repolarization for monophasic action potential signals; Q onset, J point, T peak and T end for ECG signals) and then displayed vertical spikes superimposed on the signal marking these points with high resolution for confirmation or manual correction by the observer. In the case of a biphasic T wave, the first peak was regarded as the T peak. Figure 2 shows an example of a recording of six simultaneous monophasic action potentials and nine ECG signals with their respective spikes. The three additional unipolar leads from the second beat analyzed are also included in the panels. If the signals were of

low quality, that is, if they had a flat or absent T wave or no clear-cut T peak, they were discarded at this point; U waves were occasionally seen during peffusion with d-sotalol. In this case, the T wave end was defined as the intersection of the tangent of the T wave downslope with the baseline. The processed data were then automatically written to a spread- sheet. Because all monophasic action potential recordings and ECG leads I to aVF were available in duplicate during one steady state recording, values were compared as a quality check. The average of the duplicates was taken if the difference was >5 ms, otherwise the measurements were repeated. To determine the reproducibility of measurements, 35 randomly chosen recordings were analyzed a second time by the same observer who had no knowledge of the results of the first analysis.

Monophasic action potential variables. Each of the monophasic action potential channels was analyzed for activation time, APD9o and recovery time. Activation time was defined as the interval from the pacing spike to the monophasic action potential upstroke, and APD90 as the interval between the monophasic action potential upstroke and the 90% repolarization level of the action potential. Recovery time was then calculated as activation time plus APD90. Dispersion of APD90 was defined as the range of values in all analyzable channels (APD90max - APD90min). Dispersion of recovery time was similarly calculated as the range of recovery times (maximal minus minimal recovery time). A minimum of five


analyzable monophasic action potentials, including the endocardial signal, was required for inclusion in the data analysis.

ECG variables. From each analyzable ECG channel, the intervals JT (J point to T end), JTpeak (J point to T peak), QT (Q onset to T end), OTpeak (Q onset to T peak) and T peak to T end were determined. Because T wave onset is usually very difficult to discern in the ECG tracings recorded in this model owing to a gradually upsloping ST segment, the T peak to T end interval was chosen as a measurement of half the T wave width. Accordingly, T wave areas were calculated either from the J point to the T end (total area) or from the T peak to the T end (late area). Like the body surface ECG, each lead was assumed to provide information about dispersion from a relatively "local" area of the heart. Therefore, an average between leads was calculated to determine "global" dispersion. From these data, JT dispersion (JTmax - JTmin) and QT dispersion (QTmax - QT min) were derived. The JT, JTpeak, QT and QTpeak intervals were each averaged from all leads. In addition, total T wave area (area between the curve and baseline from J point to T end) and late T wave area (area between the curve and baseline from T peak to T end) were measured. Because of the variation in QRS voltage between experiments, T wave areas were corrected by dividing them by the QRS voltage amplitude of each lead. As new ECG indexes for dispersion of repolarization, the total T wave area, the T wave area after the peak (late T wave area) and the T peak to T end interval (as a measure of T wave width) were each averaged from all measurable leads. A minimum of 9 of 12 analyzable ECG channels were required for inclusion in the data analysis. A relatively large variation in QRS voltages was observed over the 10 experiments in this study. Therefore, to normalize T wave areas in mV x ms, each was divided by the QRS voltage amplitude, resulting in corrected T wave values in ms. However, this does not mean that a corrected T wave area became a measurement of T wave width because the corrected value still accounts for variations of the T wave height in relation to QRS amplitude.

Final data analysis and statistics. From a total of eight experiments, 142 steady-state recordings were available for analysis. In 40 of these recordings, <9 of 12 ECG channels were analyzable and were not included in the analysis. In the remaining 102 recordings, 12 were obtained with the addition of d-sotalol to the perfusate at a concentration of 10 -6 mol/liter, 23 at 10 -5 tool/liter and 17 at 10 -4 mol/liter. In general, data are mean value _+ SD. The Pearson correlation coefficients with their respective p values were calculated for the relation between continuous variable pairs. These calcula- tions were performed for all available recordings with various cycle lengths, including baseline and drug recordings. The following pairs of variables were prospectively chosen for analysis: dispersion of APDgo and recovery time, each with JT dispersion, QT dispersion, total T wave area, late T wave area and T peak to T end interval. The difference in correlation coefficients was compared between selected pairs of variables by means of the method of Neil and Dunn (27). A p value -<0.05 was considered significant.

Table 1. Measurements for Electrocardiographic and Monophasic Action Potential Variables

All Recordings Baseline d-Sotalol (n = 102) (n = 50) (n = 52)

APD9o 211 ± 40 190 + 20 235 ± 42 (135-331) (135-252) (155-331)

Recovery time 242 _+ 39 221 - 21 268 ± 40 (165-358) (165-280) (183-358)

JT 229 + 57 199 ± 25 261 ± 62 (136-456) (136-272) (156-456)

JT peak 178 ± 45 158 +_ 24 201 ± 48 (86-323) (86-222) (102-323)

QT 290 ± 58 260 ± 27 318 ± 63 (189-514) (189-339) (210-514)

QT peak 239 ± 46 218 ± 26 267 ± 50 (139-381) (139-289) (152-381)

Dispersion of APD9o 31 ± 15 22 ± 7 42 ± 16 (10-96) (10-40) (15-96)

Dispersion of 41 ± 14 33 -+ 7 47 ± 20 recovery time (19-107) (24-53) (19-107)

Total corr T wave 39.1 ± 19.3 25.2 ± 6.7 54.2 ± 21.8 area (14.1-123.6) (14.1-41.8) (20.8-123.6)

Late corr T wave 11.3 ± 7.2 6.7 ± 2.1 16.2 : 9.0 area (2.6-53.1) (2 .6-11.6) (3.9-53.1)

Average T peak to T 51 ± 16 40 ± 6 61 ± 19 end interval (31-132) (31-64) (31-132)

Data presented are mean value +- SD (range). APD9o = action potential

duration at 90% repolarization; c o r r = corrected.

Results Range of monophasic action potential and ECG variables.

Table 1 shows average ECG and monophasic action potential measurements for all recordings (n = 102), baseline only (n = 50) and d-sotalol only (n = 52), together with their ranges. The cycle length behavior of repolarization variables is shown in Figure 3. Both cycle length and drug interventions increased the range of variables to allow better determination of correlation coefficients.

Figure 3. Cycle length (CL) behavior of repolarization variables. APD9o = action potential duration at 90% repolarization.

400

¢p

E

O

350

300

250

200

150

100

50

0

J j J

j / f

o JT Interval [] Q T Interval

APDgo v Recovery Time

i i

400 I000

CL [msec]

750 ZABEL ET AL. JACC Vol. 25, No. 3 ECG VARIABLES FOR DISPERSION OF REPOLARIZATION March 1, 1995:746-52

A 7O

O W E e- o

o el. ._m a

I - O

60

50

40

30

20

10

0

A A A A A / ~

" / ' ~ ( ~ ' A ~ r = 0.61, p< 0.001

A

i i i i i 20 40 60 80 100

Dispersion of APD90 [msec] 120

C 70

60

40

~ 2o

10

0

A A

A A A ~ A A

S A

~ e e ~: 0.58, p < 0.00t ." @

i [ i ~ i i i i 0 10 20 30 40 50 60 70 80 90

Dispersion of RT [msec]

B

140

E - - 120

100

~ 80 I,,- ~ 60

~. 40 I-,-

~ 20

o

D

z~ A

r = 0.81, p < 0.0001

i i t i i 20 40 60 80 100

Dispersion of APD90 [msec]

140

120

100

80

/k A

I--- 60 ~ .Z A A

O O~ 40 a ~ ~ q l ~ . 21p < 0.0001 __ ~O = 0 .8 I-- 20

100 0

120

20 40 60 80 100 120

Dispersion of RT [msec]

Figure 4. Scatterplots showing the relation between the dispersion of action potential duration at 90% repolarization (APD9o) as determined from multiple monophasic action potential recordings and the electrocardiographic variables QT dispersion (A) and average T peak to T end interval (B) and dispersion of recovery time with JT dispersion (C) and total T wave area (D). Baseline measurements are depicted by circles, measurements during d-sotalol by triangles. Linear correlation coefficients and their respective p values are depicted in each plot. The p values for statistical comparison between correlation coefficients are shown in Table 2.

Correlation between monophasic action potential and ECG dispersion indexes. Also listed in Table ] are average ECG and monophasic action potential measurements of dispersion indexes for all recordings (n = 102), for baseline only (n -- 50) and for d-sotalol only (n = 52). Figure 4 and Table 2 show the relation of these variables to dispersion of APD9o and recovery time, with their respective correlation coefficients and p values. Including all measurements performed at various cycle lengths, at baseline and during d-sotalol, a highly significant correlation was found for JT and QT dispersion with both dispersion of

APD9o and recovery time. Higher correlation coefficients were found for total T wave area, late T wave area and the average T peak to T end interval (except for the comparison between JT dispersion and total T wave area for correlation with dispersion of APD9o, which was of borderline significance).

Reproducibility of measurements of dispersion. The reproducibility of average JT and QT intervals was excellent, with r = 0.997 and r = 0.998 between the first and second measurements, respectively. The T peak to T end interval was also highly reproducible by the second measurement, with r = 0.97. For total T area and late T area, r = 0.97 and r = 0.95, respectively. Reproducibility of JT and QT dispersion was lower, with r values of 0.87 and 0.88, respectively.

D i s c u s s i o n

Correlation of ECG variables with dispersion of repolarization and recovery time. This study, in an isolated rabbit heart model, shows that the use of ECG variables enables a sufficiently exact assessment of the dispersion of repolariza-


Table 2. Correlation Coefficients and p Values for Correlations Between Monophasic Action Potential and Electrocardiographic Variables of Dispersion of Repolarization

Dispersion of Dispersion Recovery of APDgo Time p Value

/ ~ QT dispersion 0.61 0.59 <0.001 I r'--JT dispersion 0.64 0.58 . . . . . . < ' 0 . ~ 1 " " " ' 1

p = 0.09-1- L-Total corrected T wave area 0.79 0.82 ~ ; ~ q . ~ 9 ~ - P < 0.03

| Late corrected T wave area 0.81 0.81 . . . . . . <2.099) . . . . . I p < 0 . 0 4 p < 0.04 ~ T peak to T end interval 0.81 0.82 <0.0001

Brackets indicate p values for comparison between correlation coeificients. APD9o = action potential duration at 90% repolarization.

tion. The QT dispersion, the first variable proposed for this purpose (12), correlated significantly with dispersion of repolarization as defined by dispersion of recovery time and APD9o. The same was true for JT dispersion, a similar variable. In general, when the activation process in monophasic action potential or ECG measurements were included, r values were either slightly lower or unchanged. Therefore, conduction time most likely represents a constant in the determination of dispersion of repolarization that may or may not decrease the accuracy of the measurements. Clinical studies (15) that have determined QT and JT dispersion in parallel found no significant differences between these variables. Importantly, new ECG indexes for dispersion of repolarization measuring T wave width and area could improve the assessment of dispersion of ventricular repolarization. As shown in Table 2, the statistical comparison between correlation coefficients exhib- ited a statistically significant difference for three of the selected comparisons and was of borderline significance for the fourth comparison. A plausible explanation for this finding could be that T wave area and width are more accurate determinants of dispersion of repolarization because both reflect inhomogene- ity of repolarization not only at the end of repolarization, but also during the entire T wave. In addition, QT and JT dispersion are calculated from two single measurements (the maximum and minimum from a set of values). Determination of the T end is sometimes difficult, even using digital signal processing as in this study. A variable derived from two single values is expected to be more prone to statistical measurement error than average values over all ECG leads and over the entire duration of the T wave, as calculated for T wave area and width. Accordingly, this study showed a lower reproducibility for QT and JT dispersion than for total T wave area, late T wave area and the T peak to T end interval.

Isolated rabbit heart model. In the present study, unipolar Wilson leads V 1 to V 6 were added to the ECG recordings. To our knowledge, this is the first time that a 12-lead ECG simulating a body surface ECG was recorded from a perfused isolated heart while enabling simultaneous direct comparisons with signals recorded from the myocardium. The results and conclusions are based on the assumption that volume- conducted ECGs behave similarly to ECGs recorded from the body surface and that the results may be extrapolated to the

human heart, where multiple action potential recordings from the endocardium or epicardium are relatively difficult to obtain (6,7,19).

It may also be desirable to study the ventricular repolarization dispersion variables during sinus rhythm because ventricular pacing changes the activation sequence and thereby alters repolarization (28). However, this is impossible because of the rapid intrinsic rhythm of the rabbit heart, even after sinus node ablation. Furthermore, pacing over a large cycle length range, together with administration of d-sotalol in various concentrations, enabled us to vary both APD9o and the dispersion of repolarization over a wide range of values. Because pacing was performed continuously from one site, it is unlikely that the correlation of ECG variables with direct measurements of dispersion of myocardial repolarization was significantly dis- turbed. Cardiac memory, which is known to account for primary T wave changes, reaches a new steady state after 1 h of pacing from a constant ventricular site (28). Data collection was usually begun after this time. Accordingly, Higham et al. (19) reported a significant correlation between dispersion of monophasic action potentials and recovery time, even during short periods of pacing.

Implications for clinical use of ECG variables of dispersion of repolarization. The implications of accurate measurement of dispersion of repolarization from the 12-lead surface ECG arc important. Most of the variables can be measured manu- ally, whereas analysis of T wave area is most easily performed by digital signal processing, which modern ECGs can provide. Several clinical studies of QT dispersion have already shown that surface ECG variables are useful as a predictor of arrhythmias in patients with the congenital long QT syndrome (12,18) and for monitoring the effects of class III antiarrhythmic drugs, including their risk for proarrhythmia (13,15-17). Because of widespread availability of electrocardiography and ease of use, risk stratification may be enhanced in postinfarc- tion patients and survivors of sudden cardiac death.

Conclusions. It is possible to measure the dispersion of repolarization and recovery time by means of a 12-lead ECG with high accuracy. This study confirms the hypothesis that the ECG variables QT and JT dispersion, proposed for this use, exhibit highly significant correlation with dispersion of APD9o

752 ZABEL ET AL. JACC Vol. 25, No. 3 ECG VARIABLES FOR DISPERSION OF REPOLARIZATION March 1, 1995:746-52

and recovery time. However, the average T peak to T end interval, a measure of T wave width, and the calculated total T wave and late T wave areas exhibit better correlation with these variables. It seems warranted to test these new and readily available variables clinically because they may improve ECG assessment of dispersion of repolarization.

R e f e r e n c e s

1. Han J, Moe GK. Nonuniform recovery of excitability in ventricular muscle. Circ Res 1964;14:44-60.

2. Merx W, Yoon MS, Han J. The role of local disparity in conduction and recovery time on ventricular vulnerability to fibrillation. Am Heart J 1977;94:603-10.

3. Kuo CS, Munakata K, Reddy CP, Surawicz B. Characteristics and possible mechanism of ventricular arrhythmia dependent on the dispersion of action potential durations. Circulation 1983;67:1356-67.

4. Kuo CS, Reddy CP, Munakata K, Surawicz B. Mechanism of ventricular arrhythmias caused by increased dispersion of repolarization. Eur Heart J 1985;6 Suppl D:63-70.

5. Vassallo JA, Cassidy DM, Kindwall KE, Marchlinski FE, Josephson ME. Nonuniform recovery of excitability in the left ventricle. Circulation 1988; 78:1365-72.

6. Cowan JC, Hilton CJ, Griffiths CJ, et al. Sequence of epicardial repolarisa- tion and configuration of the T wave. Br Heart J 1988;60:424-33.

7. Franz MR, Bargheer K, Rafflenbeul W, Haverich A, Lichtlen PR. Monopha- sic action potential mapping in human subjects with normal electrocardio- grams: direct evidence for the genesis of the T wave. Circulation 1987;75: 379-86.

8. Sylven JC, Horacek BM, Spencer CA, Klassen GA, Montague TJ. QT interval variability on the body surface. J Electrocardiol 1984;17:179-88.

9. Mirvis DM. Spatial variation of QT intervals in normal persons and patients with acute myocardial infarction. J Am Coil Cardiol 1985;3:625-31.

10. De Ambroggi L, Bertoni T, Locati E, Stramba-Badiale M, Schwartz PJ. Mapping of body surface potentials in patients with the idiopathic long QT syndrome. Circulation 1986;74:1334-45.

11. De Ambroggi L, Negroni MS, Monza E, Bertoni T, Schwartz PJ. Dispersion of ventricular repolarization in the long QT syndrome. Am J Cardiol 1991;68:614-20.

12. Day CP, McComb JM, Campbell RW. QT dispersion: an indication of arrhythmia risk in patients with long QT intervals. Br Heart J 1990;63:342-4.

13. Day CP, McComb JM, Matthews J, Campbell RW. Reduction in QT

dispersion by sotalol following myocardial infarction. Eur Heart J 1991;12: 423-7.

14. Day CP, McComb JM, Campbell RW. QT dispersion in sinus beats and ventricular extrasystoles in normal hearts. Br Heart J 1992;67:39-41.

15. Hii JTY, Wyse DG, Gillis AM, Duff HJ, Solylo MA, Mitchell LB. Precordial QT interval dispersion as a marker of torsade de pointes: disparate effects of class Ia antiarrhythmic drugs and amiodarone. Circulation 1992;86:1376-82.

16. Cui G, Sen L, Sager P, Uppal P, Singh BN. Comparison of QT dispersion and RR interval induced by sematilide, amiodarone and sotalol [abstract]. Circulation 1992;86 Suppl I:I-393.

17. Hohnloser SH, van de Loo A, Kalusche D, Arendts W, Quart B. Does sotalol-induced alteration of QT dispersion predict effectiveness or proar- rhythmic hazards? [abstract]. Circulation 1993;88 Suppl I:I-397.

18. Priori S, Napolitano C, Diehl L, Schwartz PJ. Dispersion of the QT interval--a marker of therapeutic efficacy in the idiopathic long QT syndrome. Circulation 1994;89:1681-9.

19. Higham PD, Hilton CJ, Aitcheson JD, Furniss SS, Bourke JP, Campbell RWF. Does QT dispersion reflect dispersion of ventricular recovery? [abstract]. Circulation 1992;86 Suppl I:I-392.

20. Wilson FN, MacLeod AG, Barker PS, et al. The determination and the significance of the areas of the ventricular deflections of the electrocardiogram. Am Heart J 1934;10:46-61.

21. Harumi K, Burgess MJ, Abildskov JA. A theoretic model of the Y wave. Circulation 1966;34:657-68.

22. Spach MS, Barr RC. Origin of epicardial ST-T wave potentials in the intact dog. Circ Res 1976;39:475-87.

23. Franz MR, Bargheer K, Costard-J/iclde A, Miller DC, Lichtleo PR. Human ventricular repolarization and T wave genesis. Prog Cardiovasc Dis 1991;33: 369-84.

24. Franz MR, Cima R, Wang D, Profitt D, Kurz R. Electrophysiological effects of myocardial stretch and mechanical determinants of stretch-activated arrhythmias [published erratum appears in Circulation 1992;86:1663]. Cir- culation 1992;86:968-78.

25. Franz MR, Chin MC, Sharkey HR, Griffin JC, Scheinman MM. A new single catheter technique for simultaneous measurement of action potential duration and refractory period in vivo. J Am Coil Cardiol 1990;16:878-86.

26. Kotler B, Franz MR. Propafenone is not a Ic agent but moricizine is: new electrophysiologic and electrocardiographic data from the isolated rabbit heart [abstract]. PACE 1992;15 Suppl II:II-594.

27. Nell JJ, Dunn OJ. Equality of dependent correlation coefficients. Biometrics 1975;31:531-43.

28. Costard-J~ickle A, Goetsch B, Antz M, Franz MR. Slow and long-lasting modulation of myocardial repolarization produced by ectopic activation in isolated rabbit hearts: evidence for cardiac memory. Circulation 1989;80: 1412-20.

electrocardiographic indexes of dispersion of ventricular repolarization: an isolated heart...

Documents